**Sustainable Combustion Systems and Their Impact**

Editors

**S.M. Ashrafur Rahman Islam Md Rizwanul Fattah**

MDPI • Basel • Beijing • Wuhan • Barcelona • Belgrade • Manchester • Tokyo • Cluj • Tianjin

*Editors* S.M. Ashrafur Rahman School of Mech., Medical & Process Engineering Queensland University of Technology Brisbane Australia

Islam Md Rizwanul Fattah School of Information, Systems and Modelling University of Technology Sydney Australia

*Editorial Office* MDPI St. Alban-Anlage 66 4052 Basel, Switzerland

This is a reprint of articles from the Special Issue published online in the open access journal *Energies* (ISSN 1996-1073) (available at: www.mdpi.com/journal/energies/special issues/ Sustainable Combustion Systems).

For citation purposes, cite each article independently as indicated on the article page online and as indicated below:

LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. *Journal Name* **Year**, *Volume Number*, Page Range.

**ISBN 978-3-0365-1396-6 (Hbk) ISBN 978-3-0365-1395-9 (PDF)**

© 2021 by the authors. Articles in this book are Open Access and distributed under the Creative Commons Attribution (CC BY) license, which allows users to download, copy and build upon published articles, as long as the author and publisher are properly credited, which ensures maximum dissemination and a wider impact of our publications.

The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons license CC BY-NC-ND.

## **Contents**


## **About the Editors**

## **S.M. Ashrafur Rahman**

Dr S M Ashrafur Rahman is currently working as an Environmental Consultant at Trinity Consultants Australia. He is also working as a Visiting Fellow at Queensland University of Technology (QUT). He obtained his Doctoral degree in Mechanical Engineering from QUT in 2018. Previously, he completed his Master of Engineering Science from the University of Malaya in 2015 and Bachelor of Science in Mechanical Engineering from Bangladesh University of Engineering and Technology in 2012. His research work focuses on evaluating various bio-oils as possible supplements to petro-diesel.

#### **Islam Md Rizwanul Fattah**

Dr IMR Fattah is a Postdoctoral Research Fellow in the School of ISM, Faculty of Engineering and IT at the University of Technology Sydney (UTS), researching the effective use of waste for sustainable energy applications. He has accomplished his Ph.D. in reducing PM and soot emissions from diesel combustion from the University of New South Wales (UNSW, Sydney) in 2019. Previously, he completed his Master of Engineering Science from the University of Malaya (UM) in 2014 and Bachelor of Science in Mechanical Engineering from Bangladesh University of Engineering and Technology (BUET) in 2011. He has been actively engaged in the field since 2012 by publishing over 60 articles and gaining over 4500 citations of his works. He serves as an Editorial board member at *Energies* (MDPI). He also has peer-reviewed over 200 journal articles of many WOS-indexed journals throughout his career.

*Article*

## **Maximising Yield and Engine E**ffi**ciency Using Optimised Waste Cooking Oil Biodiesel**

**Luqman Razzaq <sup>1</sup> , Shahid Imran <sup>1</sup> , Zahid Anwar <sup>1</sup> , Muhammad Farooq <sup>1</sup> , Muhammad Mujtaba Abbas 1,\* , Haris Mehmood Khan <sup>2</sup> , Tahir Asif <sup>1</sup> , Muhammad Amjad <sup>1</sup> , Manzoore Elahi M. Soudagar <sup>3</sup> , Nabeel Shaukat <sup>2</sup> , I. M. Rizwanul Fattah 4,\* and S. M. Ashrafur Rahman 5,\***


Received: 6 October 2020; Accepted: 11 November 2020; Published: 13 November 2020

**Abstract:** In this study, waste cooking oil (WCO) was used as a feedstock for biodiesel production, where the pretreatment of WCO was performed using mineral acids to reduce the acid value. The response surface methodology (RSM) was used to create an interaction for different operating parameters that affect biodiesel yield. The optimised biodiesel yield was 93% at a reaction temperature of 57.50 ◦C, catalyst concentration 0.25 *w*/*w*, methanol to oil ratio 8.50:1, reaction stirring speed 600 rpm, and a reaction time of 3 h. Physicochemical properties, including lower heating value, density, viscosity, cloud point, and flash point of biodiesel blends, were determined using American Society for Testing and Materials (ASTM) standards. Biodiesel blends B10, B20, B30, B40, and B50 were tested on a compression ignition engine. Engine performance parameters, including brake torque (BT), brake power (BP), brake thermal efficiency (BTE), and brake specific fuel consumption (BSFC) were determined using biodiesel blends and compared to that of high-speed diesel. The average BT reduction for biodiesel blends compared to HSD at 3000 rpm were found to be 1.45%, 2%, 2.2%, 3.09%, and 3.5% for B10, B20, B30, B40, and B50, respectively. The average increase in BSFC for biodiesel blends compared to HSD at 3500 rpm were found to be 1.61%, 5.73%, 8.8%, 12.76%, and 18% for B10, B20, B30, B40, and B50, respectively.

**Keywords:** biodiesel; waste cooking oil; transesterification; response surface methodology; central composite design

## **1. Introduction**

Over the last few decades, the rapid decline of fossil fuels has become a significant problem. On the other hand, the energy demand is continuously rising owing to the rapidly expanding population coupled with the increased rate of urbanisation [1–3]. This scenario demands the adoption

of alternative energy resources to address the issues related to the energy as well as the environment [4–7]. Recently, biofuels produced from a variety of naturally occurring resources [8,9] have emerged as an alternative energy source owing to their comparable physicochemical, performance, and emission characteristics [10–12]. The main hindrance in the commercialisation of biodiesel is its production cost, and the production cost mainly depends upon the feedstock oil used. The use of edible oils for the production of biodiesel may cause an imbalance in the food chain, and has sparked the 'food vs. fuel' debate for decades. Waste cooking oil (WCO) can be a potential alternative to edible oil for biodiesel production. WCO, when disposed of without treatment, causes environmental pollution. The use of WCO for biodiesel production not only reduces the production cost of biodiesel, but also decreases the environmental burden.

Generally, biodiesel can be used as a blend (with or without additives) with diesel in diesel engines without any modification in the engines [13,14]. Biodiesel can be converted to fatty acid esters by reacting fatty acids and short-chain alcohol through a process known as transesterification [15,16]. Homogeneous alkaline catalysts such as KOH, NaOH, and CH3ONa have been used as catalysts in this process [17,18]. Methanol is suitable for the transesterification process [19]. The biodiesel yield depends upon the operating parameters of the transesterification process, which include temperature, catalyst concentration, methanol to oil ratio, reaction speed, and reaction time [20]. The operating parameters should be optimised to obtain the optimum yield of fatty acid methyl ester (FAME) [21]. The effectiveness of the transesterification process is assessed by the reaction kinetics, mass transfer, and equilibrium in the reaction mixture. The results of this process are used to predict the conversion yield and design a model which can be used for the prediction of conversion yield [22].

Performance of CI engines operated with biodiesel depends upon different factors including, but not limited to, compression ratio, injection timing and injection pressure [23,24]. A slight reduction in brake thermal power and brake torque was observed when biodiesel tested on the six cylinder DI diesel engine [25]. Biodiesel blends significantly reduced emissions such as carbon monoxide, carbon dioxide with a slight increase in oxides of nitrogen NOx [26,27]. A previous study by Nirmala et al. [28] tested pure WCO based biodiesel (WCOBD) and compared the results with that of pure diesel. WCOBD showed 4.2% higher BSFC, 3.6% lower BTE, and 10.8% lower BP than conventional diesel. Akcay et al. [29] studied hydrogen addition to intake air in conjunction with 25% WCO biodiesel blend (B25). The effects of hydrogen on the BSFC of B25 fuel were not significant compared to diesel fuel. Can [30] studied the engine performance of 5% and 10% blend of WCO with diesel and reported a slight increase in BSFC (up to 4%) and a small reduction of BTE (up to 2.8%) with the addition of the biodiesel for all tested engine loads.

Many techniques have been reported to optimise biodiesel yield. RSM is one of the most widely employed techniques [31–34]. RSM is a mathematical technique used for empirical model building. Yield is known as the response which depends upon independent variables which are the operating parameters of the transesterification process. RSM develops a suitable experimental design model to provide optimum operating conditions [35]. Jamshaid et al. [36] used the RSM technique for the optimisation of biodiesel yield from cottonseed oil and reported an optimum biodiesel yield of 98.3%. Mostafaei et al. [32] examined the effect on the biodiesel yield of the independent variables reactor diameter, ultrasound strength, and liquid height. The RSM technique was used to develop interaction among these independent variables. Anwar et al. [37] investigated the yield optimisation of second-generation biodiesel produced from Australian native stone fruit oil using the RSM technique, and thereby an optimum yield of 95.8% was obtained. This article focuses on optimising the biodiesel production process from WCO of Pakistani origin using the RSM technique that involved CCD [38].

The prevalent energy crisis has adversely affected the global economy. The economies of many developing countries, like Pakistan, have gone uncompetitive due to the shortage of usable energy. With large agriculture land and a population of over two hundred and twenty million, Pakistan is one of those economies which has the potential to generate large quantities of renewable power. Besides, waste recovery after the use of edible oils is almost non-existent in such economies. This makes the

problem more complicated. If properly treated, WCO has the potential to partially substitute some of the non-renewable fuels used for energy generation in the transportation, domestic, or industrial sectors. This present study is one such effort that investigates the possibility of using WCO in a diesel engine.

#### **2. Materials and Methods**

### *2.1. Materials*

Work reported in this study is based on WCO collected from four different restaurants. Any suspended particles in WCO were first removed using filter papers of size 12.5 mm diameter. The filtered WCO was heated at 100 ◦C for 1 h to remove the moisture content, followed by the cooling process. HSD for preparing biodiesel blends was purchased from local market. Methanol, ethanol, sulphuric acid, potassium hydroxide, and phenolphthalein were purchased from Sigma Aldrich (purity > 99%).

#### *2.2. Biodiesel Production Using WCO*

The physicochemical properties of WCO were illustrated in Table 1. Density and viscosity of WCO were observed higher than original canola oil (the source oil) due to the formation of unsaturated bonds through continuous use. This may also result in a high acid value of WCO. The acid value (AV) should be decreased before the conversion of WCO into biodiesel using mineral acids. Therefore, the raw WCO was treated with mineral acids (H2SO4, HCl, and H3PO4) through a process known as esterification for reducing its free fatty acid (FFA) contents, which in turn dictates the AV [39,40].


**Table 1.** Physicochemical properties of WCO.

The most important factor in the esterification process was the quantity of methanol. Increasing the methanol concentration in the mix would result in more effective FFA reduction. Other parameters were, reaction speed that was 600 rpm, the temperature was 60 ◦C and time for this reaction was 3 h. In this process amount of methanol used was 2.25 × FFA and the amount of sulphuric acid used was 0.05 × FFA. In this study, the maximum reduction (74.7%) in FFA was observed by treating WCO with H2SO4, which was followed by H3PO<sup>4</sup> (63.1%) and HCl (54.9%).To reduce the FFA of the WCO, the oil was treated into two steps using H2SO4. In the first step, AV was reduced from 3.9 to 1.45 mg KOH/g and in the second step it was reduced to 0.34 mg KOH/g. After this WCO was converted into biodiesel through transesterification process. The amount of catalyst used for transesterification was determined using Equation (1)

$$\text{Catalyst amount} = \frac{\text{Catalyst concentration} \times \text{Amount of WCO used}}{100} \tag{1}$$

In the presence of KOH catalyst and methanol, WCO has been converted into biodiesel by transesterification process [16,40]. In general, methanol is soluble in FAME or biodiesel but insoluble in triglycerides [41]. Methanol was added to WCO in the presence of KOH at the temperature range from (50 to 65 ◦C) with reaction times of 1–3 h which settled down overnight. Biodiesel appeared to be collected in the top layer whereas glycerine settled down in the bottom, the later was separated with a separating funnel. Transesterified biodiesel was washed with hot water continuously for removal of

impurities which included catalysts and unused methanol. Biodiesel was washed using distilled water, and this process was done until the used distilled water became transparent. After this, the rotary evaporator was used to remove the remains of water content and methanol form biodiesel. After rotary evaporation, biodiesel was filtered with Whatman filter paper to remove the traces amount of KOH catalyst. Biodiesel yield was calculated using Equation (2)

$$\text{Yield} = \frac{\text{Amount of biodies produced}}{\text{amount of WCO used}} \times 100\tag{2}$$

#### *2.3. Biodiesel Property Analysis*

The lower heating value (LHV) of biodiesel was determined using a bomb calorimeter. The flashpoint and fire point of biodiesel was measured using the Cleveland open cup apparatus (Koehler, New York, NY, USA). The FAME composition was determined using GCMS 5975C with triple I Detector. Helium gas was used as a carrier gas. For determination of the acid value of WCO, 0.5 N KOH was mixed with 50 mL distilled water and the mixture was used for titration. A mixture of 0.25 g phenathpelin and 25 mL ethyl alcohol was used as an indicator. A solution of 50 mL (95% ethyl alcohol and 5% distilled water) was prepared, and a 1 ml indicator was added into a solution of WCO. The AV of WCO was calculated using Equation (3) [40,42]

$$\text{Acid Value} = \frac{56.1 \times \text{N} \times \text{V}}{\text{W}} \tag{3}$$

where N is the normality of KOH, V is the volume of KOH and distilled water used for titration, and W is the weight of WCO used.

#### *2.4. Method for Biodiesel Yield Optimisation*

Five major operating parameters that affect biodiesel yield are temperature, catalyst concentration, and methanol to oil ratio, reaction speed and time. Design-Expert software 8.0.6 was used to design experimental conditions for the optimisation of biodiesel yield. The six-level, five factors CCD has been used in this study that requires 46 experiments. The ranges of operating parameters have been shown in Table 2.


**Table 2.** Ranges of operating parameters.

The data collected from performed experiments were analysed on Design-Expert 8.0.6 (Stat-Ease, Minnesota, MN, USA) and then interpreted. Three main analytical steps required to develop optimum conditions include regression analysis, plotting of response surface and analysis of variance (ANOVA). After optimising the yield, biodiesel was produced at a pilot scale using the optimised parameters for the engine performance test. Biodiesel blends were formulated using an electrical homogeniser. This homogeniser was rotated at 1500 rpm for 20 min to mix HSD with biodiesel in different concentrations. B10, B20, B30, B40, and B50 were prepared on volume bases where the concentration of biodiesel is specified with digits in blend name.

## *2.5. Engine Setup*

Blends of biodiesel obtained from WCO with diesel (B10, B20, B30, B40, and B50) was used to run the CI engine. The pictorial view, as well as the schematic diagram of the diesel engine test rig, is exhibited in Figure 1. The experimental setup consists of a six-cylinder, four-stroke, water-cooled, indirect injection diesel engine connected with a hydraulic dynamometer. The characteristics of the CI engine were shown in Table 3. Before collecting data, in every test, steady-state engine operation was ensured by running the engine for 20 min. Initially, the engine was operated with pure diesel to collect baseline/reference data and later on performance with biodiesel blends was tested. Engine performance (brake torque, brake power, BTE, and BSFC) at different engine speeds (1000–3500 rpm) at full load conditions were tested and reported in the following sections.

**Figure 1.** (**a**) Pictorial view and (**b**) schematic of the experimental testbed.


**Table 3.** Specification of CI engine.

#### **3. Results and Analysis**

#### *3.1. Physicochemical Properties of Biodiesel*

Table 4 presents the physicochemical properties of biodiesel sourced from the WCO. These properties have been compared with standard thermos physical properties of biodiesel as per ASTM standards. GCMS has determined the components of FAME, Table 5 illustrates the percentage composition of different long carbon chain elements.

**Table 4.** Physicochemical characteristics of biodiesel blends and HSD.


**Table 5.** FAME (*w*/*w* %) biodiesel produced from WCO.


#### *3.2. Model Fitting and ANOVA*

Table 6 represents ANOVA parameters for the quadratic polynomial model. These predicted values are obtained from model fitting techniques using Design–Expert software 8.0.6. There are various data fitting models that include linear, two factorial, cubic and quadratic. The quadratic polynomial model was used for data fitting, and the equation for this model is shown below.

$$\begin{array}{l} \text{Yield (\%)} = +84.33 - 0.16 \text{X}\_1 - 5.34 \text{X}\_2 - 7.59 \text{X}\_3 + 3.44 \text{X}\_4 + 0.97 \text{X}\_5 + 4.00 \text{X}\_1 \text{X}\_2 + 1.75 \text{X}\_1 \text{X}\_3 \\ + 5.50 \text{X}\_1 \text{X}\_4 + 2.13 \text{X}\_1 \text{X}\_5 + 6.38 \text{X}\_2 \text{X}\_3 - 0.38 \text{X}\_2 \text{X}\_4 - 0.13 \text{X}\_2 \text{X}\_5 - 0.38 \text{X}\_3 \text{X}\_4 + 1.62 \text{X}\_3 \text{X}\_5 + \\ 0.50 \text{X}\_4 \text{X}\_5 - 2.56 \text{X}\_1 \text{}^2 - 2.06 \text{X}\_2^2 - 6.81 \text{X}\_3^2 - 1.19 \text{X}\_4^2 - 2.90 \text{X}\_5^2 \end{array} \tag{4}$$

where X<sup>1</sup> is temperature, X<sup>2</sup> is catalyst concentration, X<sup>3</sup> is methanol to oil ratio, and X<sup>4</sup> is stirring speed, X<sup>5</sup> is reaction time. The model F-value of 2.70 means that the model is significant. A model F-value this large could only occur at 1.00% due to noise. "Prob > F" values less than 0.0500 show significant model terms. Values over 0.1000 are not significant in terms of the model. The experimental matrix developed by Design-Expert software which consists of CCD arrangements and responses has been shown in Table 7.


**Table 6.** ANOVA for the quadratic polynomial model.

<sup>a</sup> Significant at "prob > F" less than 0.05.; <sup>b</sup> Insignificant at "prob > F" more than 0.100.

**Table 7.** CCD arrangements and responses.



**Table 7.** *Cont.*

## *3.3. E*ff*ect of Operating Parameters*

This section explains the effects of temperature, catalyst concentration, methanol to oil ratio, reaction stirring speed and reaction time on the biodiesel yield. Figure 2a presents experimentally obtained RSM plot to investigate the effect of catalyst concentration, methanol to oil ratio, stirring speed, reaction time, at a constant temperature range. The amount of WCO sample was 20 g, and the operating temperature ranged between 50 ◦C to 65 ◦C. The catalyst concentration was varied between 0.25 *w*/*w* to 1.75 *w*/*w*. As shown in Figure 2a, biodiesel yield increased from 50 ◦C to 57.50 ◦C and decreased afterwards. The yield decreased by increasing catalyst concentration from 0.25 *w*/*w* to 1.75 *w*/*w*. Figure 2b illustrates a relation between temperature, methanol to oil ratio and percentage yield of biodiesel at a constant catalyst concentration of 1 *w*/*w*, constant stirring speed of 600 rpm and constant reaction time of 2 h. The biodiesel yield was increased from 5:1 to 8:1 and decreased afterwards. The maximum yield of 93% was obtained at 8.50:1 methanol to oil ratio. Figure 2c illustrates a relationship between temperature, stirring speed and yield. Biodiesel yield was optimised at stirring speeds between 560 to 640 rpm, and it decreased at higher stirring speeds. Figure 2d illustrates a relationship between temperature, reaction time and percentage yield of biodiesel at a constant catalyst concentration of 1 *w*/*w*, methanol to oil ratio of 8.50:1 and a constant stirring speed of 600 rpm. A slight increase in yield was observed by increasing reaction time from 1 h to 2 h. Biodiesel yield was enhanced when the temperature increased to 57.50 ◦C and then decreased by increasing the temperature further. The above-stated trends are consistent with the results already reported in the literature [43,44].

Figure 3a illustrates response surface plots as a function of catalyst concentration, methanol to oil ratio and percentage yield of biodiesel at a constant temperature of 57.50 ◦C, the rotational speed of 600 rpm and constant reaction time of 2 h. Minimum biodiesel yield was observed at 12:1 methanol to oil ratio with catalyst concentration of 1.75 *w*/*w*. Maximum biodiesel yield (93%) was observed at 8.50:1 methanol to oil ratio and 0.25 *w*/*w* catalyst concentration. Figure 3b shows that, by changing rotational speed from 400 rpm to 800 rpm, there is a sharp increase in percentage yield from 80% to 90% at a catalyst concentration of 0.25 *w*/*w*. Figure 3c illustrates that both reaction time and catalyst concentration affect the percentage yield. By increasing the reaction time up to a certain limit, and by decreasing catalyst concentration, the percentage yield of biodiesel increased.

**Figure 2.** Experimentally obtained RSM plot to investigate the effect of (**a**) catalyst concentration, (**b**) methanol to oil ratio (**c**) stirring speed (**d**) reaction time, at a constant range of temperature.

Figure 4a illustrates response surface plots as a function of methanol to oil ratio, stirring speed and percentage yield of biodiesel at a constant temperature, catalyst concentration and reaction time that are 57.50 ◦C, 1 *w*/*w* and 2 h respectively. Percentage yield sharply increased from 70% to 90% by increasing the stirring speed from 400 rpm to 800 rpm with methanol to oil ratio 8.50:1. As shown in Figure 4b, the optimum yield of biodiesel was observed at 7:1 to 8:1 methanol to oil ratio. Reaction time also affects the yield. Ghadge et al. [45] also reported the use of the same methanol to oil ratio in his investigation. Figure 5, illustrates response surface plots as a function of stirring speed, reaction time and percentage yield of biodiesel at a constant temperature, catalyst concentration, and methanol to oil ratio that are 57.50 ◦C, 1 *w*/*w* and 8.50:1 respectively. The graph shows a steady increase in yield by increasing rotational speed and reaction time.

**Figure 3.** Experimentally obtained RSM plot to investigate the effect of (**a**) methanol to oil ratio, (**b**) stirring speed and (**c**) reaction time, at a constant range of catalyst concentration.

**Figure 4.** Experimentally obtained RSM plot to investigate the effect of (**a**) stirring speed, (**b**) reaction time, at a constant range of methanol to oil ratio.

**Figure 5.** Experimentally obtained RSM plot to investigate the effect of reaction time at a constant range of stirring speed.

#### *3.4. Performance Characteristics*

#### 3.4.1. Brake Torque and Brake Power

Figure 6a,b shows the variation of brake torque and brake power with engine speed at full load condition for biodiesel blends and HSD. The trends for biodiesel blends were found to be similar to the HSD. Initially, the brake torque for HSD and biodiesel blends increases with the increase in engine speed, and it reaches a maximum value and then starts to decrease afterwards. This trend can be attributed to: firstly, the volumetric efficiency of the engine which decreases with increase in engine speed and secondly, the increase in frictional losses at higher engine speeds which results in the reduction of the brake torque. The maximum values of engine torque for both HSD and biodiesel blends were found at 3000 rpm. Biodiesel blends exhibited slightly lower brake torque than that of HSD. This can be attributed to the relatively low LHV and higher kinematic viscosity of biodiesel blends. Higher kinematic viscosity results in a greater delay in the start of injection, which also leads to poorer fuel atomisation [46]. The average torque reduction for biodiesel blends compared to HSD at 3000 rpm was found to be 1.45% for B10, 2% for B20, 2.2% for B30, 3.09% for B40, and 3.5% for B50. Brake power is derived from brake torque by multiplying with the angular speed. From Figure 6b, the highest power was observed at 3500 rpm. Again, the highest power was shown by HSD followed by B10, B20, B30, B40 and B50, respectively. Rizwanul Fattah et al. [46] studied the engine performance, and emission characteristics of Malaysian Alexandrian Laurel oil-based biodiesel (ALB) blend with diesel. They also reported a slight reduction in maximum brake powers for ALB10 and ALB20 compared to that of diesel fuel. The power outputs were 48.5 kW, 48.3 kW, and 48.2 kW, for diesel, ALB10, and ALB20, respectively. They attributed this reduction to the lower LHV and higher viscosity of blends, both of which result in poor fuel atomisation compared to that of diesel fuel.

**Figure 6.** Effect of different fuels on (**a**) BT and (**b**) BP of the engine at different speeds.

3.4.2. Brake Specific Fuel Consumption and Brake Thermal Efficiency

Figure 7a presents the variation in BSFC for biodiesel blends and HSD as a function of engine speed. Biodiesel blends have lower LHV and higher densities compared to that of HSD, resulting in higher fuel consumption for biodiesel blends. It is evident for the figure that, BSFC first decreased to the lowest at 2000 rpm, and then increased with increase in speed. The average increase in BSFC for biodiesel blends as compared to HSD at 3500 rpm is found as 1.61% for B10, 5.73% for B20, 8.8% for B30, 12.76% for B40, and 18% for B50. The increase in BSFC for biodiesel blends was due to volumetric effect of the constant fuel injection rate together with the higher kinematic viscosity and lower LHV of biodiesel and its blends, and this became more pronounced in higher biodiesel blends such as B50. Palash et al. [47] also reported a similar finding for *Aphanamixis polystachya* biodiesel (APME) blends with the average BSFC values for diesel, APME5, and APME10 of 352.96 g/kWh, 356.05 g/kWh, and 359.29 g/kWh, respectively. Figure 7b represents variation in BTE as a function of engine speed at full load conditions. Thermal efficiency varies inversely with the product of BSFC and LHV of the fuel [46]. The highest BTE for all fuels were observed at 3000 rpm at full load condition. The BTE values for HSD, B10, B20, B30, B40 and B50 are 34.97%, 32.78%, 33.36%, 33.76%, 34.15%, and 32.11%, respectively.

**Figure 7.** Effect of different fuels on (**a**) BSFC and (**b**) BTE of the engine at different speeds.

## **4. Conclusions**

Biodiesel yield was optimised at optimum operating conditions. FFAs of WCO were reduced by acid treatment. Among mineral acids, H2SO<sup>4</sup> was found to be more effective. It reduces the FFAs value of 91.36%. Transesterification of WCO with methanol was found to be very effective. A biodiesel yield of 93% was obtained with a 57.50 ◦C reaction temperature, 0.25% catalyst concentration, 8.50:1 methanol to oil ratio, 600 rpm stirring speed, and 3 h reaction time. Various physiochemical properties justify the quality of biodiesel. Oleic acid (C18:1), linoleic acid (C18:2), α-linoleic acid (C18:3), and palmitic acid (C16:0) were the major constituents of biodiesel. Performance characteristics of a diesel engine using diesel fuel and biodiesel blends B10, B20, B30, B40, and B50 were tested and compared. Engine performance in terms of brake torque, brake power, BSFC, and BTE was determined. The average torque reduction for biodiesel blends compared to HSD at 3000 rpm were found to be 1.45%, 2%, 2.2%, 3.09%, and 3.5% for B10, B20, B30, B40, and B50, respectively. The average increase in BSFC for biodiesel blends compared to HSD at 3500 rpm were found to be 1.61%, 5.73%, 8.8%, 12.76%, and 18% for B10, B20, B30, B40, and B50, respectively. Pakistan is spending a huge amount of money in importing HSD to fulfil their energy requirements. The government is looking for alternative fuels that have the potential to reduce the consumption of HSD. At present, the most suitable alternative fuel is biodiesel produced from WCO. The conversion of WCO into biodiesel not only reduces the import of HSD, but also reduces environmental concerns.

**Author Contributions:** Conceptualisation, L.R., Z.A., and S.I.; methodology, M.E.M.S.; software, M.E.M.S.; validation, M.M.A., H.M.K., and T.A.; formal analysis, M.M.A.; investigation N.S.; resources, M.M.A. and M.E.M.S.; data curation, M.E.M.S.; writing—original draft preparation, M.F.; writing—review and editing, S.M.A.R., I.M.R.F., and M.M.A.; visualisation, M.A.; supervision, L.R.; project administration, N.S.; funding acquisition, M.M.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors would like to thank to the Pakistan Council of Scientific and Industrial Research (PCSIR) for their services regarding biodiesel characterisation. Authors would also like to acknowledge 'research development fund' of School of Information, Systems and Modelling, Faculty of Engineering and Information Technology, University of Technology Sydney, Australia.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **Nomenclature**


## **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## **Coal Combustion Products Management toward a Circular Economy—A Case Study of the Coal Power Plant Sector in Poland**

## **Agnieszka Bielecka <sup>1</sup> and Joanna Kulczycka 2,3,\***


Received: 15 June 2020; Accepted: 9 July 2020; Published: 13 July 2020

**Abstract:** Coal combustion products can be considered as commercial products or waste depending on the quality of the coal, the combustion process, and the country's legislation. The circular economy can create incentives for the implementation of new business models in large power plants in cooperation with coal mines and users of coal combustion products. This is particularly important in Poland, where coal still remains the main source of energy, employing over 80,000 workers. The objective of this study was to assess the readiness for change toward a circular economy and to identify challenges, barriers, and plans at seven large power plants. To do this, a final questionnaire was developed after checking environmental reporting, a CATI survey, and brainstorming between circular economy leaders from science, industry, and non-governmental organizations. The results indicate that even if the great economic and environmental potential of coal combustion products management are understood, all requirements connected with CO<sup>2</sup> and air pollution have higher priorities. Policy shifts away from coal do not promote cooperation, but the higher acceptance of products from waste and more transparent data shows a large potential for changes toward a circular economy.

**Keywords:** circular economy; coal power plant; coal combustion products; industrial waste

## **1. Introduction**

The production volumes of coal combustion products (CCPs) are directly correlated with coal combustion in thermal power stations. Their quality, however, depends on the mineral composition of the coal, combustion efficiency, type and fineness of the coal, time of storage of the minerals in the furnace/boiler, etc. [1]. They constitute major inputs in the manufacture of construction materials and geotechnical engineering applications. Some of them i.e., pulverized fly ash, are a valuable waste residue used in concrete production [2]. However, they can also pose serious threats to air, water, and soil; over the past two decades, research has been developing methods to combine both the quality of power generation, economic viability, environmental safety, quality, and the effective use of waste products [3].

The beneficial use of CCPs varies in different countries; for example, Australia reached 5.936 million tonnes or 47% of CCPs in 2018 [4]. To facilitate precision and consistency, the members of the World Wide Coal Combustion Products Network have worked together to harmonize terminology and promote CCPs as a coherent nomenclature. This terminology is more positive and is in line with the concept of industrial ecology and circular economy (CE) principles, which is an approach that aims to reuse one industry's byproducts as another industry's raw material [5].

It is estimated that approximately 750 million tonnes of coal ash are generated every year in the coal industry throughout the world. A significant volume of coal fly ash is produced by major industrial countries, where it is one of the largest single sources of industrial solid waste [6].

Utilization rates vary considerably between countries, with the highest effective utilization rate of 99.3% being reported in Japan, while Africa/Middle East (still) has the lowest rate of 10.6%. In Poland, approximately 20 million tonnes of CCPs were generated (Table 1), of which approximately 65% was utilized in 2016 and 35% was landfilled. CCPs can substitute natural aggregate, as they are mainly used for in cement industry, in construction of new permanent disposal facilities, mining applications, in structural fills, waste stabilization, gypsum panel production, and agricultural applications. These applications can reduce greenhouse gas emissions, thus increasing energy efficiency and minimize use of primary resources, which is in line with CE strategy.

Based on the example of CCPs, Park and Chertow [7] have proposed the resource potential indicator (RPI) to promote the importance of managing wastes as resources in a closed loop. This measures the intrinsic value that a material has for reuse taking into account the technological recycling availability, economic feasibility and environmental requirements for the US market.


**Table 1.** Amount of coal combustion products (CCPs) generated in Poland in 2016 in enterprises and energy groups [8].

The strategies of enterprises from the coal power sector in Poland contain provisions that prove that these enterprises carry out activities fulfilling CE objectives, but they do not directly refer to the CE aspects, as most data are available in corporate social responsibility (CSR) reports.

However, the current way of presenting this information does not allow for the comparison of data. In the context of CCPs, there are also no provisions proving the tendency (required by the CE) to carry out activities aimed at eliminating the generation of waste, hence only part of it is managed. Moreover, there is neither sufficient transparency in the statistics nor clear rules concerning waste, byproducts, secondary material, etc. [9]. Therefore, the management of CCPs is not the most important goal for either Polish companies or Polish energy policy. They first have to meet air quality standards and try to reduce GHG emissions. As Hopkins and Purnell (2019) indicate, the politics of GHG allocation may hide a lack of progress towards a CE [10].

In Poland, according to government documents, i.e., Polish Energy Policy to 2040 (2018) [11] and the Polish National Energy and Climate Plan for the years 2021–2030 published in 2019 [12], it was stated that the share of coal in electricity generation will be systematically reduced. In 2030, it will reach the level of 56–60%, and in 2040, the downward trend will be maintained. However, it was underlined that there is great potential for the better utilization of CCPs as building materials and for a reduction in the use of primary raw materials, which is in line with a CE. It is also estimated that the renewable energy sources share in electricity production will increase to approximately 32% in 2030. These changes involve the most significant challenges for the current Polish energy sector relating to more stringent standards under the Industrial Emissions Directive (IED) [13] and the best available techniques (BAT) for the reduction of GHG and SO2, NOx, and dust emissions as well as the monitoring of hydrogen chloride, hydrogen fluoride, mercury, and ammonia. Enterprises have been allowed until 16 August 2021 to adapt to the new regulations, but the adaptation of installations to meet the stricter requirements involves the need for investment. Such solutions can have positive or negative impact on the quality of CCPs. Therefore, determining the impact of techniques focusing on reducing dust and gas emissions on the quality of CCPs, as well as possibilities of their further physicochemical processing according to end-user needs, is of extreme importance [14]. One of the good examples of this is the production of synthetic gypsum from flue gas desulphurization installations. Good quality of physical and chemical parameters of gypsum produced in the Bełchatów power plant was confirmed by laboratory tests. Therefore, about 60% of gypsum in Poland is used in different sectors, mainly for the construction market in the creation of various types of prefabricated products based on synthetic gypsum [15].

Several reviews focusing on technological issues relating to CCP utilization and application have already been published [16,17]. These analyze the economic and environmental constraints and benefits [18,19]. There is, however, a lack of publications focusing on surveys of energy companies on incentives and barriers that could have an impact on decisions on managing waste as a resource according to the CE concept. Moreover, there is lack of standards on how to measure micro-level circularity [20]. Therefore, the aim of this paper is to verify whether coal power plants were ready to implement CE models for CPP management. The research carried out in 2018 aimed to


An analysis of good practice and research conducted (patents) relating to CCP management in Poland shows the existence of a large potential on the supply side, but long-term actions stimulating and legal solutions facilitating cooperation within the value chain are still lacking.

#### **2. Materials and Methods**

In order to investigate the role of CE in LPP (large power plant), we follow Gonenc and Scholten [22], who suggested that future studies should take into account corporate intentions and actions taken which are in line with such intentions. We focus on CE, but in 2018, there were no LPPs that had a CE strategy or any indicators focusing on CE. Therefore, we first chose the most important LPPs and then searched the official corporate web sites of LPPs concerning environmental aims and data presented in CSR reports. This allowed us to identify the activities focusing on closing material loops and waste minimization that are in line with environmental strategies. Second, we selected key activities that can reflect on LPP actions directed toward implementing a CE and identified the managers responsible for its application. We contacted them by email or phone to ask about the role of the CE and the scope and responsibilities in the companies. Third, we chose 35 managers from different departments at seven LPPs who were potential leaders of the CE. Fourth, based on a literature review of surveys of the CE and after brainstorming with CE leaders from science and industry, a survey questionnaire was constructed and sent to industry representatives (Figure 1).

The research was carried out on a group of the seven largest capital groups in the Polish energy sector (with coal-fired plants), for each of which the share in the volume of electricity introduced

into the grid amounted to more than 2% (73% in total)—PGE Polska Grupa Energetyczna S.A. (36%), Tauron Polska Energia S.A. (10%), Enea S.A. (9%), PAK S.A. (6%), EDF (8%), CEZ (2%), and PGNiG (2%) [5]. At the same time, these are the largest manufacturers of CPPs. The research was conducted with the participation of 27 managers who, due to the scope of their duties, had the knowledge necessary to carry out the research.

The target group was selected from the above-mentioned companies after initially verifying the competences of the employees in departments such as Environmental Protection, Development and Quality Assurance, Development (Investments and Projects), R&D, Raw Materials and Waste, and Technological Processes. Telephone interviews (CATI) were then carried out over a period of five weeks (May–June 2018) using a questionnaire based on a five-step Likert scale (where 5 was equivalent to "strongly agree" and 1 was assigned to "strongly disagree"). The results of the research were aggregated before being presented, and the reliability of the survey was tested using the Cronbach alpha coefficient.

Aspects discussed in the survey were concerned with six topics as follows:


The detailed survey with questions is presented in Table 2. The questions (i.e., p 1\_1) were chosen based on literature, reports, and experts experience and advice.


**Table 2.** Summary presentation of the aspects assessed in the questionnaire.


**Table 2.** *Cont.*

1 through activities that will be carried out aimed at eliminating the storage of CCPs as a result of solutions enabling their further use in the economy, so that the resources used will not lose their value and their useful life will be extended. <sup>2</sup> according to the "cradle to cradle" approach.

#### *Verification of the Research Tool*

The test tool used in the research has been checked for reliability. For this purpose, an analysis of the results obtained was carried out using the Cronbach α coefficient (Table 3). This is a coefficient for the total scale, whose items can be measured on any order scale. The Cronbach α coefficient indicates what part of the variance of the total scale is the variance of the true value of that scale. A zero value means that individual scale items do not measure the true result but only generate a random error. As a result, there is no correlation between items of the total scale. In tests of the reliability of the measuring tools, the result should be within the range of 0.6–0.8.

A set of terms (in order from the highest to the lowest value of Cronbach's α coefficient) for aspects 2, 3, and 1 obtained acceptable reliability. Aspect 6, which concerned predicting the possibility of implementing CE in a 10- or 20-year perspective, Cronbach's α was 0.462, which can be explained by the fact that the respondents answered "strongly agree" for the 20-year period, whereas for the 10-year period, the vast majority only agreed. The set of aspects 4 and 5 did not receive acceptable reliability. In such a situation, factor analysis was used to indicate the number of factors (sub-aspects) that formed these aspects. Results of the analysis (selection of the optimal number of factors) were based on the screen test and the questions assigned to them were based on the varimax rotational strategy. Four factors were obtained for the set of aspect 4 and two factors for the set of aspect 5. The values of Cronbach's α coefficients for the subcategories obtained as a result of factor analysis are presented below.


**Table 3.** Internal consistency of the survey questionnaire.

As many as 11 terms of aspect 4 concerned the CE requirements for company management. Their diversity was so large that factor analysis, which was used to separate sub–aspects (factors), indicated three strong factors and one weaker one (the last one) (Table 4).

**Table 4.** Cronbach's α coefficient results for factor analysis of aspect 4.


The five terms of aspect 5 concerned the reasons for implementing a CE. The factor analysis, which was used to separate sub-aspects (factors), indicated that it was possible to divide them into two factors that could be associated with the external and internal reasons for implementing CE (Table 5).

**Table 5.** Results of the Cronbach α coefficient for factor analysis of aspect 5.


#### **3. Results**

The research undertaken has enabled


#### *3.1. Aspect 1*

The set of terms under aspect 1 concerned the evaluation of the benefits/effects related to implementing CE in a coal power company (Figure 2).

**Figure 2.** Assessment of the benefits (effects) resulting from implementing CE in coal power engineering.

As regards the assessment of the benefits of implementing CE in the coal power industry, respondents assessed such effects as the increase in efficiency of the use of resources (p1\_1), extended product life cycle (p1\_2) and establishing relations generating less environmental pressure with consumers (p1\_6) in a similar way. In relation to these terms, more than 59% of the respondents answered in the affirmative, no more than 37% of the respondents expressed doubts, while the number of negative opinions did not exceed 4% of the answers obtained. More than 80% of the respondents assessed terms such as introducing more sustainable production (p1\_3), reduction of the environmental footprint (p1\_4), strengthening standards of environmental management (p1\_9), an increase in ecological activity (p1\_10) and closing CCP landfills (p1\_11). The above-mentioned terms did not obtain negative assessments. The unanimous affirmation (positive assessment) of all respondents was shown in relation to the term generation of products instead of waste (p1\_7) and improvement in environmental quality (p1\_8). The greatest reflection among the respondents was given to the assessment of the benefits associated with the possibility of establishing less energy-intensive relations with suppliers (p1\_5) expressed by 55% of the respondents. The remaining 40% of the respondents gave this term a positive assessment and only 4% a negative one. The statement that received the highest number of negative votes (63%) and no positive assessments was the one concerning benefit resulting from the possibility of reducing the environmental costs of production (p1\_12).

#### *3.2. Aspect 2*

The set of terms under aspect 2 concerned the assessment of the risks/challenges related to implementing CE in the coal power industry (Figure 3).

As regards the assessment of the risks/challenges associated with implementing CE in the coal power industry, more than 50% of the respondents negatively assessed the risk of reducing a company's competitiveness (p2\_4). Difficulty in making the assessment relating to the increase in production costs (p2\_1) was expressed by 48% of the respondents, and the need to implement costly and risky investments (p2\_2) was expressed by 63% of the respondents.

**Figure 3.** Assessment of the risks/challenges associated with implementing CE in the coal power industry.

The risk that may be posed by implementing CE in the form of an increase in production costs (p2\_1) received a greater number of affirmative (48%) than negative (4%) opinions. However, as far as the necessity to carry out investments (p2\_2) is concerned, the opposite situation occurred as it received more negative (22%) than positive (15%) opinions. More than half (52%) of the respondents see a challenge in the need to find a way to change the current business model of the process (p2\_3), but a large proportion of them (44%) had doubts in assessing this term. The vast majority (78%) of respondents see a risk/challenge in the necessity to go beyond the "plant gates" (p2\_5). In total, 18% of the respondents expressed doubts when making the assessment in relation to this term.

#### *3.3. Aspect 3*

The set of terms under aspect 3 concerned the assessment of the barriers/obstacles to the implementation of CE in the coal power industry (Figure 4).

In terms of barriers/obstacles that respondents see in the context of implementing CE, over 40% do not agree that raw materials recovered from waste do not find buyers (p3\_3). Moreover, in relation to this term, the highest number of opinions (56%) was given to the assessment "neither agree nor disagree." About 50% of the respondents see problems with the lack of implementation of a multi–product strategy of power engineering (p3\_1) and the lack of promotion of CCPs as fully fledged products from waste (p3\_2). In relation to these two statements, an equally large number of respondents (over 44%) had difficulties in their assessment. More than 70% of respondents admitted that an obstacle to the implementation of CE is the lack of indicators and guidelines for implementing and measuring CE (p3\_4). Nearly 90% of the respondents indicated that an obstacle to implementing CE is the fact that the current CE-compliant activities are not appreciated and evaluated (p3\_5); moreover, this term was the only one that did not receive a negative assessment.

**Figure 4.** Assessment of barriers/ obstacles to the implementation of CE in the coal power industry.

#### *3.4. Aspect 4*

The set of terms under aspect 4 concerned the assessment of CE requirements in relation to the management of a coal power plant (Figure 5).

**Figure 5.** Assessment of the CE requirements in relation to the management of a coal power plant.

The analysis of the assessment of terms related to management activities indicated that more than 50% of the respondents had difficulty in assessing the need to go beyond the boundaries of the plant with responsibility for environmental protection (p4\_1) (52%), constructing an income generation logic based on the identification of the value chain (p4\_3) (70%), internalization of external environmental costs (p4\_4) (89%), creating a buyers–eco-consumers market (p4\_6) (59%), and identification of the value chain to specify quantifiable environmental benefits (p4\_10) (67%). The number of other responses expressed for these terms was always higher on the positive (affirmative) side.

None of the six terms described below received a negative assessment. These terms include cooperation within clusters (p4\_2) (85%), assessment of environmental policy in relation to suppliers and customers (p4\_5) (74%), implementation of a multi-product power engineering strategy (p4\_7) (70%), recognition of factors representing environmental costs in the expenditure structure (p4\_8) (93%), influencing the awareness and actions of other bodies (p4\_9) (59%), and integration of environmental

policy with a development strategy (p4\_11) (85%). The term about which most respondents (41%) were doubtful when expressing their opinion was influencing the awareness and actions of other bodies (p4\_9).

### *3.5. Aspect 5*

The set of terms under aspect 5 concerned the assessment of the reasons for the coal power industry to implement CE (Figure 6).

**Figure 6.** Assessment of reasons for the coal power industry to implement CE.

Respondents had an unequivocally positive assessment of four out of five reasons for implementing CE by coal power companies. These terms included awareness of the importance of environmental protection (p5\_1), social benefits (p5\_3), the possibility of reducing the product's environmental footprint (p5\_4), and economic benefits (p5\_5) received 100% of affirmative responses. Only in relation to the term expressing the possibility of reducing costs related to waste management (p5\_2) were 4% of respondents doubtful in its assessment.

## *3.6. Aspect 6*

The set of terms under aspect 6 concerned the assessment of the prospect for implementing a CE in the coal power industry (Figure 7).

**Figure 7.** Assessment of the prospect of implementing a CE in the coal power industry.

Concerning the term relating to the possibility of implementing CE in the power industry within 10 or 20 years, the respondents agreed that, in their opinion, there is such a possibility.

#### **4. Discussion**

The analysis of the assessments obtained from the respondents for aspect 1 shows that the vast majority of the respondents see the potential of environmental effects that can be achieved as a result of CE. This means that employees are aware of the potential that the implementation of this new trend may bring to the company. The reflection of the respondents shown in relation to the supply chain (benefits from establishing new relations with suppliers) may be caused by the fact that the power industry currently takes into account the quality aspect of the emerging CCPs in specifications for the purchase of raw materials (concerning the requirements for the calorific value of coal, content of ash and sulphur, etc.). Despite the positive assessments expressed in relation to most of the environmental effects generated by the implementation of CE, they do not refer to the initial stage, which is (immediately after the product design process) the stage of material procurement. The supply chain generated is of key importance and is translated into other processes and the environmental costs they create. The supply chain, through transferring the effects of the environmental impacts generated, influences the possibilities of CCP management, in relation to which the culmination of all effects generated in the preceding processes takes place. Therefore, in the opinion of the authors of this study, it can be assumed that in order to implement the CE model it is necessary to identify all links in the value chain and express them in the form of measurable indicators, which will allow the effects produced to be measured. On the other hand, the respondents fail to see that the benefits resulting from the reduction of environmental costs of production by eliminating the generation of waste may be related both to the underestimation of waste management costs (the costs of transport, storage and disposal of waste) and to the lack of internalization of external environmental costs. As a result, this may lead the respondents to believe that current relations do not require additional modifications unless they are additionally imposed by legal requirements.

The analysis of the assessments made for aspect 2 shows that they do not perceive CE as a threat related to disturbance of the position of enterprises on the market; this is due to the fact that they focus primarily on the process of energy generation. On the other hand, CCPs are perceived as a possible secondary product that may bring potential profits–but this is not the direction of the strategies currently being adopted. In the opinion of the respondents, the most important thing for enterprises is that the carrying out of a CE brings no increase in production costs. This aspect can be directly combined with the above described negative assessment expressed in relation to the benefits of implementing CE in the form of reducing environmental production costs through eliminating waste generation. This translates into further consequences, which are located in the concerns about the need to find a way to change the business model and the need to go beyond the plant gates related, for example, to the carrying out of a life cycle assessment. This may result from the lack of a consistent approach to environmental protection throughout the entire value chain generated by the company, which means that it is limited to analyses within the plant gates.

The answers obtained for aspect 3 prove that the respondents see the potential and the need to implement a multi–product power engineering strategy in the context of the possibilities that CCP management brings. However, at the same time, the potential for their sale is not utilized. The respondents also show that in their opinion activities carried out so far are CE compatible. However, the lack of guidelines for CE measurement is an obstacle in verifying this opinion.

As far as aspect 4 is concerned, the answers received confirm that requirements set by the CE, connected with the necessity to go beyond the plant gates and to analyze the impact of the activity throughout the whole life cycle determining the areas where the activities enabling the implementation of CE should be commenced. All activities where it is necessary to go beyond the plant gates and their measurement are a big barrier (e.g., introducing economic symbiosis), as are connecting with data availability and modelling.

The answers obtained for aspect 5 indicate that four out of the five reasons for implementing CE are important for coal power companies. It should be emphasized that generating economic profits was the only positively evaluated aspect for which the majority of respondents indicated the answer "I strongly agree." The respondents' doubts as to the possibility of reducing the costs of waste management may be explained by the fact that, at present, they are unable to direct their thinking towards the goal of eliminating the generation of waste in order to generate products/goods. This may be caused by focusing only on generating profits for the company from the leading activity, which is energy production (and focusing on current profits).

The answers obtained for aspect 6 indicate that the respondents envisage implementing CE in the next decade. This may be due to the decisions of the company that, through its declared environmental policy and the resulting pro-environmental attitude, will take actions related to implementing CE spontaneously. There is also an option that CE implementation will result from the need to adapt to anticipated changes in legislation.

In contemporary Polish strategic document and policy, i.e., National Energy Policy until 2040, most changes focus on the reduction emissions of CO<sup>2</sup> and other air pollutant from burning coal and less on circularity and waste management. This is a global important priority, but despite the efforts of many countries in their policies and strategies to reduce energy production from fossil fuels, a significant share of the world's energy supply, as well as the CO<sup>2</sup> emissions, is still generated from coal.

Global CO<sup>2</sup> emissions from fossil fuels combustion and processes further increased by 1.9% in 2018 compared to the previous year reaching a total of 37.9 Gt CO2. More detailed data about anthropogenic greenhouse gas emissions can be also found in Emissions Database for Global Atmospheric Research (EDGAR). Since 2015, the EU–28 share of global fossil CO<sup>2</sup> emissions has remained constant at 9.6% which is the equivalent of seven tonnes of CO2/cap/yr. The largest contributor in 2017 was Germany with a 22.4% share, followed by the United Kingdom (10.7%), Italy (10.2%), France (9.5%), and Poland (9%) [23]. Poland, however, noted a significant decrease (29%) in CO<sup>2</sup> emissions from the power industry in the years 1990–2017. In Poland, the emissions reached 8.3 tonnes of CO2/cap/yr in 2017, whereas in Germany, they reached 9.7 tonnes of CO2/cap/yr [24].

Wider and full utilization of CCPs in Poland can also have a beneficial impact on the reduction of energy use and CO<sup>2</sup> emissions in the industry. Beneficial use of CCPs as raw materials in manufacturing construction materials is well established both in literature [25] and the practice of many EU-15 countries as well as Japan, Korea, and China, where the utilization rate has reached over 70%. CPPs are mostly used for cement and concrete, road construction as well as for geotechnical and filling applications, including reclamation, as CCPs have established product standards or technical guidelines. There are, however, proposed solutions, such as


6. Analysis of the available technologies, the opportunities for their improvement, and the sources of financing in connection with the implementation of CE principles, as well as the results of research on assessing the implementation of CE in large energy concerns indicate that, along with the development of CE policy in Poland, there is a large potential for the full utilization of CCPs.

### **5. Conclusions**

From the assessment of the respondents and legal changes in EU and new energy strategy Poland, it can be concluded that the implementation of CE requires unanimity in actions and ensuring their measurability, which can be achieved by the introduction of the circular business model and its indicators.

Based on the terms of all analyzed aspects, it can be concluded that


In relation to the results obtained, it can be concluded that the respondents see potential in the modification of the current method of company management in the context of the CCPs generated, but they lack precise guidelines on how to do it (areas of action and indicators), which determines the direction and need for further research in this area. Successful implementation of the circular economy rules in Poland creates conditions for the promotion of zero waste technologies. Polish companies could analyze in more detail and assess good practices from different countries and focus on closer cooperation with scientific institution. It additionally allows them to apply for financial support for such eco-friendly investments.

**Author Contributions:** Conceptualization, J.K. and A.B.; methodology J.K.; validation, A.B. and J.K.; formal analysis, A.B.; investigation, A.B. and J.K.; resources A.B.; data curation A.B. and J.K.; writing—original draft preparation, A.B.; writing—review and editing, J.K.; visualization, A.B. and J.K.; supervision, J.K.; funding acquisition, J.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** The publishing of this work was supported by National Center for Research and Development as a part of the oto–GOZ project and AGH University of Science and Technology in Cracow, Poland.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Numerical Modelling and Experimental Verification of the Low-Emission Biomass Combustion Process in a Domestic Boiler with Flue Gas Flow around the Combustion Chamber**

#### **Przemysław Motyl 1,\* , Danuta Król 2 , Sławomir Poskrobko <sup>3</sup> and Marek Juszczak <sup>4</sup>**


Received: 8 October 2020; Accepted: 31 October 2020; Published: 9 November 2020

**Abstract:** The paper presents the results of numerical and experimental studies aimed at developing a new design of a 10 kW low-emission heating boiler fired with wood pellets. The boiler is to meet stringent requirements in terms of efficiency (η > 90%) and emissions per 10% O2: CO < 500 mg/Nm<sup>3</sup> , NOx <sup>≤</sup> 200 mg/Nm<sup>3</sup> , and dust <sup>≤</sup> 20 mg/Nm<sup>3</sup> ; these emission restrictions are as prescribed in the applicable ECODESIGN Directive in the European Union countries. An innovative aspect of the boiler structure (not yet present in domestic boilers) is the circular flow of exhaust gases around the centrally placed combustion chamber. The use of such a solution ensures high-efficiency, low-emission combustion and meeting the requirements of ECODESIGN. The results of the numerical calculations were verified and confirmed experimentally, obtaining average emission values of the limited gases CO = 91 mg/Nm<sup>3</sup> , and NOx = 197 mg/Nm<sup>3</sup> . The temperature measured in the furnace is 450–500 ◦C and in the flue it was 157–197 ◦C. The determined boiler efficiency was 92%. Numerical calculations were made with the use of an advanced CFD (Computational Fluid Dynamics) workshop in the form of the Ansys programming and a computing environment with the dominant participation of the Fluent module. It was shown that the results obtained in both experiments are sufficiently convergent.

**Keywords:** biomass combustion; pellet boiler; CFD modeling; renewable heating

## **1. Introduction**

Excessive CO<sup>2</sup> emissions to the atmosphere, the threat of smog (especially in the vicinity of large urban agglomerations), and appropriate legal regulations force the low energy industry to eliminate the use of fossil solid fuels. In the European Union countries, hard coal has almost been eliminated from the market as a fuel intended for small prosumer energy. Solid renewable fuels, wood biomass, and agro biomass remained on the market. Combustion of these fuels in low-power (domestic) boilers can undoubtedly contribute to reducing CO<sup>2</sup> emissions to the atmosphere. However, this will not make onerous and dangerous smog disappear [1–3]. To meet the care of the natural environment, the European Union Parliament imposed the obligation to conduct the biomass combustion process for low-power boilers intended for heating small objects (including residential buildings), so that

the emission of combustion products to the atmosphere meets the requirements of the ECODESIGN standard in the scope of CO < 500 mg/Nm<sup>3</sup> , NOx < 200 mg/Nm<sup>3</sup> , and dust < 40 mg/Nm<sup>3</sup> for O<sup>2</sup> = 10 mg/Nm<sup>3</sup> in the flue gas [4–6]. This standard is effective in European Union countries from 2020. The requirements of the standard (when it comes to low-power boilers fired with biomass) mainly concern the limited values of CO and NO<sup>x</sup> emissions and the conduct of the combustion process with high energy efficiency—the process high-efficiency combustion is associated with the low-emission process. Meeting the abovementioned requirements obliges boiler manufacturers (especially in countries where solid fuel boilers are widely used) to take actions aimed at continuous improvement of their products, designing a range of products adapted to legal requirements. Considering the needs of atmosphere protection and the related market requirements, in many research units actions are taken to identify biomass fuels burned in low-power boilers, model their combustion process, and increase fuel-burning efficiency to meet the standards in force in the European countries (EC). Given the stringent emission standards, the considerable potential of biomass in EC countries encourages the development of biomass combustion technology in low-power boilers [7–10]. Therefore, it is necessary to support the activities discussed in [11], where the authors justify the social and economic need for effective use of this fuel in the small domestic energy industry. Such actions are justified by the sustainable development policy implemented in the European Union countries. In [12], the efficiency of burning wood pellets in domestic boilers, with different loads and different configurations of burner settings was discussed and compared. The authors stated that high emission and energy efficiency depend (when it comes to structural changes) on the configuration of the burner—higher efficiency is obtained at the lower power supply. In the paper [13], the authors analysed the work of several hundred boilers fired with biomass pellets. These boilers were installed in residential buildings in Germany, Austria, Great Britain, and Spain. Despite the interesting operational data, it is worth noting that the monitored boilers were characterized by a typical design of the boiler furnace and convective part. The burner was located centrally at the bottom of the spacious furnace. The convective part was located behind the hearth, in a system of single-draft, self-cleaning ducts with a circular cross-section (flame tubes). The exhaust gas flowed through these channels to the flue. The furnace and convection part were separated from the surroundings with a water jacket and a layer of insulation. Such boilers (as indicated in the data contained in [13]) are widely used in EC countries. Other—relatively popular—boiler structures are presented in [14]. The research was aimed at obtaining useful data (for two different commercial low-power boilers), which concerned the intensity of sediment formation on the walls of heat exchangers. The first boiler is a two-draft boiler with a water jacket in the hearth and in the convection part, which is the second flue gas draft. In the upper part of the furnace and convective part—in order to increase the heat exchange surface—there were tufts of water tubes flushed by exhaust gases. The exhaust fumes from the convection part located behind the combustion chamber were directed to the smoke duct. The second boiler is a common one-draft structure with a flue gas/water heat exchanger located directly above the tube burner. The authors noted that the ability to settle solid particles was lower in the case of a boiler with a vertical single-tube heat exchanger. The reason for the appearance of smaller deposits in a single-pipe boiler is the stronger turbulence of flue gas flow through the pipes in which screw turbulators were additionally installed. Similar studies are presented in [15]. The researchers' interests were tubular heat exchanger boilers. They examined the impact of installing additional elements of the tubes in the heat exchanger. Based on the results, they found that an increase in the number of tubes (i.e., an increase in the heat exchange surface) has an impact on the increase in boiler efficiency, with an increase in the CO content in the flue gas. There were no changes in solid particle emissions. In the paper [16], the possibilities of modernizing the heat-flow system of a two-draft boiler with a water jacket in the furnace and in the convection part, constituting the second flue gas, were pointed out. The added fan increased the turbulence of the exhaust gas flow in the upper part of the furnace and convection chamber.

Observations of the test results based on the above review of biomass combustion technology in small boilers show that the boiler designs are based mainly on the concept of hot flue gas flow

in multi-pipe (usually two or three-pipe) channels with a water jacket. The channels are arranged according to the following scheme: the hearth, behind the hearth, successive convection channels (one behind the other) or the hearth, and other convection channels above the hearth (one above the other). These can be rectangular ducts or multi-tube flue gas ducts (smoke tubes). Tubes (smoke tubes) are also often used in the construction of flue gas/water heat exchangers with single-line flue gas flow, where the tube exchanger is located above the furnace. *Energies* **2020**, *13*, x FOR PEER REVIEW 3 of 16 gas ducts (smoke tubes). Tubes (smoke tubes) are also often used in the construction of flue gas/water heat exchangers with single-line flue gas flow, where the tube exchanger is located above the furnace. The aim of the work was to build a prototype of a 10 kW boiler with an innovative furnace design and

The aim of the work was to build a prototype of a 10 kW boiler with an innovative furnace design and convection flue gas circulation channels. The boiler is designed to burn pellets made of wood biomass in the continuous feeding mode. The aim of the work was achieved by carrying out numerical and experimental research. Numerical tests included the following: (i) modelling the height of the combustion chamber by analysing the combustion process of syngas produced from biomass in the retort burner; (ii) flue gas flow, temperature distribution and heat exchange calculation in the furnace and convection channels; and, (iii) concentration of gaseous products of CO, CO<sup>2</sup> combustion: NO<sup>x</sup> and O2. The experimental tests consisted in the verification of the obtained results from the numerical analysis in the scope of the following: (i) temperature measurement in the furnace and in the flue, (ii) measurement of gas concentration of combustion products, and (iii) determination of the boiler efficiency. The key assumption of the innovative design concept was to place the furnace with the burner in the central part of the boiler body. The adopted concept of the location of the furnace forces the circulation of hot exhaust gases in a four-pass flow. Flow I is the hearth, Flows II, III, and IV form a system of circular convection flow of exhaust gases around the hearth. Numerical calculations and then the verification of compliance of the calculation results with the results of emission measurements play an important role in the process of design development. The offered boiler is the only solution of this type with a characteristic central location of the furnace and circulating flue gas. convection flue gas circulation channels. The boiler is designed to burn pellets made of wood biomass in the continuous feeding mode. The aim of the work was achieved by carrying out numerical and experimental research. Numerical tests included the following: (i) modelling the height of the combustion chamber by analysing the combustion process of syngas produced from biomass in the retort burner; (ii) flue gas flow, temperature distribution and heat exchange calculation in the furnace and convection channels; and, (iii) concentration of gaseous products of CO, CO2 combustion: NOx and O2. The experimental tests consisted in the verification of the obtained results from the numerical analysis in the scope of the following: (i) temperature measurement in the furnace and in the flue, (ii) measurement of gas concentration of combustion products, and (iii) determination of the boiler efficiency. The key assumption of the innovative design concept was to place the furnace with the burner in the central part of the boiler body. The adopted concept of the location of the furnace forces the circulation of hot exhaust gases in a fourpass flow. Flow I is the hearth, Flows II, III, and IV form a system of circular convection flow of exhaust gases around the hearth. Numerical calculations and then the verification of compliance of the calculation results with the results of emission measurements play an important role in the process of design development. The offered boiler is the only solution of this type with a characteristic central location of the furnace and circulating flue gas.

#### **2. Materials and Methods 2. Materials and Methods**

The paper proposes a 3-stage, semi-empirical procedure for creating a numerical model of the boiler (Figure 1). After determining the target boiler power and combustion technique in the burner, an experiment was carried out to determine the composition of the synthesis gas, which is the product of gasification processes taking place in the burner. These data were used as the boundary condition for the burner model in the cylindrical test chamber. Based on the information (combustion chamber height, temperature distribution, CO and O<sup>2</sup> concentrations) adopted from calculations of syngas combustion in the cylindrical area (Figure 2), a numerical model of the actual boiler combustion chamber was made. Multi-variant simulations were carried out to suggest optimal chamber geometry. The last stage was the prototype of the boiler and the verification of model assumptions. The paper proposes a 3-stage, semi-empirical procedure for creating a numerical model of the boiler (Figure 1). After determining the target boiler power and combustion technique in the burner, an experiment was carried out to determine the composition of the synthesis gas, which is the product of gasification processes taking place in the burner. These data were used as the boundary condition for the burner model in the cylindrical test chamber. Based on the information (combustion chamber height, temperature distribution, CO and O2 concentrations) adopted from calculations of syngas combustion in the cylindrical area (Figure 2), a numerical model of the actual boiler combustion chamber was made. Multivariant simulations were carried out to suggest optimal chamber geometry. The last stage was the prototype of the boiler and the verification of model assumptions.

**Figure 1.** Scheme of 3-stage, semi-empirical modelling of the boiler. **Figure 1.** Scheme of 3-stage, semi-empirical modelling of the boiler.

**Figure 2.** Retort burner for gasification of pellets: (**a**) view of the burner structure, (**b**) burner model in a cylindrical chamber, (**c**) burner model. **Figure 2.** Retort burner for gasification of pellets: (**a**) view of the burner structure, (**b**) burner model in a cylindrical chamber, (**c**) burner model.

#### *2.1. Composition and Calorific Value of the Generator Gas*

syngas analyser.

**Figure 3.** Gasification of biomass fuel in the retort burner: (1)—syngas chamber, (2)—gasification furnace, (3)—air chamber, (4)—gasification air inlets, (5)—combustion air inlets, (6)—screw feeder, (7)—air blower, (8)—syngas purification battery, (9) measurement of syngas composition—GAS 3000 In the boiler furnace, the generator gas (syngas) will be burned, which was obtained in the process of wood biomass gasification; the wood biomass was in the form of pellets with a diameter of 3 mm and a length of about 10–15 mm (average values). The fuel characteristics of the tested biomass are shown in Table 1. Samples for analytical determinations were taken in accordance with the PN-EN 14778:2011 standard Solid biofuels–Sampling. Fuel properties were determined according to PN-EN 14774-1:2010 Solid biofuels—Determination of moisture content—Drying method—Part 1: Total moisture; PN-EN 14775:2010 Solid biofuels—Determination of Ash; PN-EN 14918: 2010 Solid biofuels—Determination of calorific value; PN-EN 15104:2011 Solid biofuels—Determination of total carbon, hydrogen and nitrogen content—Instrumental methods; PN-EN 15289: 2011 Solid biofuels—Determination of total sulphur and chlorine. The gasification process was carried out in the retort burner chamber (Figure 2a) with an added cylindrical chamber (Figure 2). The gasification factor was air. During gasification, the air demand index expressed as a fuel/air ratio F/A was 0.72, and fuel consumption F = 0.000668 g/s, air gasification consumption A = 0.000927 g/s. Syngas was obtained; its composition in the dry state is presented in Table 2. The concentrations of individual components were determined using the GAS 3000 syngas analyser. The calorific value of syngas was calculated based on Formula (1) [17]:

$$\text{LHV} = 126 \text{[CO]} + 108 \text{[H}\_2\text{]} + 359 \text{ [CH}\_4\text{]} \tag{1}$$

**Table 2.** Syngas components from biomass. **Component Unit Value** where [CO], [H2], and [CH4] percentage volumetric shares of syngas flammable components are indicated.

CO2 % 15.1 **Table 1.** Fuel properties of wood pellets.


LHV kJ/Nm3 4857 \* Expressed on a dry free basis, \*\* Expressed on a dry ash free basis.

*2.2. Numerical Modelling of Combustion Processes* Numerical modelling of the construction of a new boiler type began with the analysis of the combustion process in the burner (Figure 2a) of syngas produced in the process of wood pellets gasification. The composition of syngas obtained during the gasification process is provided in Table 2. In Figure 2b, a model of the burner in a cylindrical chamber is presented, which served as a starting point for determining the correct chamber geometry of the designed boiler. The essence of the The method of measuring the synthesis gas composition is shown in Figure 3. In order to separate the syngas from the environment, a cylindrical syngas chamber (1) was added to the burner, to which this gas flows from the furnace (2). Gas samples were collected through the syngas analyser system (9) to measure its composition. The gasification air (primary air) was supplied by the blowing fan (7) and through the ducts arranged on the perimeter of the lower part (4). This air enters the fuel layer fed continuously to the furnace (2). Mineral residue after the gasification process automatically moves to the top of the furnace, from where periodically, after sliding up the chamber (1), is removed

numerical calculations was the following: (i) determining the height at which, at 100% burner load,

outside. After removing ash, the chamber (1) returns to the sealed position and in this position the syngas samples are taken through the measuring system (8) and (9). The upper air blowing ducts (secondary air) during the synthesis gas composition test are plugged. If the syngas chamber (1) for measuring the synthesis gas composition is removed under normal boiler operation conditions, then syngas combustion will take place thanks to the secondary air supplied through the cleared channels (5) around the periphery of the upper part of the furnace (Figures 2 and 3). *Energies* **2020**, *13*, x FOR PEER REVIEW 5 of 16


**Table 2.** Syngas components from biomass.

**Figure 3.** Gasification of biomass fuel in the retort burner: (1)—syngas chamber, (2)—gasification furnace, (3)—air chamber, (4)—gasification air inlets, (5)—combustion air inlets, (6)—screw feeder, (7)—air blower, (8)—syngas purification battery, (9) measurement of syngas composition—GAS 3000 **Figure 3.** Gasification of biomass fuel in the retort burner: (1)—syngas chamber, (2)—gasification furnace, (3)—air chamber, (4)—gasification air inlets, (5)—combustion air inlets, (6)—screw feeder, (7)—air blower, (8)—syngas purification battery, (9) measurement of syngas composition—GAS 3000 syngas analyser.

#### *2.2. Numerical Modelling of Combustion Processes*

syngas analyser.

in a cylindrical chamber, (**c**) burner model.

**Table 2.** Syngas components from biomass. **Component Unit Value** CO2 % 15.1 CO % 18.4 H2 % 14.2 CH4 % 3.3 N2 % 49.0 LHV kJ/Nm3 4857 *2.2. Numerical Modelling of Combustion Processes* Numerical modelling of the construction of a new boiler type began with the analysis of the Numerical modelling of the construction of a new boiler type began with the analysis of the combustion process in the burner (Figure 2a) of syngas produced in the process of wood pellets gasification. The composition of syngas obtained during the gasification process is provided in Table 2. In Figure 2b, a model of the burner in a cylindrical chamber is presented, which served as a starting point for determining the correct chamber geometry of the designed boiler. The essence of the numerical calculations was the following: (i) determining the height at which, at 100% burner load, all combustible particles undergo combustion in the mixture of synthesis gas and air; and, (ii) determining the temperature distribution that governs the efficiency and low-emission combustion. Numerical calculations were made assuming a syngas flow corresponding to 10 kW with an air excess factor λ = 2. The composition of syngas corresponded to the experimental (real) data given in Table 2. The air flow was calculated according to the following relationships (2) and (3):

$$\text{O}\_{1} = \frac{1}{2}(\text{H}\_{2} + \text{CO}) + 2\text{CH}\_{4} - \text{O}\_{2}[\text{m}^{3}\text{O}\_{2}/\text{m}^{3}\text{ of } fuel] \tag{2}$$

$$\mathbf{L}\_{\mathbf{l}} = \frac{\mathbf{O}\_{\mathbf{l}}}{0.21} \mathbf{[m^3 \ of \ air/m^3 \ of \ fuel]} \tag{3}$$

In the second step of numerical modelling, a full boiler model with flue gas flow and positioning of the stake according to the original concept was made.

The numerical grid of the cylindrical combustion chamber with the burner was built of 622,000 polyhedral cells, while the numerical grid of the full boiler model consists of 945.000 polyhedral cells. In order to simulate the gas phase reactions within the full geometry of a biomass furnace in CFD, models for turbulence and radiation were selected [18]. The implementation of the realizable k-ε (enhanced wall treatment) turbulence model enables the Reynolds turbulent equations to be closed. The discrete ordinates radiation model was used to simulate the radiation heat transfer. Combustion reactions are complex and simplified into a series of global reactions with CFD modelling. The hybrid Finite-Rate/Eddy-Dissipation model [18] with the typical reactions scheme can be summarized in the form of kinetics rates [19,20]:

$$\text{CH}\_4 + 1.5\text{O}\_2 \rightarrow \text{CO} + 2\text{H}\_2\text{O Ar (s} - 1) = 5.012 \times 10^{11} \text{ E'r (J/kmol)} = 2.0 \times 10^8 \tag{4}$$

$$\text{CO} + 0.5\text{O}\_2 \rightarrow \text{CO}\_2 \text{ Ar (s} - 1) = 2.239 \times 10^{12} \text{ E'r (J/kmol)} = 1.7 \times 10^8 \tag{5}$$

$$\text{H}\_2 + 0.5\text{O}\_2 \rightarrow \text{H}\_2\text{O Ar (s} - 1) = 9.870 \times 10^8 \text{ E'r (J/kmol)} = 3.1 \times 10^7 \tag{6}$$

The gas phase absorption coefficients are calculated using the weighted-sum-of-grey gases model (WSGGM). Boundary conditions are as described in Table 3. Water cooling was installed on the wall in the cylindrical chamber model. The necessary settings were fixed according to the documentation for the Ansys Fluent packet. In a similar method of modelling, the operation of a biomass gasifying retort burner is presented in [21].


**Table 3.** Boiler's full-load data (N—standard temperature and pressure: 20◦C, 1013 hPa).

#### *2.3. Experimental Set up and Procedure*

Measurements were carried out in the boiler installation belonging to the Poznan University of Technology, Division of Heating, Air Conditioning and Air Protection. Experiments were performed in almost real-life conditions of small power boilers. Boiler installation with measurement equipment is presented in Figure 4.

Pollutant concentrations in the flue gas and flue gas temperature were measured using the Vario Plus (manufacturer—MRU GmbH Germany) flue gas analyser. CO and CxH<sup>y</sup> concentrations were measured using the infrared procedure. Oxygen (O2), nitric oxide (NO), nitrogen dioxide (NO2) concentrations were measured with electrochemical cells. The gas analyser Vario Plus also calculated the air excess ratio and chimney loss for the boiler.

the air excess ratio and chimney loss for the boiler.

concentrations were measured with electrochemical cells. The gas analyser Vario Plus also calculated

Heat received by the boiler water and boiler heat output were measured with a Kamstrup heat

**Figure 4.** Boiler installation with measurement equipment. **Figure 4.** Boiler installation with measurement equipment.

**3. Results and Discussions** *3.1. Results of Numeric Simulations* Heat received by the boiler water and boiler heat output were measured with a Kamstrup heat meter. Fuel-mas was measured with Sartorius weight. The boiler heat efficiency was calculated as heat transferred to the boiler water, divided by fuel-mas, multiplied by fuel lower heating value.

paths are given in Figures 5–11. Table 4 presents the temperature and flue gas composition values in

#### Examples of the results of temperature distribution, combustion gas products, and exhaust gas **3. Results and Discussions**

#### the outlet plane to the chimney from the target boiler model. Analysing the results (Figure 5c), it *3.1. Results of Numeric Simulations*

should be noted that only in the central part of the cylindrical chamber are there favourable conditions for the rapid burning of carbon monoxide (the temperature exceeds 882 K, i.e., the flash point of CO and other compounds, i.e., CH4 and H2). The observed CO concentration may result in the formation inside the boiler chamber, of low temperature and low O2 content zones, which are unfavourable for reducing flammable compounds. Consequently, this will lead to their emission through the chimney to the atmosphere. Based on the results of the numerical calculations of syngas combustion (syngas composition Table 2) in the burner (Figure 5), with the assumed geometry of the burner (Figure 6), the height of the combustion chamber of the boiler h = 285 mm was determined. Figure 7 shows the calculated exhaust gas path from the burner through the combustion chamber located in the central part of the boiler. It should be emphasized that, in the proposed geometry of the combustion chamber in the boiler (Figures 6 and 2a), the effect of flue gas swirling was obtained (Figure 7), which promotes afterburning of carbon monoxide and also intensifies heat exchange with the water jacket. The combustion of carbon monoxide is additionally favoured by the high temperature zone, which exceeds the flash point of CO of 973 K (700 °C) in the cross section (Figures 8 and 9). The upward stream of hot exhaust gases visible in Figures 7, 9, and 10 should therefore involve the use of a protective layer of refractory material, e.g., chamotte, etc. The geometry of the Examples of the results of temperature distribution, combustion gas products, and exhaust gas paths are given in Figures 5–11. Table 4 presents the temperature and flue gas composition values in the outlet plane to the chimney from the target boiler model. Analysing the results (Figure 5c), it should be noted that only in the central part of the cylindrical chamber are there favourable conditions for the rapid burning of carbon monoxide (the temperature exceeds 882 K, i.e., the flash point of CO and other compounds, i.e., CH<sup>4</sup> and H2). The observed CO concentration may result in the formation inside the boiler chamber, of low temperature and low O<sup>2</sup> content zones, which are unfavourable for reducing flammable compounds. Consequently, this will lead to their emission through the chimney to the atmosphere. Based on the results of the numerical calculations of syngas combustion (syngas composition Table 2) in the burner (Figure 5), with the assumed geometry of the burner (Figure 6), the height of the combustion chamber of the boiler h = 285 mm was determined. Figure 7 shows the calculated exhaust gas path from the burner through the combustion chamber located in the central part of the boiler. It should be emphasized that, in the proposed geometry of the combustion chamber in the boiler (Figures 6 and 2a), the effect of flue gas swirling was obtained (Figure 7), which promotes afterburning of carbon monoxide and also intensifies heat exchange with the water jacket. The combustion of carbon monoxide is additionally favoured by the high temperature zone, which exceeds the flash point of CO of 973 K (700 ◦C) in the cross section (Figures 8 and 9). The upward stream of hot exhaust gases visible in Figures 7, 9 and 10 should therefore involve the use of a protective layer of refractory material, e.g., chamotte, etc. The geometry of the proposed combustion chamber thus ensures a long residence time (from 1 to 2 s) of gas particles in the high temperature zone (Figure 10). Figure 11 shows numerical analysis of deflector applications that is often used by boiler manufacturers. The obtained results clearly indicate that, in this way, the effect of flue

gas swirling and its internal circulation is reduced, which is favourable for the low-emission nature of the boiler operation. Therefore, in the structure proposed by the authors, this classic solution should be abandoned. The risk of high temperatures affecting the top wall of the chamber can be reduced by a protective layer of refractory material. CO % of mole fraction 0.01 O2 % of mole fraction 10.5 H20 % of mole fraction 1.6 CO2 % of mole fraction 10.9 CO % of mole fraction 0.01 O2 % of mole fraction 10.5 H20 % of mole fraction 1.6 CO2 % of mole fraction 10.9

CH4 % of mole fraction 0.0

CH4 % of mole fraction 0.0

**Table 4.** Conditions in the chimney flue.

**Table 4.** Conditions in the chimney flue.

*Energies* **2020**, *13*, x FOR PEER REVIEW 8 of 16

*Energies* **2020**, *13*, x FOR PEER REVIEW 8 of 16

proposed combustion chamber thus ensures a long residence time (from 1 to 2 s) of gas particles in the high temperature zone (Figure 10). Figure 11 shows numerical analysis of deflector applications that is often used by boiler manufacturers. The obtained results clearly indicate that, in this way, the effect of flue gas swirling and its internal circulation is reduced, which is favourable for the lowemission nature of the boiler operation. Therefore, in the structure proposed by the authors, this classic solution should be abandoned. The risk of high temperatures affecting the top wall of the

proposed combustion chamber thus ensures a long residence time (from 1 to 2 s) of gas particles in the high temperature zone (Figure 10). Figure 11 shows numerical analysis of deflector applications that is often used by boiler manufacturers. The obtained results clearly indicate that, in this way, the effect of flue gas swirling and its internal circulation is reduced, which is favourable for the lowemission nature of the boiler operation. Therefore, in the structure proposed by the authors, this classic solution should be abandoned. The risk of high temperatures affecting the top wall of the

chamber can be reduced by a protective layer of refractory material.

chamber can be reduced by a protective layer of refractory material.

**Figure 5.** (**a**) Temperature distribution (K); (**b**) O2 (mole fraction); (**c**) CO (mole fraction) for the 1st phase of numeric modelling. **Figure 5.** (**a**) Temperature distribution (K); (**b**) O<sup>2</sup> (mole fraction); (**c**) CO (mole fraction) for the 1st phase of numeric modelling. **Figure 5.** (**a**) Temperature distribution (K); (**b**) O2 (mole fraction); (**c**) CO (mole fraction) for the 1st phase of numeric modelling.

**Figure 6.** Geometry of furnace modelling. **Figure 6.** Geometry of furnace modelling. *Energies* **2020 Figure 6.** Geometry of furnace modelling. , *13*, x FOR PEER REVIEW 9 of 16

**Figure 7.** Gas particle paths—temperature scale (K). **Figure 7.** Gas particle paths—temperature scale (K).

**Figure 8.** CO concentration (mole fraction).

**Figure 9.** Temperature distribution (K) with determined temperature zone of above 880 K.

*Energies* 

*3.2. Real Model of the Boiler*

**Figure 7.** Gas particle paths—temperature scale (K).

**Figure 7.** Gas particle paths—temperature scale (K).

*Energies* **2020**, *13*, x FOR PEER REVIEW 9 of 16

**Figure 8.** CO concentration (mole fraction). **Figure 8.** CO concentration (mole fraction). **Figure 8.** CO concentration (mole fraction).

**Figure 9.** Temperature distribution (K) with determined temperature zone of above 880 K. **Figure 9.** Temperature distribution (K) with determined temperature zone of above 880 K. **2020**, *13*, x FOR PEER REVIEW 10 of 16

**Figure 10.** Tracks of gas particles escaping from the burner - time scale (s). **Figure 10.** Tracks of gas particles escaping from the burner—time scale (s).

**Figure 11.** Temperature distribution (K) and gas paths in the boiler model with deflector.

Based on the results obtained from the numerical model of syngas combustion in the retort burner and the numerical model of the boiler, a prototype of a 10 kW boiler with a retort burner, a centrally located combustion chamber, and exhaust gas circulation around the combustion chamber were constructed. The real model of the boiler is shown in Figure 12. The boiler is built of the combustion chamber, where the burner (6) burning fuel in the form of wood pellets is placed at the bottom (the fuel characteristics are presented in Table 1). The furnace is fenced from below from the flue gas circulation chamber (4) and the convection chamber I (2) and the convection chamber II (3) slot parts. It is fenced off by a sliding cast iron or steel ceramic grate with heat-resistant properties (5) located under the burner (6). The grate (5) tightly closes the furnace (1) so that the fumes are not drawn downwards to the circulation chamber (4). In order to remove the ash from biomass gasification, the grate (5) is manually extended, and after removing the ash, it is moved to a tight position. The ash is poured into chamber (4), from where it is removed periodically outwards through the cleaning duct (8). Hot flue gases cool due to natural draft. They then flow from the furnace, through the convection, the flue gas circulation chamber (4) and the convection chamber II, into the smoke conduit (11) and escape to the atmosphere through the chimney. The chimney draft is regulated by a damper (7). The combustion and convection chambers are enclosed with a water jacket (9) (flue gas/water heat exchanger). The exhaust gas circulation process (which is significant) is forced by the temperature difference in the natural course of the furnace (1), convection chamber I (2), recirculation chamber (4) (exhaust gas recurrence) under the burner (6), and convection chamber II (3). Cooled flue gas is directed to the chimney via the smoke conduit (11). Fuel to the retort burner is dispensed from the tank through stub (10). Inspection hatches (12) and (13) periodically allow observation and removal of impurities that can be deposited on the heating surfaces of the furnace

**Figure 10.** Tracks of gas particles escaping from the burner - time scale (s).

**Figure 11.** Temperature distribution (K) and gas paths in the boiler model with deflector. **Figure 11.** Temperature distribution (K) and gas paths in the boiler model with deflector.


**Table 4.** Conditions in the chimney flue.

#### located under the burner (6). The grate (5) tightly closes the furnace (1) so that the fumes are not drawn downwards to the circulation chamber (4). In order to remove the ash from biomass *3.2. Real Model of the Boiler*

*3.2. Real Model of the Boiler*

gasification, the grate (5) is manually extended, and after removing the ash, it is moved to a tight position. The ash is poured into chamber (4), from where it is removed periodically outwards through the cleaning duct (8). Hot flue gases cool due to natural draft. They then flow from the furnace, through the convection, the flue gas circulation chamber (4) and the convection chamber II, into the smoke conduit (11) and escape to the atmosphere through the chimney. The chimney draft is regulated by a damper (7). The combustion and convection chambers are enclosed with a water jacket (9) (flue gas/water heat exchanger). The exhaust gas circulation process (which is significant) is forced by the temperature difference in the natural course of the furnace (1), convection chamber I (2), recirculation chamber (4) (exhaust gas recurrence) under the burner (6), and convection chamber II (3). Cooled flue gas is directed to the chimney via the smoke conduit (11). Fuel to the retort burner is dispensed from the tank through stub (10). Inspection hatches (12) and (13) periodically allow observation and removal of impurities that can be deposited on the heating surfaces of the furnace Based on the results obtained from the numerical model of syngas combustion in the retort burner and the numerical model of the boiler, a prototype of a 10 kW boiler with a retort burner, a centrally located combustion chamber, and exhaust gas circulation around the combustion chamber were constructed. The real model of the boiler is shown in Figure 12. The boiler is built of the combustion chamber, where the burner (6) burning fuel in the form of wood pellets is placed at the bottom (the fuel characteristics are presented in Table 1). The furnace is fenced from below from the flue gas circulation chamber (4) and the convection chamber I (2) and the convection chamber II (3) slot parts. It is fenced off by a sliding cast iron or steel ceramic grate with heat-resistant properties (5) located under the burner (6). The grate (5) tightly closes the furnace (1) so that the fumes are not drawn downwards to the circulation chamber (4). In order to remove the ash from biomass gasification, the grate (5) is manually extended, and after removing the ash, it is moved to a tight position. The ash is poured into chamber (4), from where it is removed periodically outwards through the cleaning duct (8). Hot flue gases cool due to natural draft. They then flow from the furnace, through the convection, the flue gas circulation chamber (4) and the convection chamber II, into the smoke conduit (11) and escape to the atmosphere through the chimney. The chimney draft is regulated by a damper (7). The combustion and convection chambers are enclosed with a water jacket (9) (flue gas/water heat exchanger). The exhaust gas circulation process (which is significant) is forced by the temperature difference in the natural course of the furnace (1), convection chamber I (2), recirculation chamber (4) (exhaust gas recurrence) under the burner (6), and convection chamber II (3). Cooled flue gas is directed to the chimney via the smoke conduit (11). Fuel to the retort burner is dispensed from the tank through stub (10). Inspection hatches (12) and (13) periodically allow observation and removal of impurities that can be deposited on the heating surfaces of the furnace (1) and convection chambers (2) and (3). It should be added that, during stable boiler operation, no intrusive sediment is present on these surfaces. In addition to the exhaust gas circulation process, the central location of the combustion chamber (1) is noteworthy. mm (Figure 12).

This design solution enables the combustion process in chamber (1) to be carried out in adiabatic conditions with a uniformly high temperature distribution, which promotes the efficient combustion of flammable gaseous particles and soot. An expression of thermal processes occurring in the combustion chamber is the absence of tarry (black) deposits on the boiler's heating surfaces. The flue gas/water heat exchange surface is ca. 1.2 m<sup>2</sup> , which translates into relatively small external dimensions of the boiler 10 kW − 700 × 780 × 550 mm (Figure 12). with its own electrically heated automatic ignition device. Air supply was provided to the furnace by a fan integrated with the furnace. The boiler lacked an automatic device with an oxygen probe (lambda sensor) that would measure oxygen concentration in the flue gas downstream the boiler. Air stream could be modified manually by fan speed regulation, ranging from 10 to 100% of its maximum value. Secondary air entered the furnace through a perforated plate located at the bottom of pellet furnace. The fan stopped operating when boiler water temperature reached its maximum value of 85 °C and reinitiated after the temperature decreased 5 °C below this maximum value.

modifying operation and stand-by time of the screw pellet dispenser. The pellet furnace was equipped

*Energies* **2020**, *13*, x FOR PEER REVIEW 11 of 16

(1) and convection chambers (2) and (3). It should be added that, during stable boiler operation, no intrusive sediment is present on these surfaces. In addition to the exhaust gas circulation process, the central location of the combustion chamber (1) is noteworthy. This design solution enables the combustion process in chamber (1) to be carried out in adiabatic conditions with a uniformly high temperature distribution, which promotes the efficient combustion of flammable gaseous particles and soot. An expression of thermal processes occurring in the combustion chamber is the absence of tarry (black) deposits on the boiler's heating surfaces. The flue gas/water heat exchange surface is ca. 1.2 m2, which translates into relatively small external dimensions of the boiler 10 kW − 700 × 780 × 550

The fuel stream was measured several times using a weighing device. Temperature in the combustion chamber was measured about 0.10 m above the flame with a radiation shielded thermocouple PtRhPt connected additionally with a temperature meter for value comparison. Heat received by the boiler water and boiler heat output were measured with an ultrasonic heat meter. All the obtained data (measured continuously) were transmitted to a personal computer via a data acquisition system. For each test run, parameter values were collected every 5 s, and averaged and mean values were calculated. This time interval of data gathering was optimal, since the measured values were not changing very quickly. Mean values for several consecutive test runs were used to obtain the overall mean value for each type of experiment. Uncertainty intervals were calculated for all measurement results with a 0.95 probability. Pellets were supplied from the hopper by means of a fixedspeed screw feeder and gravitationally fell into a small chamber. Subsequently, a horizontal fixed-speed screw pellet dispenser of each furnace introduced pellets into the burning region. Both devices were

**Figure 12.** View of the boiler with 10 kW flue gas circulation: 1—combustion chamber, 2—convection chamber I, 3—convection chamber II, 4—combustion gas circulation chamber, 5—airtight sliding grate closing the combustion chamber, 6–10 kW retort burner, 7—thrust adjustment slide, 8—ash **Figure 12.** View of the boiler with 10 kW flue gas circulation: 1—combustion chamber, 2—convection chamber I, 3—convection chamber II, 4—combustion gas circulation chamber, 5—airtight sliding grate closing the combustion chamber, 6–10 kW retort burner, 7—thrust adjustment slide, 8—ash removal cleaning duct, 9—water jacket (flue gas/water heat exchanger), 10—fuel feed (pellet), 11—flue (exhaust outlet), 12—inspection hatch of the furnace and convection chamber II, 13—inspection hatch for convection chamber II, 14—return water connector, and 15—feed water connector.

The fuel stream was measured several times using a weighing device. Temperature in the combustion chamber was measured about 0.10 m above the flame with a radiation shielded thermocouple PtRhPt connected additionally with a temperature meter for value comparison. Heat received by the boiler water and boiler heat output were measured with an ultrasonic heat meter. All the obtained data (measured continuously) were transmitted to a personal computer via a data acquisition system. For each test run, parameter values were collected every 5 s, and averaged and mean values were calculated. This time interval of data gathering was optimal, since the measured values were not changing very quickly. Mean values for several consecutive test runs were used to obtain the overall mean value for each type of experiment. Uncertainty intervals were calculated for all measurement results with a 0.95 probability. Pellets were supplied from the hopper by means of a fixed-speed screw feeder and gravitationally fell into a small chamber. Subsequently, a horizontal fixed-speed screw pellet dispenser of each furnace introduced pellets into the burning region. Both devices were synchronized and always worked simultaneously. The fuel stream could be regulated manually by modifying operation and stand-by time of the screw pellet dispenser. The pellet furnace was equipped with its own electrically heated automatic ignition device. Air supply was provided to the furnace by a fan integrated with the furnace. The boiler lacked an automatic device with an oxygen probe (lambda sensor) that would measure oxygen concentration in the flue gas downstream the boiler. Air stream could be modified manually by fan speed regulation, ranging from

10 to 100% of its maximum value. Secondary air entered the furnace through a perforated plate located at the bottom of pellet furnace. The fan stopped operating when boiler water temperature reached its maximum value of 85 ◦C and reinitiated after the temperature decreased 5 ◦C below this maximum value.

#### *3.3. Verification of the Numerical Model in Real Conditions*

In experiments carried out with the boiler, we strived to achieve a steady state in order to analyse the behaviour of the boiler. Parameters such as temperature and CO2, O<sup>2</sup> and CO emissions were compared with those collected from numerical simulations, performed using the model (Figures 13 and 14). Analysing the data of the average temperature values in the chimney and CO2, O<sup>2</sup> emissions from a 2-h measurement and analogous values obtained from the numerical simulation, one can notice differences in the range of 2.7–7%. Experimental data show slight oscillations of the measured quantities, which is a derivative of the nature of a pellet-operated heating installation. Steady-state results are obtained from the numerical model. This is the reason for slight differences in values, but it does not affect the general nature of the simulation results obtained, based on which the research was continued. The boiler geometry adopted for numerical calculations was mapped (Figure 6), the design (executive) documentation was prepared, and the prototype presented in Figure 12 was constructed. The fuel used was wood pellets with the fuel characteristics given in Table 1. The results of the wood pellet combustion tests refer to the working boiler in optimal, determined, full load conditions, i.e., = 10 kW. Under these conditions, the boiler obtained the highest efficiency. It was, on average, η = 92.6% at the fuel calorific value LHV = 16,165 kJ/kg, with fuel consumption 0.000668 g/s and the excess air ratio was λ = 2. The efficiency estimation was based on Formula (7) *Energies* **2020**, *13*, x FOR PEER REVIEW 13 of 16 emission depending on the excess air factor (λ) and the nominal boiler power. It should be noted that the proposed combustion technique allows one to obtain average values of NOx concentrations below 200 mg/Nm3 (converted into 10% of oxygen in the flue gas) for the nominal power of 10 kW and to ensure parameters are in line with the requirements of the ECODESIGN Directive. This is confirmed,

$$\eta = \frac{\dot{Q}}{\dot{B} \cdot LHV} \tag{7}$$

where . *Q*—heat stream generated in the boiler (kJ/s), . *B*—fuel stream (kg/s), *LHV*—calorific value of fuel (kJ/kg), η—boiler energy efficiency. seems fully justified to equip boiler installations with a responsive mechanism of reacting to instantaneous emissions by adjusting operating parameters to temporary operating conditions.

nitrogen oxide formation. It should be emphasized that, with restrictive exhaust emission standards, it

not only by the simulation results, but also by experimental measurements. Even with a power cut of

**Figure 13.** Measured values of gaseous combustion products emission CO, CO2, and O2. **Figure 13.** Measured values of gaseous combustion products emission CO, CO2, and O2.

**Figure 14.** The measured value of the flue gas temperature in the chimney duct.

**Figure 13.** Measured values of gaseous combustion products emission CO, CO2, and O2.

emission depending on the excess air factor (λ) and the nominal boiler power. It should be noted that the proposed combustion technique allows one to obtain average values of NOx concentrations below 200 mg/Nm3 (converted into 10% of oxygen in the flue gas) for the nominal power of 10 kW and to ensure parameters are in line with the requirements of the ECODESIGN Directive. This is confirmed, not only by the simulation results, but also by experimental measurements. Even with a power cut of 40%, NOx emission remains stable. Increasing the excess air factor to 2.5 should enable a further reduction of NOx emissions by lowering the flame temperature and limiting the thermal mechanism of nitrogen oxide formation. It should be emphasized that, with restrictive exhaust emission standards, it seems fully justified to equip boiler installations with a responsive mechanism of reacting to

instantaneous emissions by adjusting operating parameters to temporary operating conditions.

**Figure 14. Figure 14.** The measured value of The measured value of the flue gas temperature in the chimney duct. the flue gas temperature in the chimney duct.

For example, in six attempts to burn wood pellets (when the percentage of oxygen and carbon dioxide in the waste gas was O<sup>2</sup> = 10.7% and CO<sup>2</sup> = 9.93%), the content in the carbon monoxide flue gas was on average CO = 91 mg/m<sup>3</sup> . The emission values obtained through the experiment confirmed the reliability of the previously obtained results using the numerical method presented in Figures 5 and 8. The obtained emission and efficiency parameters result from the previously carried out numerical analysis of combustion in the retort burner shown in Figure 2.

The prototype of the boiler proposed in this paper is characterized (according to numerical calculations) by a simple structure with a wide possibility of modernization in the direction of the following: (i) obtaining higher energy efficiency; (ii) low NO<sup>x</sup> emissions; (iii) increasing power, e.g., to 30 kW by increasing the flue gas/water heat exchange surface with unchanged dimensions; and, (iv) increasing the efficiency of the boiler in condensing operation with the possibility of adding (in a simple way) an additional heat exchanger condensing water in the flue gas.

The temperature measurement in the chimney duct (Figure 14) indicated a value of 430 K to 470 K (157–197 ◦C) and the temperature measurement in the combustion chamber indicated 723 K to 773 K (450–500 ◦C). Both values were confirmed by numerical calculations.

Figures 13 and 15 show the concentrations of the gaseous components of the flue gas CO, CO2, O2, and NO<sup>x</sup> measured in the flue (Figure 4) during the 2 h of boiler operation. The low emission of the boiler (which is related to the achievement of complete combustion conditions in it) refers to low concentrations of CO emissions, the values of which are approximately 100 mg/Nm<sup>3</sup> . Low CO emission is associated with the proposed combustion chamber system and the circular circulation of the exhaust gases in the convection channels. Such a thermal-flow system ensures swirling of exhaust gases and turbulence, which promotes the mixing of combustible compounds and oxygen under high temperature conditions. Under these conditions, the instantaneous NO<sup>x</sup> concentration values were in the range of 100–300 mg/Nm<sup>3</sup> , while the average value was 197 mg/Nm<sup>3</sup> . It was consistent with the simulation results, i.e., the value of 187 mg/Nm<sup>3</sup> . Low, similar values of NOx concentration can be obtained by the classic air and fuel staging method [22,23]. Figure 16 shows a simulation of the NO<sup>x</sup> emission depending on the excess air factor (λ) and the nominal boiler power. It should be noted that the proposed combustion technique allows one to obtain average values of NO<sup>x</sup> concentrations below 200 mg/Nm<sup>3</sup> (converted into 10% of oxygen in the flue gas) for the nominal power of 10 kW and to ensure parameters are in line with the requirements of the ECODESIGN Directive. This is

confirmed, not only by the simulation results, but also by experimental measurements. Even with a power cut of 40%, NO<sup>x</sup> emission remains stable. Increasing the excess air factor to 2.5 should enable a further reduction of NO<sup>x</sup> emissions by lowering the flame temperature and limiting the thermal mechanism of nitrogen oxide formation. It should be emphasized that, with restrictive exhaust emission standards, it seems fully justified to equip boiler installations with a responsive mechanism of reacting to instantaneous emissions by adjusting operating parameters to temporary operating conditions. *Energies* **2020**, *13*, x FOR PEER REVIEW 14 of 16 *Energies* **2020**, *13*, x FOR PEER REVIEW 14 of 16

**Figure 15.** Emissions of nitrogen oxides NOx measured in the chimney. **Figure 15.** Emissions of nitrogen oxides NOx measured in the chimney.

**Figure 16.** Numerical simulation of the combined effect of nominal power and λ on NOx concentration **Figure 16.** Numerical simulation of the combined effect of nominal power and λ on NOx concentration in the flue gases.

#### **Figure 16.** Numerical simulation of the combined effect of nominal power and λ on NOx concentration in the flue gases. in the flue gases. **4. Conclusions**

and

and

**4. Conclusions 4. Conclusions** The presented research works aimed at building a prototype of a high-efficiency and low-The presented research works aimed at building a prototype of a high-efficiency and low-emission heating boiler with a capacity of 10 kW included the following:

The presented research works aimed at building a prototype of a high-efficiency and lowemission heating boiler with a capacity of 10 kW included the following: emission heating boiler with a capacity of 10 kW included the following: (1) Experimental and numerical modelling of the wood pellet combustion process in a retort burner,

As a result of the performed work, it was proved that the values according to the adopted numerical models turned out to be consistent with the results obtained on the test stand (where the

As a result of the performed work, it was proved that the values according to the adopted numerical models turned out to be consistent with the results obtained on the test stand (where the

(3) Experimental verification of the results obtained in the numerical model.

(3) Experimental verification of the results obtained in the numerical model.


As a result of the performed work, it was proved that the values according to the adopted numerical models turned out to be consistent with the results obtained on the test stand (where the efficiency and emission tests of the boiler prototype were carried out). The obtained results gave credibility to the innovative concept of boiler construction adopted in the numerical model. A characteristic feature of the innovative design is the centrally located irradiated combustion chamber (furnace) and the circulation of hot exhaust gases in convection channels arranged around the chamber. Such arrangement of the furnace and convection channels ensures (which has been demonstrated numerically and experimentally) quasi-adiabatic combustion at a constant temperature. In the entire volume of the combustion chamber (especially in its upper part and the first convection sequence), it ensures effective mixing of hot exhaust gases with air. This results in low concentrations in the exhaust gases (CO—91 mg/Nm<sup>3</sup> on average, NOx—197 mg/Nm<sup>3</sup> ) and is conducive to achieving high efficiency, the average value of which is 92%. The given values refer to the 10% share of O<sup>2</sup> in the flue gas. Noteworthy is the low, allowable (according to ECODESIGN) NO<sup>x</sup> emission value. With a power cut of 40%, NO<sup>x</sup> emission remains stable. Increasing the excess air factor to 2.5 should enable a further reduction of NO<sup>x</sup> emissions by lowering the flame temperature and limiting the thermal mechanism of nitrogen oxide formation. It should be emphasized that, with restrictive emission standards, it seems fully justified to equip boiler installations with a responsive mechanism of reacting to instantaneous emissions by adjusting operating parameters to temporary operating conditions. The obtained results showed that the adopted innovative, non-standard boiler structure with a capacity of 10 kW can be successfully applied for low-emission heating of residential buildings.

The results obtained from the tests of the 10 kW boiler (prototype) indicate the possibility of further research on improving the structure, e.g.,: (i) maintaining its dimensions by increasing the heating surface in convection channels, e.g., with ribbing, placing a bundle of water pipes in the space above the combustion chamber; and, (ii) the reduction of NOx emissions by returning portion of the hot exhaust gases from e.g., the lower (horizontal) convection duct to the combustion chamber.

**Author Contributions:** Conceptualization, P.M., D.K. and S.P.; Formal analysis, D.K. and S.P.; Investigation, M.J.; Methodology, P.M., D.K. and S.P.; Software, P.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Low-Cost Organic Adsorbents for Elemental Mercury Removal from Lignite Flue Gas**

**Marta Marczak-Grzesik 1,2,\*, Stanisław Budzy ´n <sup>1</sup> , Barbara Tora <sup>3</sup> , Szymon Szufa <sup>4</sup> , Krzysztof Kogut <sup>1</sup> and Piotr Burmistrz <sup>1</sup>**


**Abstract:** The research presented by the authors in this paper focused on understanding the behavior of mercury during coal combustion and flue gas purification operations. The goal was to determine the flue gas temperature on the mercury emissions limits for the combustion of lignites in the energy sector. The authors examined the process of sorption of mercury from flue gases using fine-grained organic materials. The main objectives of this study were to recommend a low-cost organic adsorbent such as coke dust (CD), corn straw char (CS-400), brominated corn straw char (CS-400-Br), rubber char (RC-600) or granulated rubber char (GRC-600) to efficiently substitute expensive dust-sized activated carbon. The study covered combustion of lignite from a Polish field. The experiment was conducted at temperatures reflecting conditions inside a flue gas purification installation. One of the tested sorbents—tire-derived rubber char that was obtained by pyrolysis—exhibited good potential for Hg<sup>0</sup> into Hg2+ oxidation, resulting in enhanced mercury removal from the flue. The char characterization increased elevated bromine content (mercury oxidizing agent) in comparison to the other selected adsorbents. This paper presents the results of laboratory tests of mercury sorption from the flue gases at temperatures of 95, 125, 155 and 185 ◦C. The average mercury content in Polish lignite was 465 <sup>µ</sup>g·kg−<sup>1</sup> . The concentration of mercury in flue gases emitted into the atmosphere was 17.8 <sup>µ</sup>g·m−<sup>3</sup> . The study analyzed five low-cost sorbents with the average achieved efficiency of mercury removal from 18.3% to 96.1% for lignite combustion depending on the flue gas temperature.

**Keywords:** lignite; anthropogenic emission; mercury removal; flue gases purification; low-cost asorbents

## **1. Introduction**

Ecotoxic elements, especially mercury, are particularly hazardous substances among pollutants with no physiological relevance for living organisms. It is generally known that mercury exposure can inflict various health issues, especially neurological, immunological, behavioral and sensory issues [1,2]. The mentioned afflictions were diagnosed in consumers of contaminated fish (Minamata Disease in Japan) and crops (Iraq, Guatemala and Russia). Mercury has an adverse impact on health, as well as long atmospheric lifetime and propensity for deposition in the aquatic environment and in living tissue. Due to the global distribution of mercury, the US Environmental Protection Agency [3] has classified it and its compounds as an air quality threat. Coal combustion release, which has constantly increased following growing worldwide energy demand, is considered as one of the most significant anthropogenic origins of mercury [3–5]. Therefore, due to the element's (wide) exposition, to avoid adverse harmful effects to the respiratory, nervous and immune systems, it is crucial to decrease atmospheric mercury emissions. Conducted research confirms

**Citation:** Marczak-Grzesik, M.; Budzy ´n, S.; Tora, B.; Szufa, S.; Kogut, K.; Burmistrz, P. Low-Cost Organic Adsorbents for Elemental Mercury Removal from Lignite Flue Gas. *Energies* **2021**, *14*, 2174. https:// doi.org/10.3390/en14082174

Academic Editor: S.M. Ashrafur Rahman

Received: 15 March 2021 Accepted: 11 April 2021 Published: 13 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

the severe influence on the environment and warrants worldwide actions for the reduction of emissions [6–8].

In Poland, solid fuel combustion, mainly coal, constitutes the dominating source of mercury emission, exceeding 80% of the country's share [9,10]. Under the adoption of 2010/75/UE Directive (IED—Industrial Emissions Directive) that includes the permitted industrial emissions, the Polish energy sector has introduced the Best Available Technologies (BAT) methods with a permissible emission limit of 1–10 <sup>µ</sup>g·Nm−<sup>3</sup> , depending on the fuel used and plant size [10,11]. At present, the vast majority of units in Polish power plants do not meet the BAT-defined standards for mercury, which imposes the need for technological progress. As new emerging technologies such as clean coal technologies or deep desulfurization and denitrogenation do not meet the European standards for mercury emission, auxiliary means of mercury removal are needed to prevent power plant closures due to non-compliance with regulations. Among the techniques of standalone mercury removal systems, injection of powdered activated carbon (PAC) to flue gasses due to the high specific surface of the material is commonly utilized in the United States of America [12,13].

The efficiency of mercury removal using powdered sorbent injection depends on both physical and chemical properties of the material, flue gas temperature (inversely proportional, i.e. increases in temperature result in sorption decreases) and flue gas constituents (presence of halogen compounds and sulfur trioxide (SO3)). Furthermore, mercury speciation plays a vital role as the Hg2+ form of mercury has good affinity for sorbents capture, whereas Hg<sup>0</sup> is practically not adsorbed. Therefore, oxidation of the mercury in order to increase the share of Hg2+ is an established practice [14–17].

The main disadvantage of activated carbon sorbent usage is the material price, which results in global research [18–21] aimed at the search for new low-cost adsorbents with comparable sorptive properties that can be used as a replacement (low-cost adsorbents). The authors suggest that the use of waste materials with no industrial application can be the most beneficial in both economic and environmental terms [22]. For example, Guangqian L. et al. [23] developed waste-derived sorbents from biomass and brominated flame retarded plastic for mercury removal from coal-fired flue gas. Charpenteau C. et al. [24] proposed coal fly ash as low-cost material. The authors of [25,26] analyzed six low-cost sorbents with the average achieved efficiency of mercury removal of 30.6–92.9% for sub-bituminous coal and 22.8–80.3% for lignite combustion. One of the main objectives of this study was to recommend a low-cost organic sorbent such as coke dust to efficiently substitute expensive activated carbon for mercury removal from flue gas. Other researchers have attempted to develop alternative low-cost yet efficient adsorbents utilizing agricultural and industrial wastes [27–29]. Biochars play a significant role in addressing the current demands of adsorbents for various applications [30,31]. Initial research on specially prepared chars from rubber wastes has proven beneficial for the capture of both mercury [32] and other heavy metals [33]. Considering the amount of mercury released by the energy sector [34–40], the research was organized to determine the potential of rubber char as a sorbent for capturing mercury from flue gas at different temperatures and compare it to other possible adsorbents as an alternative for expensive activated carbons.

The purpose of this study was the application of dust-sized sorbents for reduction of mercury emissions from flue gas. The main objective of this study was to recommend a low-cost organic adsorbent such as coke dust (CD), corn straw char (CS-400), brominated corn straw char (CS-400-Br), rubber char (RC-600) or granulated rubber char (GRC-600) to efficiently substitute expensive dust-sized activated carbon. The study covered combustion of lignite from a Polish field. The experiment was conducted at temperatures reflecting conditions inside a flue gas purification installation.

#### **2. Materials and Methods**

### *2.1. Sorbents Origins*

To compare the efficiency of sorbents, studies consisted of five types of low-cost organic sorbents in three different forms: coke-derived char dust (CD), biomass chars (CS, CS-Br) and rubber-derived chars (RC and GRC). Commercially available activated carbon (AC) was included as industrial flue gas mercury sorbent of choice for comparative reasons. The sorbents used are described in Table 1.

**Table 1.** Description of low-cost organic adsorbent.


LCOA, Low-Cost Organic Adsorbent.

*2.2. Sorbents Analysis*

The scope of the sorbent analysis listed in Section 2.1 includes:


Values of these parameters were determined for the air-dried basis of the sample.

#### *2.3. Mercury Adsorption System*

The stand for measuring mercury sorption from flue gases generated by the combustion of solid fuels is shown in Figure 1. The bench-scale measurement setup consists of tube furnace (3) along with temperature and gas flow regulation with quartz tube (4) as the combustion chamber, gas source (1) and sorbent holder (8). The utilized equipment allows for controllable gas heating at specified points (9). The measurement procedure

*Energies* **2021**, *14*, 2174 4 of 15

by the sorbent.

involves air-fueled coal sample (5) combustion and analysis of flue-gas mercury captured by the sorbent. bustion zone was 20 min. Additionally, LCOA was placed a sorbent trap (7), which measured the concentration of Hg emitted into the atmosphere.

involves air-fueled coal sample (5) combustion and analysis of flue-gas mercury captured

The experiment was conducted with predetermined temperature conditions for sample combustion and sorbent temperature with similar combustion times and airflow. One gram of the sample was positioned in a ceramic boat-shaped crucible and progressively transported to the center of the combustion zone with a metal rod (2). The standardized measurement time after the crucible was introduced to the center of the com-

**Figure 1.** Test stand schematic: (**1**) gas source, air; (**2**) metal rod; (**3**) tube furnace; (**4**) quartz tube; (**5**) coal sample; (**6**) flue mercury adsorption device; (**7**) sorbent trap; (**8**) sorbent holder; (**9**) flue gases heating furnace; and (**10**) set of impingers with mercury survey meter EMP-3 of impingers for mercury speciation measurement**. Figure 1.** Test stand schematic: (**1**) gas source, air; (**2**) metal rod; (**3**) tube furnace; (**4**) quartz tube; (**5**) coal sample; (**6**) flue mercury adsorption device; (**7**) sorbent trap; (**8**) sorbent holder; (**9**) flue gases heating furnace; and (**10**) set of impingers with mercury survey meter EMP-3 of impingers for mercury speciation measurement.

> A more detailed description of the experimental conditions is presented in [25]. For the combustion process, lignite was used. It underwent ultimate and proximate analysis, with the additional steps of mercury and chlorine determination in accordance with the methodology described in Section 2.2. The lignite sample was prepared in accordance with the ISO standards [49]. The characteristic of the used lignite is shown in Section 3.2. The experiment was conducted with predetermined temperature conditions for sample combustion and sorbent temperature with similar combustion times and airflow. One gram of the sample was positioned in a ceramic boat-shaped crucible and progressively transported to the center of the combustion zone with a metal rod (2). The standardized measurement time after the crucible was introduced to the center of the combustion zone was 20 min. Additionally, LCOA was placed a sorbent trap (7), which measured the concentration of Hg emitted into the atmosphere.

> Mercury Speciation Testing To determine mercury speciation in flue gas samples, two sets of impingers (three impingers in each set) and a mercury survey meter EMP-3 (continuous mercury analyzer) were used. For the period of mercury speciation ascertainment, the flue mercury ad-A more detailed description of the experimental conditions is presented in [25]. For the combustion process, lignite was used. It underwent ultimate and proximate analysis, with the additional steps of mercury and chlorine determination in accordance with the methodology described in Section 2.2.

> sorption device (Figure 1, Point 6) was disengaged. Flue gases flowed by a series of impingers (Figure 1, Point 10). Raw flue gas was channeled into two flow streams. In the The lignite sample was prepared in accordance with the ISO standards [49]. The characteristic of the used lignite is shown in Section 3.2.

#### first vessel out of the first gas washing bottle setup, the Hg2+ mercury from the first Mercury Speciation Testing

stream was reduced to Hg<sup>0</sup> with 10% SnCl<sup>2</sup> solution, followed by acidic gas scrubbing with 10% KOH solution and subsequent moisture removal with the third vessel. The purified flue gas stream was then analyzed with an EMP-3 detector for total mercury content determination. The second gas washing bottle setup had an analogous configuration to the first setup, with the difference being the first vessel was used to capture Hg2+ To determine mercury speciation in flue gas samples, two sets of impingers (three impingers in each set) and a mercury survey meter EMP-3 (continuous mercury analyzer) were used. For the period of mercury speciation ascertainment, the flue mercury adsorption device (Figure 1, Point 6) was disengaged. Flue gases flowed by a series of impingers (Figure 1, Point 10). Raw flue gas was channeled into two flow streams. In the first vessel out of the first gas washing bottle setup, the Hg2+ mercury from the first stream was reduced to Hg<sup>0</sup> with 10% SnCl<sup>2</sup> solution, followed by acidic gas scrubbing with 10% KOH solution and subsequent moisture removal with the third vessel. The purified flue gas stream was then analyzed with an EMP-3 detector for total mercury content determination. The second gas washing bottle setup had an analogous configuration to the first setup, with the difference being the first vessel was used to capture Hg2+ mercury out of the second stream with 10% KCl solution. Consequently, the EMP-3 detector analyzed the purified

flue gas from the second stream for elementary mercury. The concentration of Hg2+ was determined based on the difference between total and elementary mercury.

#### *2.4. Methodology of Mercury Adsorption*

To check the reliability of the tests performed, the balance of mercury in the laboratory installation was calculated for each experiment in accordance with the following model:

$$m\_{\text{c}} \cdot \text{C}\_{0} - m\_{\text{ashi}} \cdot \text{C}\_{\text{ash}} = m\_{\text{sorb}} \cdot \left(\text{C}\_{\text{Hg}}'' - \text{C}\_{\text{Hg}}'\right) + m\_{\text{trap}} \cdot \left(\text{C}\_{\text{trap}}'' - \text{C}\_{\text{trap}}'\right) \tag{1}$$

where *<sup>m</sup><sup>c</sup>* is the mass of combusted coal (kg); *<sup>C</sup>*<sup>0</sup> is the mercury content in coal (µg·kg−<sup>1</sup> ); *mash* is the mass of ash from coal combustion (kg); *Cash* is the Hg content in the ash (kg); *msorb* is the weight of tested sorbent (kg); *C* 0 *Hg* and *C* 00 *Hg* are the mercury concentrations in sorbent before and after sorption (µg·kg−<sup>1</sup> ); *mtrap* is the mass of sorbent trap [kg]; and *C* 0 *trap* and *C* 00 *trap* are the mercury concentrations in the trap before and after the experiment (µg·kg−<sup>1</sup> ).

Based on the obtained data and measurements of flue gas mercury concentration before and after adsorption, the Hg adsorption capacities (*q*, µg·g −1 ) of the sorbent samples were calculated using below equation:

$$q = \frac{\mathcal{Q}\_{fg}}{m\_{\rm sorb}} \int\_0^t (\mathcal{C}\_{fg}' - \mathcal{C}\_{fg}') \, dt \tag{2}$$

where *Qf g* is the gas flow rate (m<sup>3</sup> ·min−<sup>1</sup> ) and *C* 00 *f g* and *C* 0 *f g* are the inlet and outlet Hg concentrations in flue gases (µg·Nm−<sup>3</sup> ) at combustion time t (min).

Additionally, mercury removal efficiency of tested sorbents (MR, %) was calculated as:

$$MR = \frac{\mathcal{C}\_{Hg}^{\prime} - \mathcal{C}\_{Hg}^{\prime}}{\mathcal{C}\_0 - \mathcal{C}\_{ash}} \cdot 100\% \tag{3}$$

#### **3. Results and Discussion**

#### *3.1. Sorbent Characteristics*

AC contained 6.1 <sup>µ</sup>g Hg·kg−<sup>1</sup> (Table 2). Relatively small quantities of mercury were also found in CD (8.9 <sup>µ</sup>g·kg−<sup>1</sup> ) and CS-400 (2.7 <sup>µ</sup>g·kg−<sup>1</sup> ) CD and AC were obtained in the carbonization process, therefore they contained minimal quantities of volatiles: 3.2 wt% for CD and 15% for AC. RC-600 and GRC-600 had high mercury content, at 158 and <sup>73</sup> <sup>µ</sup>g·kg−<sup>1</sup> , respectively, as well as ash (19.8% and 19.6%, respectively). Commercial activated carbon (AC) and rubber waste chars (RC-600 and GRC-600) recorded from 3.5 to 18 times higher sulfur content than other examined sorbents. Sorbents had bromine content in the range of 50–730 ppm.

**Table 2.** Properties of the analyzed sorbents.


Mad, moisture in the air-dried basis; Vad, volatile matter in the air-dried basis; Aad, ash in an air-dried basis; Clad, chlorine in the air-dried basis; Cad, carbon in the air-dried basis; Had, hydrogen in the air-dried basis; Sad, sulfur in the air-dried basis; Brad, bromine in the air-dried basis; Hgad, mercury in the air-dried basis.

Coke and rubber waste chars are microporous materials with a moderately developed mesoporous and poor microporous structure (Table 3). The specific surfaces (SBET) of CD,

RC-600 and GRC-600 have several dozen m<sup>2</sup> ·g −1 , while in AC this parameter reaches 670.5 m<sup>2</sup> ·g −1 . The specific surface area of corn straw char is 4.8 m<sup>2</sup> ·g −1 . However, the mesoporous structure is an essential parameter for mercury adsorption because, due to the particle size of mercury, mesopores are considered to be dominant areas of its deposition. RC-600 and GRC-600 sorbents have the highest mesoporous surface among the analyzed samples: 0.38 and 0.17 cm<sup>3</sup> ·g −1 , respectively. Therefore, these samples underwent further qualitative and quantitative phase analysis. of CD, RC-600 and GRC-600 have several dozen m<sup>2</sup> ·g**−**<sup>1</sup> , while in AC this parameter reaches 670.5 m<sup>2</sup> ·g**−**<sup>1</sup> . The specific surface area of corn straw char is 4.8 m<sup>2</sup> ·g**−**<sup>1</sup> . However, the mesoporous structure is an essential parameter for mercury adsorption because, due to the particle size of mercury, mesopores are considered to be dominant areas of its deposition. RC-600 and GRC-600 sorbents have the highest mesoporous surface among the analyzed samples: 0.38 and 0.17 cm<sup>3</sup> ·g**−**<sup>1</sup> , respectively. Therefore, these samples underwent further qualitative and quantitative phase analysis.

Coke and rubber waste chars are microporous materials with a moderately developed mesoporous and poor microporous structure (Table 3). The specific surfaces (SBET)

GRC-600 2.3 3.2 19.6 76.3 0.95 2.66 0.073 0.075 73.0 Mad, moisture in the air-dried basis; Vad, volatile matter in the air-dried basis; Aad, ash in an air-dried basis; Clad, chlorine in the air-dried basis; Cad, carbon in the air-dried basis; Had, hydrogen in the air-dried basis; Sad, sulfur in the air-dried basis; Brad, bromine in the air-dried basis; Hgad,


**Table 3.** Parameters of porous structure for LCOA. **Table 3.** Parameters of porous structure for LCOA**.**

mercury in the air-dried basis.

*Energies* **2021**, *14*, 2174 6 of 15

Qualitative and Quantitative Phase Analysis of Rubber Waste towards as a Sorbent Qualitative and Quantitative Phase Analysis of Rubber Waste Towards as a Sorbent

Based on conducted experimental studies presented in a previous publication [25], sorbents of rubber waste were the most promising material for flue gas mercury capture in lignite combustion processes. Therefore, only this material was selected for Scanning Electron Microscopy (SEM) using Energy Dispersive Spectroscopy (EDS system). Based on conducted experimental studies presented in a previous publication [25], sorbents of rubber waste were the most promising material for flue gas mercury capture in lignite combustion processes. Therefore, only this material was selected for Scanning Electron Microscopy (SEM) using Energy Dispersive Spectroscopy (EDS system).

The results of the chemical analysis for selected RC-600 sample points (Figure 2) are shown in Table 4. The char contains 78.5–80.8% carbon, most of which comes from carbon black used for tire production. Due to the point-based nature of the measurement, this amount does not correspond to the average content of carbon for the representative sample (Table 2). Another probable origin is tire pyrolytic oil that carbonized during rubber processing. The mineral content in the char is assumed to be a result of industry additives present in the tire manufacturing such as fillers, plasticizers, vulcanizing activators and crosslinking additives. The used matter consists of SiO2, ZnO CaO, Al2O3, Na2O and Fe2O<sup>3</sup> (Table 4). The results of the chemical analysis for selected RC-600 sample points (Figure 2) are shown in Table 4. The char contains 78.5–80.8% carbon, most of which comes from carbon black used for tire production. Due to the point-based nature of the measurement, this amount does not correspond to the average content of carbon for the representative sample (Table 2). Another probable origin is tire pyrolytic oil that carbonized during rubber processing. The mineral content in the char is assumed to be a result of industry additives present in the tire manufacturing such as fillers, plasticizers, vulcanizing activators and crosslinking additives. The used matter consists of SiO2, ZnO CaO, Al2O3, Na2O and Fe2O<sup>3</sup> (Table 4).

**Figure 2.** SEM image of RC-600.


**Table 4.** Analysis EDS of selected points of RC-600. **Table 4.** Analysis EDS of selected points of RC-600.

**Figure 2.** SEM image of RC-600.

*Energies* **2021**, *14*, 2174 7 of 15

Figure 3 shows example SEM images of the microstructure of RC-600 and GRC-600 chars. The picture presents small carbon particles aggregated into larger formations as the main constituent of the samples. Figure 3 shows example SEM images of the microstructure of RC-600 and GRC-600 chars. The picture presents small carbon particles aggregated into larger formations as the main constituent of the samples.

**Figure 3.** SEM images of: (**a**) RC-600; and (**b**) GRC-600. **Figure 3.** SEM images of: (**a**) RC-600; and (**b**) GRC-600.

#### *3.2. Properties of Lignite 3.2. Properties of Lignite*

The characteristics of the combusted coal are shown in Table 5. Lignite contained an average of 465 μg·kg−1 of mercury. The average content of halogens such as Cl and Br (supporting factors in the oxidation of mercury from Hg<sup>0</sup> to Hg2+) in lignite was equal to 30 ppm Cl and 3.9 ppm Br. The average sulfur content for the lignite was 1.8 wt%; it also contained 23.7 wt% ash, 41.20 wt% volatiles and 12.9 wt% moisture. The characteristics of the combusted coal are shown in Table 5. Lignite contained an average of 465 <sup>µ</sup>g·kg−<sup>1</sup> of mercury. The average content of halogens such as Cl and Br (supporting factors in the oxidation of mercury from Hg<sup>0</sup> to Hg2+) in lignite was equal to 30 ppm Cl and 3.9 ppm Br. The average sulfur content for the lignite was 1.8 wt%; it also contained 23.7 wt% ash, 41.20 wt% volatiles and 12.9 wt% moisture.

**Table 5.** Characteristic of the coal used in the experiment. **Table 5.** Characteristic of the coal used in the experiment.


#### Lignite 12.9 41.20 23.7 43.5 4.90 1.80 2.42 3.9 30 465.0 *3.3. Determination Mercury Speciation in Flue Gases during Lignite Combustion*

*3.3. Determination Mercury Speciation in Flue Gases during Lignite Combustion* Figure 4 presents a Sankey diagram for mercury released from bench-scale lignite combustion with determination for various species of Hg at the flue gases temperature of 95 °C. Due to the experimental process, ash was a solid residue after coal combustion, Figure 4 presents a Sankey diagram for mercury released from bench-scale lignite combustion with determination for various species of Hg at the flue gases temperature of 95 ◦C. Due to the experimental process, ash was a solid residue after coal combustion, and nearly all the mercury contained in the coal changed to the flue gas in gaseous forms (Hg<sup>0</sup> and Hg2+). Only 2% remained in the ashes in the Hgp form. The proportion of Hg<sup>0</sup> in the analyzed flow was relatively high, reaching 70%. The Hg<sup>0</sup> :Hg2+ ratio (5:2) was determined by the chemical composition of the fuel. Lignite is characterized by low content

of chlorine and bromine (Table 5), which leads to oxidation of Hg<sup>0</sup> to Hg2+. This effect, however, can be partially slowed down by relatively high presence of calcium (2.42 wt%), which is capable of chemical deactivation of chlorine due to their chemical affinity [50]. The resulting fuel composition determines the share of Hg2+ to be 28% or lower. Due to behavioral differences between both forms, Hg<sup>0</sup> is generally more difficult to remove from flue gas by the adsorption process. Therefore, the determining factor in the selection of sorbent and process conditions for mercury removal should be the chemical composition of the sorbents as well as the sorption temperature. chlorine and bromine (Table 5), which leads to oxidation of Hg<sup>0</sup> to Hg2+. This effect, however, can be partially slowed down by relatively high presence of calcium (2.42 wt%), which is capable of chemical deactivation of chlorine due to their chemical affinity [50]. The resulting fuel composition determines the share of Hg2+ to be 28% or lower. Due to behavioral differences between both forms, Hg<sup>0</sup> is generally more difficult to remove from flue gas by the adsorption process. Therefore, the determining factor in the selection of sorbent and process conditions for mercury removal should be the chemical composition of the sorbents as well as the sorption temperature.

and nearly all the mercury contained in the coal changed to the flue gas in gaseous forms (Hg<sup>0</sup> and Hg2+). Only 2% remained in the ashes in the Hgp form. The proportion of Hg<sup>0</sup>

mined by the chemical composition of the fuel. Lignite is characterized by low content of

in

:Hg2+ ratio (5:2) was deter-

*Energies* **2021**, *14*, 2174 8 of 15

the analyzed flow was relatively high, reaching 70%. The Hg<sup>0</sup>

**Figure 4.** Sankey diagrams of mercury distribution (%). **Figure 4.** Sankey diagrams of mercury distribution (%).

#### *3.4. Hg Adsorption Performance during Combustion of Lignite*

*3.4. Hg Adsorption Performance during Combustion of Lignite* Table 6 shows the Hg adsorption capacity (q) of the tested sorbents and the efficiency of Hg removal from flue gas (MR) at process temperature of 95 °C. The commercial activated carbon, currently used in active flue gas mercury removal methods, was the most efficient. AC removed the mercury almost entirely. The sorption efficiency of CD was also high at 93.8%. RC-600 and GRC-600 presented mercury removal during combustion of lignite of 81.5% and 65.7%, respectively. The analyses showed corn straw char to be the worst sorptive material during lignite combustion. CS-400 decreased the concentration of mercury in flue gas by only 32.4%. CS-400-Br was more efficient in mercury Table 6 shows the Hg adsorption capacity (q) of the tested sorbents and the efficiency of Hg removal from flue gas (MR) at process temperature of 95 ◦C. The commercial activated carbon, currently used in active flue gas mercury removal methods, was the most efficient. AC removed the mercury almost entirely. The sorption efficiency of CD was also high at 93.8%. RC-600 and GRC-600 presented mercury removal during combustion of lignite of 81.5% and 65.7%, respectively. The analyses showed corn straw char to be the worst sorptive material during lignite combustion. CS-400 decreased the concentration of mercury in flue gas by only 32.4%. CS-400-Br was more efficient in mercury removal (MR at 50%).

**Table 6.** Efficacy assessment of LCOA for Hg sorption.

removal (MR at 50%).


As shown in Figure 5, the Hg concentration in raw flue gas was 17.8 <sup>µ</sup>g·m−<sup>3</sup> (in the process of lignite combustion with a Hg content of 465 <sup>µ</sup>g·kg−<sup>1</sup> ). Most of the mercury from flue gases was removed by AC and CD: their Hg adsorption capacities were 102.6 and 100.2 µg·g −1 , while the concentration of Hg in flue gases was decreased by 17.1 and 6.7 <sup>µ</sup>g·m−<sup>3</sup> , respectively, for AC and CD (Figure 5). Application of corn straw chars resulted in incomplete mercury removal. Hg adsorption capacity for CS-400 and CS-400-Br reached 34.3 and 53.4 µg·g −1 , respectively, which caused the reduction of mercury emission to 12.1 and 8.9 <sup>µ</sup>g·m−<sup>3</sup> . The low sorptive capability of CS-400 can be explained by its specific surface of 4.8 m<sup>2</sup> ·g −1 . A higher value of q was obtained for rubber chars; the adsorption capacity was 87.0 µg·g −1 for RC-600 and 70.2 µg·g −1 for GRC-600. As a result, Hg concentration in flue gases was reduced to 3.3 and 6.1 <sup>µ</sup>g·m−<sup>3</sup> , respectively. As shown in Figure 5, the Hg concentration in raw flue gas was 17.8 μg·m**−**<sup>3</sup> (in the process of lignite combustion with a Hg content of 465 μg·kg**−**<sup>1</sup> ). Most of the mercury from flue gases was removed by AC and CD: their Hg adsorption capacities were 102.6 and 100.2 μg·g**−**<sup>1</sup> , while the concentration of Hg in flue gases was decreased by 17.1 and 6.7 μg·m**−**<sup>3</sup> , respectively, for AC and CD (Figure 5). Application of corn straw chars resulted in incomplete mercury removal. Hg adsorption capacity for CS-400 and CS-400-Br reached 34.3 and 53.4 μg·g**−**<sup>1</sup> , respectively, which caused the reduction of mercury emission to 12.1 and 8.9 μg·m**−**<sup>3</sup> . The low sorptive capability of CS-400 can be explained by its specific surface of 4.8 m<sup>2</sup> ·g**−**<sup>1</sup> . A higher value of q was obtained for rubber chars; the adsorption capacity was 87.0 μg·g**−**<sup>1</sup> for RC-600 and 70.2 μg·g**−**<sup>1</sup> for GRC-600. As a result, Hg concentration in flue gases was reduced to 3.3 and 6.1 μg·m**−**<sup>3</sup> , respectively.

CS-400 34.3 32.4 CS-400-Br 53.4 50.1 RC-600 87.0 81.5 GRC-600 70.2 65.7

*Energies* **2021**, *14*, 2174 9 of 15

**Figure 5.** Concentration of mercury in raw flue gas and flue gases cleaned by LCOA. **Figure 5.** Concentration of mercury in raw flue gas and flue gases cleaned by LCOA.

#### *3.5. Effect of Flue Gas Temperature on Mercury Sorption Ability 3.5. Effect of Flue Gas Temperature on Mercury Sorption Ability*

The next stage of the study considered the influence of the adsorption temperature on the effectiveness of LCOA in Hg removal. The average mercury removal effectiveness for AC, CD and CS-400 decreased as the temperature increased, with the highest reading at 95 °C and the lowest at 185 °C (Table 7). The change in adsorption temperature had a different effect on RC-600, GRC-600 and CS-400-Br. The average mercury removal efficiency increased as the adsorption temperature increased, with the lowest results at 95 °C and the highest at 185 °C. The next stage of the study considered the influence of the adsorption temperature on the effectiveness of LCOA in Hg removal. The average mercury removal effectiveness for AC, CD and CS-400 decreased as the temperature increased, with the highest reading at 95 ◦C and the lowest at 185 ◦C (Table 7). The change in adsorption temperature had a different effect on RC-600, GRC-600 and CS-400-Br. The average mercury removal efficiency increased as the adsorption temperature increased, with the lowest results at 95 ◦C and the highest at 185 ◦C.


**Table 7.** Efficacy assessment of LCOA for Hg sorption at different temperatures.

The average mercury concentration in flue gases during lignite combustion was 17.8 <sup>µ</sup>g·m−<sup>3</sup> (Figure 6), whereas, after the adsorption process in the temperature range (95–185 ◦C), the readings showed 1.1–5.8 <sup>µ</sup>g·m−<sup>3</sup> for AC, 0.7–3.8 <sup>µ</sup>g·m−<sup>3</sup> for CS and 12.1–14.6 <sup>µ</sup>g·m−<sup>3</sup> for CS-400. CS-400 presented poor Hg removal performance at all tested temperatures, while the brominated CS-400-Br showed progression in Hg removal efficiency as temperature increased, similarly to RC-600 and GRC-600 (Figure 6).

The deciding factor for Hg removal efficiency can be the share of Hg<sup>0</sup> in the flue. For AC, CD and CS-400, the Hg removal efficiency decreased with further increase in the adsorption temperature, whereas, for CS-400-Br, RC-600, GRC-600, the efficiency increased.

It was speculated that high adsorption temperature causes Hg desorption from the sorbent surface. However, the experiment showed that sorption temperature increase led to a higher share of Hg<sup>0</sup> in the flue gas. The higher Hg removal of CS-400-Br, RC-600 and GRC-600 as the temperature raised can be connected to higher Br and Cl intrinsic content.

The deciding factor for Hg removal efficiency can be the share of Hg<sup>0</sup> Effect of Bromine Content in the Sorbent on Hg<sup>0</sup> Oxidation

AC, CD and CS-400, the Hg removal efficiency decreased with further increase in the adsorption temperature, whereas, for CS-400-Br, RC-600, GRC-600, the efficiency increased. It was speculated that high adsorption temperature causes Hg desorption from the Flue gases from lignite combustion are characterized by the high share of unoxidized Hg<sup>0</sup> mercury. It is a volatile constituent of the flue with its ratio versus Hg2+ increasing with sorption temperature. As Hg<sup>0</sup> content increases, the average Hg removal efficiency for AC, CD and CS was the highest for 95 ◦C and the lowest for 185 ◦C.

in the flue. For

sorbent surface. However, the experiment showed that sorption temperature increase led to a higher share of Hg<sup>0</sup> in the flue gas. The higher Hg removal of CS-400-Br, RC-600 and GRC-600 as the temperature raised can be connected to higher Br and Cl intrinsic content. AC, CD and CS-400 sorbency was determined by low bromine content, which resulted in lesser Hg2+ and therefore higher Hg<sup>0</sup> amounts. The surface of the sorbent allows for Hg2+ adsorption, which can explain the pattern of higher sorptive properties of AC, CD and CS-400 at higher sorption temperatures, as the Hg0 content increases.

Effect of Bromine Content in the Sorbent on Hg<sup>0</sup> Oxidation Flue gases from lignite combustion are characterized by the high share of unoxidized Hg<sup>0</sup> mercury. It is a volatile constituent of the flue with its ratio versus Hg2+ increasing with sorption temperature. As Hg<sup>0</sup> content increases, the average Hg removal efficiency for AC, CD and CS was the highest for 95 °C and the lowest for 185 °C. AC, CD and CS-400 sorbency was determined by low bromine content, which resulted in lesser Hg2+ and therefore higher Hg<sup>0</sup> amounts. The surface of the sorbent allows for Hg2+ adsorption, which can explain the pattern of higher sorptive properties of AC, On the other hand, RC-600, GRC-600 and CS-400-Br were characterized by 2–14 times higher bromine content. Sorbent-present Br oxidized Hg0 in the flue to Hg2+ form at higher sorption temperatures, which obtained higher Hg2+ concentrations, along with higher removal capabilities [50]. The average Hg removal for RC-600, GRC-600 and CS-400-Br followed the ascending order: 95 ◦C < 125 ◦C < 155 ◦C < 185 ◦C. The proposed Hg<sup>0</sup> oxidation and Hg2+ adsorption mechanism on RC-600, GRC-600 and CS-400-Br surfaces are presented in Figure 7. Hg<sup>0</sup> was first oxidized at the activated site on the sorbent, followed by Hg2+ sorption.

On the other hand, RC-600, GRC-600 and CS-400-Br were characterized by 2–14 times higher bromine content. Sorbent-present Br oxidized Hg0 in the flue to Hg2+ form at higher sorption temperatures, which obtained higher Hg2+ concentrations, along with higher removal capabilities [50]. The average Hg removal for RC-600, GRC-600 and CS-400-Br followed the ascending order: 95 °C < 125 °C < 155 °C < 185 °C. The proposed Hg<sup>0</sup> oxidation and Hg2+ adsorption mechanism on RC-600, GRC-600 and CS-400-Br surfaces are presented in Figure 7. Hg<sup>0</sup> was first oxidized at the activated site on the sorbent,

CD and CS-400 at higher sorption temperatures, as the Hg0 content increases.

followed by Hg2+ sorption.


**Figure 7.** Effect of bromine content in the sorbent on Hg<sup>0</sup> oxidation**. Figure 7.** Effect of bromine content in the sorbent on Hg<sup>0</sup> oxidation. **Figure 7.** Effect of bromine content in the sorbent on Hg<sup>0</sup> oxidation**.**

To further assess the application prospect of such rubber waste-derived sorbents, a comparison of the accumulative Hg adsorption capacity of RC-600 and GRC-600 with other sorbents was conducted under similar experimental conditions (Figure 8). It demonstrated that longer sorption times (10–60 min, every 10 min) led to increases in the accumulative properties for bromine-present sorbents. No time influence was observed for bromine-free (AC, CD and CS-400) sorbents. To further assess the application prospect of such rubber waste-derived sorbents, a comparison of the accumulative Hg adsorption capacity of RC-600 and GRC-600 with other sorbents was conducted under similar experimental conditions (Figure 8). It demonstrated that longer sorption times (10–60 min, every 10 min) led to increases in the accumulative properties for bromine-present sorbents. No time influence was observed for bromine-free (AC, CD and CS-400) sorbents. To further assess the application prospect of such rubber waste-derived sorbents, a comparison of the accumulative Hg adsorption capacity of RC-600 and GRC-600 with other sorbents was conducted under similar experimental conditions (Figure 8). It demonstrated that longer sorption times (10–60 min, every 10 min) led to increases in the accumulative properties for bromine-present sorbents. No time influence was observed for bromine-free (AC, CD and CS-400) sorbents.

**Figure 8.** Accumulative Hg adsorption capacity for the tested sorbents. **Figure 8.** Accumulative Hg adsorption capacity for the tested sorbents. **Figure 8.** Accumulative Hg adsorption capacity for the tested sorbents.

#### **4. Conclusions 4. Conclusions**

**4. Conclusions** This manuscript presents the results of mercury adsorption from lignite flue gas by the active method. For this purpose, we used six low-cost organic adsorbents (AC, CD, CS-400, CS-400-Br, RC-600 and GRC-600). The presented results allow drawing the following conclusions: the efficiency of sorbents for removal of mercury from flue gases at 95 °C decreased successively: AC (96.1%) and CD (93.8%), followed by RC-600 (81.5%) and GRC-600 (65.7%). The CS-400-Br exhibited better Hg removal performance compared to virgin biochar CS-400. The doping of sorbents with bromine resulted in a higher share of Hg0 oxidation to Hg2+ species. Low mercury removal efficiency by CS-400 was This manuscript presents the results of mercury adsorption from lignite flue gas by the active method. For this purpose, we used six low-cost organic adsorbents (AC, CD, CS-400, CS-400-Br, RC-600 and GRC-600). The presented results allow drawing the following conclusions: the efficiency of sorbents for removal of mercury from flue gases at 95 °C decreased successively: AC (96.1%) and CD (93.8%), followed by RC-600 (81.5%) and GRC-600 (65.7%). The CS-400-Br exhibited better Hg removal performance compared to virgin biochar CS-400. The doping of sorbents with bromine resulted in a higher share of Hg0 oxidation to Hg2+ species. Low mercury removal efficiency by CS-400 was This manuscript presents the results of mercury adsorption from lignite flue gas by the active method. For this purpose, we used six low-cost organic adsorbents (AC, CD, CS-400, CS-400-Br, RC-600 and GRC-600). The presented results allow drawing the following conclusions: the efficiency of sorbents for removal of mercury from flue gases at 95 ◦C decreased successively: AC (96.1%) and CD (93.8%), followed by RC-600 (81.5%) and GRC-600 (65.7%). The CS-400-Br exhibited better Hg removal performance compared tovirgin biochar CS-400. The doping of sorbents with bromine resulted in a higher share of Hg0 oxidation to Hg2+ species. Low mercury removal efficiency by CS-400 was caused by its low mesopore volume (0.01 cm<sup>3</sup> ·g −1 ). The Hg removal efficiency for AC, CD and

biochars (CS-400) decreased with the increase of temperature, caused by a higher amount of sorption inactive Hg<sup>0</sup> mercury in the flue gas.

CS-400-Br, RC-600 and GRC-600 enhanced their mercury adsorption capacity with an increase in the temperature. These sorbents had a higher bromine content (2–14 times) than other sorbents. Therefore, it can be confirmed that bromine had positive effects on Hg<sup>0</sup> removal due to its influence on better Hg<sup>0</sup> oxidation as well as adsorption on the free sites of the surface. In this study, a novel sorbent was created with one-step pyrolysis of tire waste for Hg<sup>0</sup> removal from lignite flue gas. This method could combine municipal solid waste disposal and mercury sorbent preparation in one process. The Hg adsorption capacities of RC-600 and GRC-600 were close to those of commercial activated carbons.

**Author Contributions:** Conceptualization, M.M.-G. and S.B.; methodology, S.B. and M.M.-G.; validation, K.K.; formal analysis, S.B. and M.M.-G.; investigation, S.B. and M.M.-G.; resources, S.B., M.M.-G. and S.S.; data curation, M.M.-G.; writing—original draft preparation, M.M.-G.; writing—review and editing, M.M.-G., S.B., P.B., B.T. and S.S.; visualization, M.M.-G.; and funding acquisition, M.M.-G. and S.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was co-financed from the Research Subsidy of the AGH University of Science and Technology for the Faculty of Energy and Fuels (No. 16.16.210.476) and by the National Centre for Research and Development (NCRD) Poland within the LIDER X edition research program. The research and development project is entitled "Prediction of Hg and As distribution during the process of subbituminous and lignite coals combustion in pulverised coal-fired boiler and its flue gas cleanup with use of regression models and neural networks" (grant No. LIDER/33/0183/L-10/18/NCBR/2019). This work was supported by research infrastructure of the AGH Center of Energy.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


*Article*

## **A Study of Combustion Characteristics of Two Gasoline–Biodiesel Mixtures on RCEM Using Various Fuel Injection Pressures**

#### **Ardhika Setiawan <sup>1</sup> , Bambang Wahono <sup>1</sup> and Ocktaeck Lim 2,\***


Received: 22 May 2020; Accepted: 15 June 2020; Published: 24 June 2020

**Abstract:** Experimental research was conducted on a rapid compression and expansion machine (RCEM) that has characteristics similar to a gasoline compression ignition (GCI) engine, using two gasoline–biodiesel (GB) blends—10% and 20% volume—with fuel injection pressures varying from 800 to 1400 bar. Biodiesel content lower than GB10 will result in misfires at fuel injection pressures of 800 bar and 1000 bar due to long ignition delays; this is why GB10 was the lowest biodiesel blend used in this experiment. The engine compression ratio was set at 16, with 1000 µs of injection duration and 12.5 degree before top dead center (BTDC). The results show that the GB20 had a shorter ignition delay than the GB10, and that increasing the injection pressure expedited the autoignition. The rate of heat release for both fuel mixes increased with increasing fuel injection pressure, although there was a degradation of heat release rate for the GB20 at the 1400-bar fuel injection rate due to retarded in-cylinder peak pressure at 0.24 degree BTDC. As the ignition delay decreased, the brake thermal efficiency (BTE) decreased and the fuel consumption increased due to the lack of air–fuel mixture homogeneity caused by the short ignition delay. At the fuel injection rate of 800 bar, the GB10 showed the worst efficiency due to the late start of combustion at 3.5 degree after top dead center (ATDC).

**Keywords:** RCEM; GCI; gasoline; biodiesel; fuel injection pressure

## **1. Introduction**

In recent years, economic improvements have increased consumer purchasing power, leading to increased demand for vehicles and electric devices around the world, which will significantly increase the use of fossil fuels. Because fossil fuel sources are finite, many governments have been conducting research to utilize renewable energy as an alternative to the use of fossil fuels. Biodiesel is such an alternative fuel; it can decrease the usage of conventional petroleum diesel over the long term, since it is produced from animal fat and vegetable oil [1].

The compression ignition (CI) engine system is a very promising candidate for this research. It injects fuel near the top dead center (TDC: the position of an engine's piston when it is at the very top of its stroke), which means that only compressed air is used in the compression step. This produces an ideal cycle in the system during combustion, making the performance of the CI engine better than the spark ignition (SI) engine. A high compression ratio can also be applied to CI engines, making them popular in industrial engines and transportation vehicles due to the high energy that can be produced.

In the last two decades, many renewable sources have been popular as energy resources. For the transportation system, alternate fuels—such as vegetable oils, biodiesel and simple alcohol (ethanol and methanol) blends with diesel—have been extensively explored. However, biodiesel has been widely developed as a replacement for diesel fuel. Blends of 20% biodiesel and lower can be used in diesel equipment with no—or only minor—modifications, although certain manufacturers do not extend warranty coverage if equipment is damaged by these blends. The 6% biodiesel to 20% biodiesel blends are covered by the American society for testing and materials (ASTM) D7467 specification. Biodiesel can also be used in its pure form (B100) but may require certain engine modifications to avoid maintenance and performance problems. The Volkswagen Group has released a statement indicating that several of its vehicles are compatible with B5 and B100 made from rapeseed oil and compatible with the European norms (EN) 14214 standard. The use of the specified biodiesel type in its cars will not void any warranty [2]. Mercedes Benz does not allow diesel fuels containing greater than 5% biodiesel (B5) due to concerns about "production shortcomings". Any damages caused by the use of such non-approved fuels will not be covered by the Mercedes-Benz Limited Warranty. The use of biodiesel is not only limited to diesel cars, but is also for airplanes, trains, generators, etc.

However, the diesel fuel that is mainly used for CI produces high emissions, which has become a major environmental concern that has led to the issuance of strict emissions regulations in various countries. Gasoline has lower emissions and has become an alternative fuel to replace diesel. Putrasari et al. [3] found that the thermal efficiency and combustion duration of mixed gasoline–biodiesel is almost the same as diesel. Many researchers have demonstrated that adding biodiesel content to conventional petroleum diesel will decrease pollution to the environment but also decrease the brake power and thermal efficiency because of increased knocking [4]. Gasoline compression ignition (GCI) engines are promising because of their low emissions and high thermal efficiency characteristics, attracting considerable research interest [5,6].

Low-temperature combustion (LTC) can be achieved by using fuel with a low cetane number or high volatility [7,8]. LTC will reduce NOx emissions and improve air–fuel mixing, which lowers the probability of attaining a fuel-rich region and will ultimately reduce particulate matter (PM) [9]. Gasoline is a good choice for LTC, having a high octane number that creates high autoignition resistance. This leads to a longer ignition delay, resulting in a more homogeneous air–fuel mixture. Additionally, the capability of fuel to evaporate depends on its volatility. As a result, enhanced air–fuel mix and decreased local equivalence ratio were obtained.

The low lubricity [10] and higher vapor pressure characteristics of pure gasoline will damage a conventional common rail system over long-term usage [11]. Biodiesel has great potential to overcome the low lubricity problem of gasoline when applied in a GCI. Additionally, given its high oxygen content, adding biodiesel will enhance the combustion process [12]. Adams et al. [13] investigated the combustion behavior of GCI engines, using a mixed fuel of gasoline and biodiesel. They demonstrated that adding biodiesel to conventional petroleum diesel increases the combustion stability and decreases the required intake temperature.

Improving engine performance and efficiency has been actively researched for years. Ahmed et al. [14] studied the effect of a nanomaterial additive on diesel-n-heptane blends in a diesel engine and found an increase in the engine performance. Mehmet [15] investigated the effect of fuel injection pressure on a diesel engine and found that higher fuel injection pressure decreased the brake-specific fuel consumption (BSFC) and enhanced the brake thermal efficiency (BTE) at low speed. Varying the injection pressure seems to be a promising strategy to improve combustion characteristics; it is the main factor that determines fuel stratification inside the chamber and highly affects the combustion process. Higher fuel injection pressure delivers fuel as smaller droplets, resulting in a higher surface area to volume ratio. This enhances the vapor capability of the fuel and produces more complete combustion.

This experiment analyzes the effect of high injection pressure on a GCI engine that uses gasoline as the main fuel. Generally, the GCI engine has an engine structure similar to a diesel engine and uses common rails and injectors with standard diesel fuel. Further study related to high injection pressure using gasoline needs to be done. The purpose of this study was to investigate the combustion characteristics of a GCI engine. A rapid compression expansion machine (RCEM), which has similar characteristics as a CI engine, was used to represent the GCI engine. Gasoline mixed with a small quantity of biodiesel by volume was used with variable high-pressure fuel injections. The purpose of adding biodiesel content to the gasoline was to increase the cetane number to overcome the automatic ignition resistance of gasoline. The injection rate was tested to investigate the effects of the fuel injection pressure on the different densities of gasoline and biodiesel. By varying the injection pressure for the two different fuel mixtures, the combustion characteristics could be analyzed.

#### **2. Methodology**

#### *2.1. Fuel Preparation*

The two kinds of mixed fuel were investigated: GB10 (90% gasoline and 10% biodiesel) and GB20 (80% gasoline and 20% biodiesel). Biodiesel content lower than GB10 resulted in misfires at fuel injection pressures of 800 bar and 1000 bar due to the long ignition delay. The gasoline–biodiesel mixtures were prepared in glass containers and were stirred for approximately 5 to 15 min to ensure homogeneity. Table 1 shows the chemical composition of biodiesel produced from soybeans [16]. It was necessary to perform our experiments immediately after the mixing process because—given the different densities of the gasoline and the biodiesel—phase separation and crystalline colloid would occur fairly rapidly. Water coolant was added to the fuel containers by means of a stainless steel tube to prevent fuel overheating. The physical properties of the fuels were tested in the laboratory and are shown in Table 2 [3].


**Table 1.** The chemical composition of soybean oil [16].

**Table 2.** Physical properties of fuels [3].


#### *2.2. Test Engine and System Setup*

The experiment was carried out on an RCEM that was designed to replicate the phenomena in one cycle of a CI system. In an RCM/RCEM, the single-shot, rapid compression of a test fuel can be studied *2.2. Test Engine and System Setup* 

in a well-defined and controlled environment without the complex fluid dynamics characteristics of a typical internal combustion engine [17,18]. Figure 1 shows the schematic of the RCEM. It is moved by a 22-kW electric motor and has a 100-mm bore and a 420-mm stroke. Adjusting the screw in the base of the crankshaft will change the compression ratio, which can range from 10 to 23. Temperature sensors were installed in the TDC, body, and BTDC to ensure the uniformity of the initial temperature, which could reach a maximum of 393 K. A Kistler 6052CU20 pressure transducer, connected with a Kistler 5018 amplifier, was used to measure the in-cylinder pressure in the engine. An Autonics E40S8-1800-3-T-24 rotary encoder was used to measure the crank angle position. Both sensors were connected to a Dewetron DEWE-800-CA to log the data. A Bosch 0445110327 seven-hole fuel injector was combined with an injector controller (a Zenobalti ZB-8035 multi-stage injection device) and a common rail solenoid injector peak. To control the injection duration and injection timing, an encoder interface ZB-100 and a ZB-5100 hold driver were used. studied in a well-defined and controlled environment without the complex fluid dynamics characteristics of a typical internal combustion engine [17,18]. Figure 1 shows the schematic of the RCEM. It is moved by a 22-kW electric motor and has a 100-mm bore and a 420-mm stroke. Adjusting the screw in the base of the crankshaft will change the compression ratio, which can range from 10 to 23. Temperature sensors were installed in the TDC, body, and BTDC to ensure the uniformity of the initial temperature, which could reach a maximum of 393 K. A Kistler 6052CU20 pressure transducer, connected with a Kistler 5018 amplifier, was used to measure the in-cylinder pressure in the engine. An Autonics E40S8-1800-3-T-24 rotary encoder was used to measure the crank angle position. Both sensors were connected to a Dewetron DEWE-800-CA to log the data. A Bosch 0445110327 sevenhole fuel injector was combined with an injector controller (a Zenobalti ZB-8035 multi-stage injection device) and a common rail solenoid injector peak. To control the injection duration and injection timing, an encoder interface ZB-100 and a ZB-5100 hold driver were used.

one cycle of a CI system. In an RCM/RCEM, the single-shot, rapid compression of a test fuel can be

*Energies* **2020**, *13*, x FOR PEER REVIEW 4 of 12

**Test Item Unit Test Method Gasoline B100** 

Pour Point °C ASTM D6749:2002 −57 1

Density (15 °C) kg/m3 ISO 12185:2003 712.7 882.3

**Figure 1.** Rapid compression expansion (RCEM) engine setup. **Figure 1.** Rapid compression expansion (RCEM) engine setup.

#### *2.3. Test Procedure 2.3. Test Procedure*

The RCEM was able to run up to 1200 RPM, but—given its large bore and stroke—it was considered dangerous to run the engine at such a high speed. A gearbox was installed to reduce the engine speed to 250 RPM, which was maintained as a parameter in this experiment. The fuel injection pressures were varied at 800, 1000, 1200, and 1400 bar, while the fuel injection rate was set at 1000 µs. The RCEM was able to run up to 1200 RPM, but—given its large bore and stroke—it was considered dangerous to run the engine at such a high speed. A gearbox was installed to reduce the engine speed to 250 RPM, which was maintained as a parameter in this experiment. The fuel injection pressures were varied at 800, 1000, 1200, and 1400 bar, while the fuel injection rate was set at 1000 µs.

To investigate the effect of fuel injection pressure on the different fuel physical characteristics, an injection rate test was conducted. Figure 2 shows the schematic of the fuel measuring system in a glass vessel. A 1.5 bar air pressure was maintained in the vessel by injecting N<sup>2</sup> to create backpressure, as presented in Figure 3. A heating element was used to control the initial temperature within the cylinder and was set at 323 K. Single-fuel injection was controlled at 12.5◦ before top dead center (BTDC).

(BTDC).

To investigate the effect of fuel injection pressure on the different fuel physical characteristics, an injection rate test was conducted. Figure 2 shows the schematic of the fuel measuring system in a glass vessel. A 1.5 bar air pressure was maintained in the vessel by injecting N2 to create backpressure, as presented in Figure 3. A heating element was used to control the initial temperature within the

To investigate the effect of fuel injection pressure on the different fuel physical characteristics, an injection rate test was conducted. Figure 2 shows the schematic of the fuel measuring system in a glass vessel. A 1.5 bar air pressure was maintained in the vessel by injecting N2 to create backpressure, as presented in Figure 3. A heating element was used to control the initial temperature within the

*Energies* **2020**, *13*, x FOR PEER REVIEW 5 of 12

**Figure 2.** Schematic diagram of the fuel injection test rate measurement system. **Figure 2.** Schematic diagram of the fuel injection test rate measurement system. **Figure 2.** Schematic diagram of the fuel injection test rate measurement system.

**Figure 3.** Fuel injection rates of GB10 and GB20. **Figure 3.** Fuel injection rates of GB10 and GB20. **Figure 3.** Fuel injection rates of GB10 and GB20.

#### **3. Results and Discussion 3. Results and Discussion 3. Results and Discussion**

#### *3.1. Effect of Fuel Injection Pressure on the Fuel Injection Flow Rate 3.1. Effect of Fuel Injection Pressure on the Fuel Injection Flow Rate 3.1. E*ff*ect of Fuel Injection Pressure on the Fuel Injection Flow Rate*

Figure 3 shows the fuel injection flow rates of the GB10 and GB20 with fuel injection pressure at 600 to 1400 bar. The injection rates between 600 and 1000 bar show clear differences between the GB10 and GB20. The GB10 had higher injection rates than the GB20 due to its lower content of biodiesel, reaching 0.01, 0.013, 0.015, 0.019, and 0.021 mL/cycle, compared to the GB20, which had injection rates of 0.009, 0.011, 0.014, 0.017, and 0.02 mL/cycle, respectively. When the injection pressure exceeded 1000 bar, there were random differences between the GB10 and GB20. The different densities and viscosities of the gasoline and biodiesel caused the different fuel injection rates; a higher-viscosity fuel (i.e., with a higher biodiesel content) will have a lower fuel injection rate. The viscosity affects Figure 3 shows the fuel injection flow rates of the GB10 and GB20 with fuel injection pressure at 600 to 1400 bar. The injection rates between 600 and 1000 bar show clear differences between the GB10 and GB20. The GB10 had higher injection rates than the GB20 due to its lower content of biodiesel, reaching 0.01, 0.013, 0.015, 0.019, and 0.021 mL/cycle, compared to the GB20, which had injection rates of 0.009, 0.011, 0.014, 0.017, and 0.02 mL/cycle, respectively. When the injection pressure exceeded 1000 bar, there were random differences between the GB10 and GB20. The different densities and viscosities of the gasoline and biodiesel caused the different fuel injection rates; a higher-viscosity fuel (i.e., with a higher biodiesel content) will have a lower fuel injection rate. The viscosity affects Figure 3 shows the fuel injection flow rates of the GB10 and GB20 with fuel injection pressure at 600 to 1400 bar. The injection rates between 600 and 1000 bar show clear differences between the GB10 and GB20. The GB10 had higher injection rates than the GB20 due to its lower content of biodiesel, reaching 0.01, 0.013, 0.015, 0.019, and 0.021 mL/cycle, compared to the GB20, which had injection rates of 0.009, 0.011, 0.014, 0.017, and 0.02 mL/cycle, respectively. When the injection pressure exceeded 1000 bar, there were random differences between the GB10 and GB20. The different densities and viscosities of the gasoline and biodiesel caused the different fuel injection rates; a higher-viscosity fuel (i.e., with a higher biodiesel content) will have a lower fuel injection rate. The viscosity affects the fuel injection rate by increasing the viscous friction force and the hydraulic force at the leakage passage. In contrast, the density of the fuel affects only the hydraulic force. We found that higher fuel injection pressure decreased the effects of density and viscosity on the fuel rate, and the difference in injected fuel capacity is similar. These results were in agreement with the previous study by Kim et al. [19],

who stated that increasing the fuel injection pressure will decrease the effects of the fuel density because the discharge coefficient becomes steady at a high Reynolds number. passage. In contrast, the density of the fuel affects only the hydraulic force. We found that higher fuel injection pressure decreased the effects of density and viscosity on the fuel rate, and the difference in injected fuel capacity is similar. These results were in agreement with the previous study by Kim et al. [19], who stated that increasing the fuel injection pressure will decrease the effects of the fuel

the fuel injection rate by increasing the viscous friction force and the hydraulic force at the leakage

#### *3.2. E*ff*ect of Fuel Injection Pressure on Combustion Characteristics* density because the discharge coefficient becomes steady at a high Reynolds number.

#### 3.2.1. In-Cylinder Pressure and Heat Release Rate *3.2. Effect of Fuel Injection Pressure on Combustion Characteristics*

Figure 4 shows the in-cylinder pressures of GB10 and GB20 at different fuel injection pressures. At an 800 bar fuel injection pressure, GB10 produced much lower in-cylinder pressure than the GB20, i.e., 25.78 bar vs. 33.51 bar, respectively. However, GB20 showed less improvement in in-cylinder pressure than the GB10 when a higher fuel injection pressure was applied. At 1400-bar fuel injection, there was only a slight difference of in-cylinder pressure between the GB10 and GB20, i.e., 40.18 bar and 40.22 bar, respectively. From Table 2, we can see that gasoline has a higher heating value than biodiesel. Therefore, fuel with more gasoline is expected to produce greater in-cylinder pressure. However, even with a lower fuel heat value, GB20 at fuel injection pressure 800–1000 bar produces higher in-cylinder pressure than GB10. As shown in Figure 5, the GB10 ignition delay at 800- and 1000-bar fuel injection pressures occurred for a very long time. The start of combustion (SOC) occurs after top dead center (ATDC), after which the pressure in the cylinder decreases due to volume expansion, resulting in reduced in-cylinder pressure generated during combustion. These phenomena were in agreement with Raeie et al. [20], who studied ignition delay by varying the injection timing. At fuel injection pressures of 1200 and 1400 bar, the fuel injection rate of the GB20 was higher than the GB10. 0.026 and 0.0322 mL/cycle, respectively, for the GB10 and 0.0262 and 0.0329 mL/cycle, respectively, for the GB20. This produced higher heating values for the GB20 at those two fuel injection rates. 3.2.1. In-Cylinder Pressure and Heat Release Rate Figure 4 shows the in-cylinder pressures of GB10 and GB20 at different fuel injection pressures. At an 800 bar fuel injection pressure, GB10 produced much lower in-cylinder pressure than the GB20, i.e., 25.78 bar vs. 33.51 bar, respectively. However, GB20 showed less improvement in in-cylinder pressure than the GB10 when a higher fuel injection pressure was applied. At 1400-bar fuel injection, there was only a slight difference of in-cylinder pressure between the GB10 and GB20, i.e., 40.18 bar and 40.22 bar, respectively. From Table 2, we can see that gasoline has a higher heating value than biodiesel. Therefore, fuel with more gasoline is expected to produce greater in-cylinder pressure. However, even with a lower fuel heat value, GB20 at fuel injection pressure 800–1000 bar produces higher in-cylinder pressure than GB10. As shown in Figure 5, the GB10 ignition delay at 800- and 1000-bar fuel injection pressures occurred for a very long time. The start of combustion (SOC) occurs after top dead center (ATDC), after which the pressure in the cylinder decreases due to volume expansion, resulting in reduced in-cylinder pressure generated during combustion. These phenomena were in agreement with Raeie et al. [20], who studied ignition delay by varying the injection timing. At fuel injection pressures of 1200 and 1400 bar, the fuel injection rate of the GB20 was higher than the GB10. 0.026 and 0.0322 mL/cycle, respectively, for the GB10 and 0.0262 and 0.0329 mL/cycle, respectively, for the GB20. This produced higher heating values for the GB20 at those two fuel injection rates.

**Figure 4. Figure 4. (a)** In-cylinder pressures of GB10 and (b) In-cylinder pressure of GB20. (**a**) In-cylinder pressures of GB10 and (**b**) In-cylinder pressure of GB20.

*Energies* **2020**, *13*, x FOR PEER REVIEW 7 of 12

**Figure 5.** Ignition delays of GB10 and GB20. **Figure 5.** Ignition delays of GB10 and GB20. **Figure 5.** Ignition delays of GB10 and GB20.

As shown in Figure 6, an increase in fuel injection pressure increases the rate of heat release for both fuel mixes, which affects the pressure in the cylinder. With increasing fuel injection pressure, the first stage of heat release or low-temperature heat release (LTHR) also increases. At fuel injection pressures of 1200 and 1400 bar, a significant increase in LTHR is shown for both fuel mixtures. Higher fuel injection pressure will increase spray capability where smaller fuel drops will be produced. This condition will make the fuel more reactive to rising temperatures and pressures, thereby increasing the premixed maximum of HRR. However, at the fuel injection pressure of 1400 bar, the increase in the heat release rate of the GB10 was higher than that of GB20. This was because the early SOC caused in-cylinder peak pressure to occur at 0.24° BTDC, retarding the combustion phasing (CA50) and reducing the heat release rate. As shown in Figure 6, an increase in fuel injection pressure increases the rate of heat release for both fuel mixes, which affects the pressure in the cylinder. With increasing fuel injection pressure, the first stage of heat release or low-temperature heat release (LTHR) also increases. At fuel injection pressures of 1200 and 1400 bar, a significant increase in LTHR is shown for both fuel mixtures. Higher fuel injection pressure will increase spray capability where smaller fuel drops will be produced. This condition will make the fuel more reactive to rising temperatures and pressures, thereby increasing the premixed maximum of HRR. However, at the fuel injection pressure of 1400 bar, the increase in the heat release rate of the GB10 was higher than that of GB20. This was because the early SOC caused in-cylinder peak pressure to occur at 0.24◦ BTDC, retarding the combustion phasing (CA50) and reducing the heat release rate. As shown in Figure 6, an increase in fuel injection pressure increases the rate of heat release for both fuel mixes, which affects the pressure in the cylinder. With increasing fuel injection pressure, the first stage of heat release or low-temperature heat release (LTHR) also increases. At fuel injection pressures of 1200 and 1400 bar, a significant increase in LTHR is shown for both fuel mixtures. Higher fuel injection pressure will increase spray capability where smaller fuel drops will be produced. This condition will make the fuel more reactive to rising temperatures and pressures, thereby increasing the premixed maximum of HRR. However, at the fuel injection pressure of 1400 bar, the increase in the heat release rate of the GB10 was higher than that of GB20. This was because the early SOC caused in-cylinder peak pressure to occur at 0.24° BTDC, retarding the combustion phasing (CA50) and reducing the heat release rate.

**Figure 6.** (**a**) Heat release rates of GB10 and (**b**) Heat release rate of GB20. **Figure 6.** (**a**) Heat release rates of GB10 and (**b**) Heat release rate of GB20. **Figure 6.** (**a**) Heat release rates of GB10 and (**b**) Heat release rate of GB20.

The ignition delay of the engine when operated using GB10 and GB20 fuels is presented in

The ignition delay of the engine when operated using GB10 and GB20 fuels is presented in

3.2.2. Combustion Duration and Ignition Delay

3.2.2. Combustion Duration and Ignition Delay

### 3.2.2. Combustion Duration and Ignition Delay

The ignition delay of the engine when operated using GB10 and GB20 fuels is presented in Figure 6. In this work, the ignition delay is defined as the time interval of fuel injected until combustion occurs. The second peak of the pressure rise rate is defined as the SOC, as described by Lee et al. [21]. Thus, it can be concluded that the ignition delay will be shortened when higher fuel injection pressure is applied in the engine. Increasing the content of biodiesel also affects the ignition process, making it occur faster. These phenomena are caused by the higher biodiesel content in GB20 than in GB10, which means GB20 has a higher cetane number. Generally, the cetane number is related to the autoignition characteristics of the fuel; the GB20 with its higher biodiesel content and higher cetane number will shorten the ignition delay. This is consistent with Adams [13]. A higher cetane number will also enhance the fuel capability to autoignite against compression, resulting in higher engine performance. Increasing the biodiesel content in a gasoline–biodiesel fuel mix will increase the cetane number, which will improve the autoignition speed and engine performance, as proposed by Putrasari et al. [3]. Meanwhile, an increase in fuel injection pressure increases cavitation—the main breakup—and, consequently, fuel atomization shortens the ignition delay as well as the duration of combustion. *Energies* **2020**, *13*, x FOR PEER REVIEW 8 of 12 combustion occurs. The second peak of the pressure rise rate is defined as the SOC, as described by Lee et al. [21]. Thus, it can be concluded that the ignition delay will be shortened when higher fuel injection pressure is applied in the engine. Increasing the content of biodiesel also affects the ignition process, making it occur faster. These phenomena are caused by the higher biodiesel content in GB20 than in GB10, which means GB20 has a higher cetane number. Generally, the cetane number is related to the autoignition characteristics of the fuel; the GB20 with its higher biodiesel content and higher cetane number will shorten the ignition delay. This is consistent with Adams [13]. A higher cetane number will also enhance the fuel capability to autoignite against compression, resulting in higher engine performance. Increasing the biodiesel content in a gasoline–biodiesel fuel mix will increase the cetane number, which will improve the autoignition speed and engine performance, as proposed by Putrasari et al. [3]. Meanwhile, an increase in fuel injection pressure increases cavitation—the main breakup—and, consequently, fuel atomization shortens the ignition delay as well as the duration of combustion.

Figure 7 shows the CA10, 50 and 90, while Figure 8 shows combustion duration of GB10 and GB20 with diferent fuel injection pressures. The combustion phasing can be observed in the CA10, CA50, and CA90 graphs. CA50 shows a very consistent advance when higher fuel injection pressure is applied. The start and the end of combustion can be observed from CA10 and CA90, respectively. The combustion duration of GB10 mostly has a lower burn duration than GB20. Fuel with a higher calorific value will produce a shorter total duration of combustion [22]. By increasing fuel injection pressure, shorter combustion duration due to smaller fuel droplets makes fuel more combustible. However, when the rate of increase in the first stage of heat release increases outside the main HRR, the duration of combustion shows an increase. When the first phase of the heat release phase is completed after TDC, the duration of combustion continues to decrease. Figure 7 shows the CA10, 50 and 90, while Figure 8 shows combustion duration of GB10 and GB20 with diferent fuel injection pressures. The combustion phasing can be observed in the CA10, CA50, and CA90 graphs. CA50 shows a very consistent advance when higher fuel injection pressure is applied. The start and the end of combustion can be observed from CA10 and CA90, respectively. The combustion duration of GB10 mostly has a lower burn duration than GB20. Fuel with a higher calorific value will produce a shorter total duration of combustion [22]. By increasing fuel injection pressure, shorter combustion duration due to smaller fuel droplets makes fuel more combustible. However, when the rate of increase in the first stage of heat release increases outside the main HRR, the duration of combustion shows an increase. When the first phase of the heat release phase is completed after TDC, the duration of combustion continues to decrease.

**Figure 7. Figure 7.** Combustion phasing of CA10, 50 and 90. Combustion phasing of CA10, 50 and 90.

*Energies* **2020**, *13*, x FOR PEER REVIEW 9 of 12

*Energies* **2020**, *13*, x FOR PEER REVIEW 9 of 12

**Figure 8.** Burn duration. **Figure 8.** Burn duration. **Figure 8.** Burn duration.

#### *3.3. Brake Thermal Efficiency and BSFC 3.3. Brake Thermal E*ffi*ciency and BSFC 3.3. Thermal and BSFC*

Figure 9 shows the brake thermal efficiency (BTE) of the engine using GB10 and GB20 with fuel injection pressures of 800 to 1400 bar. Both fuel blends showed the best improvement of BTE when the SOC occurred near TDC: 31.04% at 1000 bar for GB10 and 55.14% at 800 bar for GB20. Decreases in BTE were shown by both fuel mixes when the fuel injection pressure increased. Figure 10 shows the brake-specific fuel consumption (BSFC) of the engine using GB10 and GB20 with fuel injection pressures of 800 to 1400 bar. GB10 seemed to have a higher BSFC than GB20 in every experimental test: it had the highest fuel consumption, reaching 545 g/kWh at an 800-bar fuel injection pressure, compared to GB20's 360 g/kWh at a 1400-bar fuel injection pressure. Figure 9 shows the brake thermal efficiency (BTE) of the engine using GB10 and GB20 with fuel injection pressures of 800 to 1400 bar. Both fuel blends showed the best improvement of BTE when the SOC occurred near TDC: 31.04% at 1000 bar for GB10 and 55.14% at 800 bar for GB20. Decreases in BTE were shown by both fuel mixes when the fuel injection pressure increased. Figure 10 shows the brake-specific fuel consumption (BSFC) of the engine using GB10 and GB20 with fuel injection pressures of 800 to 1400 bar. GB10 seemed to have a higher BSFC than GB20 in every experimental test: it had the highest fuel consumption, reaching 545 g/kWh at an 800-bar fuel injection pressure, compared to GB20's 360 g/kWh at a 1400-bar fuel injection pressure. Figure 9 shows the brake thermal efficiency (BTE) of the engine using GB10 and GB20 with fuel injection pressures of showed the best improvement BTE the SOC occurred near TDC: 31.04% at 1000 bar for GB10 and 55.14% at 800 bar for GB20. Decreases in BTE were shown by both fuel mixes when the fuel injection pressure increased. Figure 10 shows the brake-specific fuel consumption (BSFC) of the engine using GB10 and GB20 with fuel injection pressures of 800 to 1400 bar. GB10 seemed to have a higher BSFC than GB20 in every experimental test: it had the highest fuel consumption, reaching 545 g/kWh at an 800-bar fuel injection pressure, compared to GB20's 360 g/kWh at a 1400-bar fuel injection pressure.

**Figure 9.** Brake thermal efficiency. **Figure 9. Figure 9.** Brake thermal efficiency. Brake thermal efficiency.

*Energies* **2020**, *13*, x FOR PEER REVIEW 10 of 12

**Figure 10.** Brake specific fuel consumption. **Figure 10.** Brake specific fuel consumption.

Increasing the fuel injection pressure will increase the BSFC for all fuel blends. As discussed above, increasing the fuel injection pressure will decrease the ignition delay. A longer ignition delay gives the fuel a chance to mix with the air more perfectly, leading to better quality combustion. However, a too-long or too-short ignition delay will make the engine misfire, and a short ignition delay will produce low combustion quality and efficiency due to insufficient air–fuel mixing. The ignition timing and burn duration will also affect the efficiency of the combustion. However, a tooearly SOC will produce backpressure, decreasing the piston speed before it reaches the TDC. A toolate SOC would lead to a decrease in efficiency because of volumetric effects and delayed fuel conversion processes. This statement is in agreement with Haris [23], who studied the effect of ignition delay (ID) on performance, emission, and combustion characteristics of 2-Methyl Furan-Unleaded gasoline blends. He found that retarded ignition will decrease in BTE. For GB10, the SOC at an 800-bar fuel injection pressure occurred after 4.6° ATDC, as shown in Figure 7, resulting in the lowest thermal and fuel efficiency compared with the other samples. The best efficiency was achieved when the SOC occurred synchronously with TDC, which for GB10 was 2.1° CA ATDC at a 1000-bar fuel injection pressure. For GB20, the closest SOC was achieved at an 800-bar fuel injection pressure, which was 2.6° CA BTDC. Meanwhile, by increasing the value of the cetane number, the efficiency also increased. Yakup [24] states that improvement in engine performance by increasing the cetane number may be attributed to improving the combustion process, as the reason for the highest Increasing the fuel injection pressure will increase the BSFC for all fuel blends. As discussed above, increasing the fuel injection pressure will decrease the ignition delay. A longer ignition delay gives the fuel a chance to mix with the air more perfectly, leading to better quality combustion. However, a too-long or too-short ignition delay will make the engine misfire, and a short ignition delay will produce low combustion quality and efficiency due to insufficient air–fuel mixing. The ignition timing and burn duration will also affect the efficiency of the combustion. However, a too-early SOC will produce backpressure, decreasing the piston speed before it reaches the TDC. A too-late SOC would lead to a decrease in efficiency because of volumetric effects and delayed fuel conversion processes. This statement is in agreement with Haris [23], who studied the effect of ignition delay (ID) on performance, emission, and combustion characteristics of 2-Methyl Furan-Unleaded gasoline blends. He found that retarded ignition will decrease in BTE. For GB10, the SOC at an 800-bar fuel injection pressure occurred after 4.6◦ ATDC, as shown in Figure 7, resulting in the lowest thermal and fuel efficiency compared with the other samples. The best efficiency was achieved when the SOC occurred synchronously with TDC, which for GB10 was 2.1◦ CA ATDC at a 1000-bar fuel injection pressure. For GB20, the closest SOC was achieved at an 800-bar fuel injection pressure, which was 2.6◦ CA BTDC. Meanwhile, by increasing the value of the cetane number, the efficiency also increased. Yakup [24] states that improvement in engine performance by increasing the cetane number may be attributed to improving the combustion process, as the reason for the highest efficiency is shown by GB20 in all tests.

#### efficiency is shown by GB20 in all tests. **4. Conclusions**

**4. Conclusions**  This study of the combustion characteristics of two gasoline–biodiesel mixtures using various This study of the combustion characteristics of two gasoline–biodiesel mixtures using various fuel injection pressures was conducted to gain more understanding of the effects of high fuel injection pressure in a GCI engine. The following conclusions can be drawn from the results of this study.

fuel injection pressures was conducted to gain more understanding of the effects of high fuel injection pressure in a GCI engine. The following conclusions can be drawn from the results of this study. The injection rates for 600 and 1000 bar show a clear difference between GB10 and GB20, caused by the differences in density and viscosity between gasoline and biodiesel. GB10 results in higher The injection rates for 600 and 1000 bar show a clear difference between GB10 and GB20, caused by the differences in density and viscosity between gasoline and biodiesel. GB10 results in higher injection rates than GB20 due to the higher content of biodiesel. However, a higher fuel injection pressure will deter the effects of density and viscosity differences on the fuel rate and injected fuel capacity.

injection rates than GB20 due to the higher content of biodiesel. However, a higher fuel injection pressure will deter the effects of density and viscosity differences on the fuel rate and injected fuel capacity. If the SOC is closer to the TDC it will maximize combustion, so that it produces high in-cylinder pressure. However, when the SOC exceeds TDC, volumetric inefficiency will cause a pressure drop

If the SOC is closer to the TDC it will maximize combustion, so that it produces high in-cylinder

because GB10 at fuel injection pressures of 800 and 1000 bar has a lower in-cylinder pressure than GB20 while having a higher fuel heat value.

An increase in fuel injection will result in an increase in LTHR. Both fuel mixes showed a significant increase in fuel injection pressure above 1200 bar, which exceeded the main HRR and caused an increase of combustion duration. However, the duration of combustion will still decrease if the end of the first stage of heat release is complete after TDC.

The ignition delay decreased as the fuel injection pressure and the cetane content in the fuel increased. With both GB10 and GB20, increasing the fuel injection pressure above 800 bar produced low efficiency due to the shortened ignition delay. At a fuel injection pressure of 800 bar, GB10 produced the slowest ignition delay or SOC at 3.5◦ ATDC. At a fuel injection pressure of 1400 bar, GB20 produced the shortest ignition delay, leading to a retarded in-cylinder peak pressure which occurred at 0.24◦ BTDC, resulting in the lowest efficiency of both fuel mixes.

This study specifically studies the effects of applying high fuel injection pressure in GCI engines with gasoline as the main fuel on combustion characteristics. The idea of increasing efficiency by reducing fuel particles using high fuel injection pressure is not proven to increase efficiency in this system. Not only wasting power, high fuel pressure on gasoline will easily increase the fuel temperature due to fuel friction in the common rail. Fuel with high volatility, such as gasoline, requires a long ignition delay to increase combustion performance. However, further research needs to be done to find out the maximum ignition delay that can be applied to the GCI engine to produce optimal performance.

**Author Contributions:** Conceptualization, A.S.; methodology, A.S. and B.W.; softwere, A.S.; resource, A.S. and B.W.; writing original draft preparation, A.S.; writing review and editing, A.S.; O.L. supervised the research, advised on the research gap and objective, proofread the paper, and guided the writing process as well as reviewing the presented concepts and outcomes. All authors have read and agreed to the published version of the manuscript.

**Acknowledgments:** This research was financially supported by the Energy Technology Development Project of the Korea Energy Technology Evaluation and Planning (20182010106370, Demonstration Research Project of Clean Fuel DME Engine for Fine Dust Reduction), Republic of Korea.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## **Comparative Assessment of Spray Behavior, Combustion and Engine Performance of ABE-Biodiesel**/**Diesel as Fuel in DI Diesel Engine**

**Sattar Jabbar Murad Algayyim 1,2,3,\* and Andrew P. Wandel <sup>2</sup>**


Received: 22 November 2020; Accepted: 7 December 2020; Published: 10 December 2020

**Abstract:** This study investigates the impact of an acetone-butanol-ethanol (ABE) mixture on spray parameters, engine performance and emission levels of neat cottonseed biodiesel and neat diesel blends. The spray test was carried out using a high-speed camera, and the engine test was conducted on a variable compression diesel engine. Adding an ABE blend can increase the spray penetration of both neat biodiesel and diesel due to the low viscosity and surface tension, thereby enhancing the vaporization rate and combustion efficiency. A maximum in-cylinder pressure value was recorded for the ABE-diesel blend. The brake power (BP) of all ABE blends was slightly reduced due to the low heating values of ABE blends. Exhaust gas temperature (EGT), nitrogen oxides (NOx) and carbon monoxide (CO) emissions were also reduced with the addition of the ABE blend to neat diesel and biodiesel by 14–17%, 11–13% and 25–54%, respectively, compared to neat diesel. Unburnt hydrocarbon (UHC) emissions were reduced with the addition of ABE to diesel by 13%, while UHC emissions were increased with the addition of ABE to biodiesel blend by 25–34% compared to neat diesel. It can be concluded that the ABE mixture is a good additive blend to neat diesel rather than neat biodiesel for improving diesel properties by using green energy for compression ignition (CI) engines with no or minor modifications.

**Keywords:** acetone-butanol-ethanol mixture; biodiesel; spray visualization; emissions

## **1. Introduction**

The rapid depletion of fossil fuel reserves, population growth and the increase in air pollution from internal combustion engines using fossil fuels have motivated the search for an alternative biofuel such as biodiesel and alcohol [1,2]. Acetone-butanol-ethanol (ABE), a butanol intermediate product fermentation, has shown potential as an additive fuel blend for conventional diesel due to a reduction in the recovery cost requirements of butanol separation [3–5]. Another benefit of using ABE is that it is produced from renewable sources such as agricultural waste [6–8]. Furthermore, a variety of biomass types can be used as a source of ABE fermentation [3].

The ABE blend has attracted researchers' attention because it is a renewable fuel that reduces dependence on fossil fuels and decreases diesel engine emissions [9,10]. Researchers have experimentally tested ABE mixtures with several investigations [11,12] assessing ABE blend performance under different operating conditions. Luo et al. [13] investigated the sooting tendency of ABE fuels blended with diesel. They found that ABE-diesel has a lower sooting tendency than the butanol-diesel blend because it possesses higher oxygen content and lower carbon content for the same blend ratio. Ma et al. [14] tested droplet evaporation of an ABE blend. The ABE mixture addition enhanced the droplet evaporation speed, and thus reduced the droplet lifetime. Therefore, droplet clouds have a significant impact on the propagation of turbulent flame. The pre-evaporation rate and droplet size are important parameters in controlling burning velocity. Droplet size and the overall number of droplets also have a substantial impact on ignition success, so the high evaporation rate resulting from the additions of an ABE mixture could improve combustion rates [15]. Recently, a study [16] investigated the impact of a butanol-acetone (BA) mixture as an additive for biodiesel fuel on spray and combustion characteristics. The experimental results revealed that all BA mixtures enhanced spray penetration, offered some improvement in brake power and reduced emission levels (UHC, CO and NOx). The abovementioned studies support the advantages of using ABE in compression ignition (CI) engines [3,17,18]. Recent research by the authors [5] has also investigated ABE-diesel blends and found that the studied ABE blend reduced exhaust emissions.

To the best of our knowledge, comparative assessment of the ABE mixture as an additive to neat biodiesel (cottonseed) and neat diesel, and the related spray characteristics and engine performance, have not been fully studied. The main goal of this paper is to evaluate and compare the macroscopic spray parameters and engine performance of 10% ABE blended with neat biodiesel and diesel as fuel in a direct injection (DI) diesel engine.

#### **2. Materials and Methods**

#### *2.1. Fuel Preparation*

Cottonseed biodiesel was prepared from cottonseed oil via transesterification. The fatty acid compositions of the cottonseed biodiesel (chemical profiles) were determined using Flame Ionization Detector-Equipped Gas Chromatography (FID-GC) [19]. The analytical grades of normal butanol (nB) and acetone (A) were used with 99.8% purity, and ethanol was used with 100% purity. All alcohol blends were obtained from Chem-Supply Australia. The ABE blend was mixed with a ratio of A:B:E (3:6:1) by volume and used to simulate ABE fermentation [3]. ABE (10%) was blended with 90% neat cottonseed biodiesel (Bd), which is referred to as ABE10Bd90. ABE (10%) was also blended with 90% neat diesel (D) creating ABE10D90. Blend density was measured according to ASTM 1298 [20,21]. The blends' dynamic viscosities were measured according to the ASTM 445-01 [22] using a Brookfield Viscometer (DV-II+Pro Extra, AMETEK Brookfield, Middleboro, MA, USA). The kinematic viscosity was then calculated [23]. The blends' calorific values were measured using a digital oxygen bomb calorimeter (XRY-1A, Shanghai Changji Geological Instrument Co., Ltd., Shanghai, China) following ASTM D240 [24,25]. The properties of the neat fuel are listed in Table 1, and the measured blends' properties are presented in Figure 1.


**Table 1.** Fuel properties [3,19,21,24].

*Energies* **2020**, *13*, x FOR PEER REVIEW 3 of 12

**Figure 1.** Measured fuel properties. (**a**) Density of test blends; (**b**) Viscosity of test blends; (**c**) Calorifc values of test blends. **Figure 1.** Measured fuel properties. (**a**) Density of test blends; (**b**) Viscosity of test blends; (**c**) Calorifc values of test blends.

#### *2.2. Spray Test Setup 2.2. Spray Test Setup*

The spray test was conducted at atmospheric condition. The setup was consistent with those described in previous work [4,16,20]. Figure 2 shows a schematic diagram of the spray setup system. The spray images were captured using a Photron SA3 high-speed camera. The injector driver specifications, injection setup and camera specification are presented Table 2. Image processing methods were the same as those employed in [4,16,20,26] using the MATLAB (2015R, The MathWorks, Natick, MA, USA) program. The spray test was conducted at atmospheric condition. The setup was consistent with those described in previous work [4,16,20]. Figure 2 shows a schematic diagram of the spray setup system. The spray images were captured using a Photron SA3 high-speed camera. The injector driver specifications, injection setup and camera specification are presented Table 2. Image processing methods were the same as those employed in [4,16,20,26] using the MATLAB (2015R, The MathWorks, Natick, MA, USA) program.

**Figure 2.** Schematic of spray setup system. **Figure 2.** Schematic of spray setup system.

**Table 2.** Specification of visualization system [4,20]. **Table 2.** Specification of visualization system [4,20].


2000 fps

**Injection setup** 

A Nikon AF Micro-Nikkor lens with a focal length of 60 mm and a maximum aperture of f/2.8D with filter size 62 mm was connected to the camera


Repeat time 3

**Table 2.** *Cont.*

## *2.3. Engine Test Setup*

The engine test consisted of a single-cylinder diesel engine, and a Coda gas analyzer used to measure emissions (Figure 3). Table 3 contains the engine specifications. The engine equipment and the Coda gas analyzer's accuracy ranges were used as described in previous work [16,25]. Specific fuel consumption (SFC) for each fuel test was measured using a flow rate meter because of the differences in density and heating values. Therefore, for each engine run, the test blend properties were included to measure the amount of fuel injected. Brake thermal efficiency (BTE) was calculated from the measured data. *2.3. Engine Test Setup* The engine test consisted of a single-cylinder diesel engine, and a Coda gas analyzer used to measure emissions (Figure 3). Table 3 contains the engine specifications. The engine equipment and the Coda gas analyzer's accuracy ranges were used as described in previous work [16,25]. Specific fuel consumption (SFC) for each fuel test was measured using a flow rate meter because of the differences in density and heating values. Therefore, for each engine run, the test blend properties were included to measure the amount of fuel injected. Brake thermal efficiency (BTE) was calculated from the measured data.

**Figure 3.** Schematic diagram of test set-up. **Figure 3.** Schematic diagram of test set-up.



1400, 2000 and 2600 RPM

Engine speeds test @ full load

**3. Results and Discussion** 

*3.1. Spray Characteristics*

Compression ratio test 19:1

atomization.

#### **3. Results and Discussion** *Energies* **2020**, *13*, x FOR PEER REVIEW 5 of 12

#### *3.1. Spray Characteristics Energies* **2020**, *13*, x FOR PEER REVIEW 5 of 12 Spray images of the test fuels are displayed in Figure 4. These images are a sample of three

Spray images of the test fuels are displayed in Figure 4. These images are a sample of three images recorded in the test. The results are presented in Figures 5 and 6. Spray tip penetration (S) and spray cone angle ( Spray images of the test fuels are displayed in Figure 4. These images are a sample of three images recorded in the test. The results are presented in Figures 5 and 6. Spray tip penetration (S) and spray cone angle (ѳ) were averaged from three images and six plumes. ) were averaged from three images and six plumes. *Energies* **2020**, *13*, x FOR PEER REVIEW 5 of 12 Spray images of the test fuels are displayed in Figure 4. These images are a sample of three images recorded in the test. The results are presented in Figures 5 and 6. Spray tip penetration (S) images recorded in the test. The results are presented in Figures 5 and 6. Spray tip penetration (S) and spray cone angle (ѳ) were averaged from three images and six plumes.

and spray cone angle (ѳ) were averaged from three images and six plumes.


**Figure 4.** Spray comparison images of test blends. **Figure 4.** Spray comparison images of test blends. **Figure 4.** Spray comparison images of test blends.

**Figure 4.** Spray comparison images of test blends. Spray tip penetration of test fuel blends under different injection conditions is presented in Figure 5 Spray penetration was improved for all ABE test blends because of the low surface tension and viscosity of the ABE blends. The spray penetration of ABE-Bd and ABE-D blends increased, respectively, by 3–5% and 4–5% compared to neat biodiesel and diesel. The low viscosity and boiling point of alcohol can result in improved atomization and evaporation behavior of diesel and biodiesel [26,27]. Therefore, the reaction rate increased. Compared to other test fuels, the spray penetration of Spray tip penetration of test fuel blends under different injection conditions is presented in Figure 5 Spray penetration was improved for all ABE test blends because of the low surface tension and viscosity of the ABE blends. The spray penetration of ABE-Bd and ABE-D blends increased, respectively, by 3–5% and 4–5% compared to neat biodiesel and diesel. The low viscosity and boiling point of alcohol can result in improved atomization and evaporation behavior of diesel and biodiesel [26,27]. Therefore, the reaction rate increased. Compared to other test fuels, the spray penetration of biodiesel was clearly shown to be lower due to its high viscosity [26]. Engine power reduction and fuel consumption increments may have occurred because of lower penetration and poorer atomization. Spray tip penetration of test fuel blends under different injection conditions is presented in Figure 5 Spray penetration was improved for all ABE test blends because of the low surface tension and viscosity of the ABE blends. The spray penetration of ABE-Bd and ABE-D blends increased, respectively, by 3–5% and 4–5% compared to neat biodiesel and diesel. The low viscosity and boiling point of alcohol can result in improved atomization and evaporation behavior of diesel and biodiesel [26,27]. Therefore, the reaction rate increased. Compared to other test fuels, the spray penetration of biodiesel was clearly shown to be lower due to its high viscosity [26]. Engine power reduction and fuel consumption increments may have occurred because of lower penetration and poorer atomization. Spray tip penetration of test fuel blends under different injection conditions is presented in Figure 5 Spray penetration was improved for all ABE test blends because of the low surface tension and viscosity of the ABE blends. The spray penetration of ABE-Bd and ABE-D blends increased, respectively, by 3–5% and 4–5% compared to neat biodiesel and diesel. The low viscosity and boiling point of alcohol can result in improved atomization and evaporation behavior of diesel and biodiesel [26,27]. Therefore, the reaction rate increased. Compared to other test fuels, the spray penetration of biodiesel was clearly shown to be lower due to its high viscosity [26]. Engine power reduction and fuel consumption increments may have occurred because of lower penetration and poorer atomization.

**Figure 5.** Spray tip penetration of test blends. **Figure 5.** Spray tip penetration of test blends.

**Figure 5.** Spray tip penetration of test blends.

The spray cone angle of test blends is presented in Figure 6 under different times after start of injection (ASOI). The increase in injection pressure leads to a slight widening of the spray cone angle. However, the biodiesel fuel presented the maximum spray cone angle due to its high viscosity. In general, at 300 bar injection pressure, the spray cone angle of ABE-D/Bd blends increased at ASOI up to 0.75 ms, while the spray cone angle of neat biodiesel was higher at ASOI 0.5 ms. This result is consistent with the findings of previous work [16,26,27]. The spray cone angle of test blends is presented in Figure 6 under different times after start of injection (ASOI). The increase in injection pressure leads to a slight widening of the spray cone angle. However, the biodiesel fuel presented the maximum spray cone angle due to its high viscosity. In general, at 300 bar injection pressure, the spray cone angle of ABE-D/Bd blends increased at ASOI up to 0.75 ms, while the spray cone angle of neat biodiesel was higher at ASOI 0.5 ms. This result is consistent with the findings of previous work [16,26,27]. *Energies* **2020**, *13*, x FOR PEER REVIEW 6 of 12 injection (ASOI). The increase in injection pressure leads to a slight widening of the spray cone angle. general, at 300 bar injection pressure, the spray cone angle of ABE-D/Bd blends increased at ASOI up consistent with the findings of previous work [16,26,27].

*Energies* **2020**, *13*, x FOR PEER REVIEW 6 of 12

**Figure 6.** Spray cone angle of test blends. **Figure 6.** Spray cone angle of test blends. **Figure 6.** Spray cone angle of test blends.

#### *3.2. Engine Performance 3.2. Engine Performance*

3.2.2. BP and BTE

#### 3.2.1. Maximum in-Cylinder Pressure 3.2.1. Maximum in-Cylinder Pressure

Figure 7 presents the relationship between the maximum in-cylinder pressure trace of the test fuels. The ABE10D90 blend gave a maximum peak of in-cylinder pressure at 5 bars higher than neat diesel due to the high oxygen content and low cetane number (CN) of the ABE blend. This resulted in increased ignition time and rapid in-cylinder pressure. However, an increase in the engine speed resulted in reduced in-cylinder pressure by about 10 bars. In-cylinder pressure was improved with the addition of ABE to the biodiesel blend. Spray and combustion characteristics enhanced the ABEbiodiesel blend due to a reduction in biodiesel viscosity. Figure 7 presents the relationship between the maximum in-cylinder pressure trace of the test fuels. The ABE10D90 blend gave a maximum peak of in-cylinder pressure at 5 bars higher than neat diesel due to the high oxygen content and low cetane number (CN) of the ABE blend. This resulted in increased ignition time and rapid in-cylinder pressure. However, an increase in the engine speed resulted in reduced in-cylinder pressure by about 10 bars. In-cylinder pressure was improved with the addition of ABE to the biodiesel blend. Spray and combustion characteristics enhanced the ABE-biodiesel blend due to a reduction in biodiesel viscosity. 3.2.1. Maximum in-Cylinder Pressure of fuels. The ABE10D90 blend gave a maximum peak of in-cylinder pressure at 5 bars higher than neat in increased ignition time and rapid in-cylinder pressure. However, an increase in the engine speed the addition of ABE to the biodiesel blend. Spray and combustion characteristics enhanced the ABE-

**Figure 7.** Maximum in-cylinder pressure test fuels. **Figure 7.** Maximum in-cylinder pressure test fuels. 3.2.2. BP and BTE **Figure 7.** Maximum in-cylinder pressure test fuels.

#### 3.2.2. BP and BTE *Energies* **2020**, *13*, x FOR PEER REVIEW 7 of 12 *Energies* **2020**, *13*, x FOR PEER REVIEW 7 of 12

The engine was connected to an electrical dynamometer, which was used to measure engine brake power output (BP) at various engine speeds. Brake thermal efficiency (BTE) is the ratio between the brake powers of the engine and the fuel energy supplied to the engine. The engine was connected to an electrical dynamometer, which was used to measure engine brake power output (BP) at various engine speeds. Brake thermal efficiency (BTE) is the ratio between the brake powers of the engine and the fuel energy supplied to the engine. The engine was connected to an electrical dynamometer, which was used to measure engine brake power output (BP) at various engine speeds. Brake thermal efficiency (BTE) is the ratio between

Figure 8 shows the variation of BP with engine speed according to test fuel. Both the neat biodiesel (Bd) and ABE-Bd blend revealed a lower value of BP due to low heating values (Table 1 and Figure 1). ABE-D/Bd blends had a higher combustion efficiency because of their high oxygen content, which improved the combustion rate when used as an additive blend. Algayyim et al. [6] investigated the effect of BA-diesel blends in a diesel engine. The experimental results showed that BTE increased because of the addition of BA to the diesel blend. These increments in BTE were achieved because of increased oxygen content in the blend (Figure 9). Oxygen helped to improve combustion efficiency, particularly during the diffusion combustion phase. Another factor influencing the BTE was the cetane number. ABE-diesel/biodiesel fuel blends have a lower cetane number than diesel and biodiesel, causing longer ignition delay, and a wider range in the fraction of fuel burned in the premixed mode, which elevates BTE [23–25]. Figure 8 shows the variation of BP with engine speed according to test fuel. Both the neat biodiesel (Bd) and ABE-Bd blend revealed a lower value of BP due to low heating values (Table 1 and Figure 1). ABE-D/Bd blends had a higher combustion efficiency because of their high oxygen content, which improved the combustion rate when used as an additive blend. Algayyim et al. [6] investigated the effect of BA-diesel blends in a diesel engine. The experimental results showed that BTE increased because of the addition of BA to the diesel blend. These increments in BTE were achieved because of increased oxygen content in the blend (Figure 9). Oxygen helped to improve combustion efficiency, particularly during the diffusion combustion phase. Another factor influencing the BTE was the cetane number. ABE-diesel/biodiesel fuel blends have a lower cetane number than diesel and biodiesel, causing longer ignition delay, and a wider range in the fraction of fuel burned in the premixed mode, which elevates BTE [23–25]. the brake powers of the engine and the fuel energy supplied to the engine. Figure 8 shows the variation of BP with engine speed according to test fuel. Both the neat biodiesel (Bd) and ABE-Bd blend revealed a lower value of BP due to low heating values (Table 1 and Figure 1). ABE-D/Bd blends had a higher combustion efficiency because of their high oxygen content, which improved the combustion rate when used as an additive blend. Algayyim et al. [6] investigated the effect of BA-diesel blends in a diesel engine. The experimental results showed that BTE increased because of the addition of BA to the diesel blend. These increments in BTE were achieved because of increased oxygen content in the blend (Figure 9). Oxygen helped to improve combustion efficiency, particularly during the diffusion combustion phase. Another factor influencing the BTE was the cetane number. ABE-diesel/biodiesel fuel blends have a lower cetane number than diesel and biodiesel, causing longer ignition delay, and a wider range in the fraction of fuel burned in the premixed mode, which elevates BTE [23–25].

**Figure 8.** Brake power (BP) of test fuels. **Figure 8.** Brake power (BP) of test fuels. **Figure 8.** Brake power (BP) of test fuels.

**Figure 9.** Thermal efficiency of test fuels. **Figure 9.** Thermal efficiency of test fuels. **Figure 9.** Thermal efficiency of test fuels.

3.2.3. EGT and NOx Formation

3.2.3. EGT and NOx Formation

conventional diesel (Figure 11) due to the high combustion temperature associated with biodiesel

Figures 10 and 11 show the relationship between the EGT and NOx emissions of the test fuels at

Figures 10 and 11 show the relationship between the EGT and NOx emissions of the test fuels at

## 3.2.3. EGT and NO<sup>x</sup> Formation

Figures 10 and 11 show the relationship between the EGT and NO<sup>x</sup> emissions of the test fuels at various engine speeds. The ABE-Bd blend showed a significant reduction in EGT compared to neat biodiesel at all engine speeds. Neat biodiesel showed higher NO<sup>x</sup> emissions compared to conventional diesel (Figure 11) due to the high combustion temperature associated with biodiesel reaction [28–30]. Adding a blend such as acetone, butanol and ethanol creates lower boiling points, which result in an increased evaporation rate and combustion efficiency. Therefore, EGT will be reduced. EGT and NO<sup>x</sup> emissions were reduced with the ABE blend to neat diesel and biodiesel by 14–17% and 11–13%, respectively. This result agrees with other work [30–35]. *Energies* **2020**, *13*, x FOR PEER REVIEW 8 of 12 reaction [28–30]. Adding a blend such as acetone, butanol and ethanol creates lower boiling points, which result in an increased evaporation rate and combustion efficiency. Therefore, EGT will be reduced. EGT and NO*x* emissions were reduced with the ABE blend to neat diesel and biodiesel by 14–17% and 11–13%, respectively. This result agrees with other work [30–35]. *Energies* **2020**, *13*, x FOR PEER REVIEW 8 of 12 reaction [28–30]. Adding a blend such as acetone, butanol and ethanol creates lower boiling points, which result in an increased evaporation rate and combustion efficiency. Therefore, EGT will be reduced. EGT and NO*x* emissions were reduced with the ABE blend to neat diesel and biodiesel by 14–17% and 11–13%, respectively. This result agrees with other work [30–35].

**Figure 10.** Exhaust gas temperature (EGT) of test fuels. **Figure 10.** Exhaust gas temperature (EGT) of test fuels. **Figure 10.** Exhaust gas temperature (EGT) of test fuels.

**Figure 11.** NOx emissions of test fuels. **Figure 11.** NOx emissions of test fuels. **Figure 11.** NOx emissions of test fuels.

#### 3.2.4. UHC, CO and CO2 Emissions 3.2.4. UHC, CO and CO2 Emissions 3.2.4. UHC, CO and CO<sup>2</sup> Emissions

The use of the ABE-D blend decreased UHC emissions compared to neat diesel at medium and high engine speeds (Figure 12). This reduction occurred because the boiling point of the ABE blend was low, which improved the vaporization rate and promoted combustion performance. The difference in droplet lifetime between ABE (3.25 s/mm2) and diesel (3.75 s/mm2) affected the reaction time of ABE blends at 823 K, which resulted in increased mixing time and led to complete reaction resulting in decreased UHC emissions [7]. ABE-biodiesel blends increased the UHC emissions due The use of the ABE-D blend decreased UHC emissions compared to neat diesel at medium and high engine speeds (Figure 12). This reduction occurred because the boiling point of the ABE blend was low, which improved the vaporization rate and promoted combustion performance. The difference in droplet lifetime between ABE (3.25 s/mm2) and diesel (3.75 s/mm2) affected the reaction time of ABE blends at 823 K, which resulted in increased mixing time and led to complete reaction resulting in decreased UHC emissions [7]. ABE-biodiesel blends increased the UHC emissions due The use of the ABE-D blend decreased UHC emissions compared to neat diesel at medium and high engine speeds (Figure 12). This reduction occurred because the boiling point of the ABE blend was low, which improved the vaporization rate and promoted combustion performance. The difference in droplet lifetime between ABE (3.25 s/mm<sup>2</sup> ) and diesel (3.75 s/mm<sup>2</sup> ) affected the reaction time of ABE blends at 823 K, which resulted in increased mixing time and led to complete reaction resulting in decreased UHC emissions [7]. ABE-biodiesel blends increased the UHC emissions due to the high

emissions were increased by 25–34% when ABE was added to biodiesel blends. All neat biodiesel and ABE-biodiesel/diesel blends presented lower CO emissions at all engine speeds due to the high oxygen content (Figure 13). This result agrees with previous work [34–37]. Moreover, the oxygen content of ABE-Bd blend was higher than biodiesel resulting in a significant reduction in CO

[36,37]. UHC emissions were reduced by 13% when ABE was added to diesel. However, UHC emissions were increased by 25–34% when ABE was added to biodiesel blends. All neat biodiesel and ABE-biodiesel/diesel blends presented lower CO emissions at all engine speeds due to the high oxygen content (Figure 13). This result agrees with previous work [34–37]. Moreover, the oxygen content of ABE-Bd blend was higher than biodiesel resulting in a significant reduction in CO

to the high oxygen atom (biodiesel and ABE), which altered the electronic structure. Almost all the

to the high oxygen atom (biodiesel and ABE), which altered the electronic structure. Almost all the

oxygen atom (biodiesel and ABE), which altered the electronic structure. Almost all the C-H bonds of the ABE-Bd blend are less active for reaction compared to hydrocarbon fuels (fossil fuels), which means more time is required to complete the reaction and increases UHC emissions [36,37]. UHC emissions were reduced by 13% when ABE was added to diesel. However, UHC emissions were increased by 25–34% when ABE was added to biodiesel blends. All neat biodiesel and ABE-biodiesel/diesel blends presented lower CO emissions at all engine speeds due to the high oxygen content (Figure 13). This result agrees with previous work [34–37]. Moreover, the oxygen content of ABE-Bd blend was higher than biodiesel resulting in a significant reduction in CO emissions compared to neat biodiesel and ABE-diesel blends. CO<sup>2</sup> emissions were slightly increased of ABE10Bd90 blends (Figure 14). It is clear that ABE0-D and ABE-Bd blends produced less CO emissions compared to diesel due to the lower carbon-to-hydrogen ratio [31,33,37]. *Energies* **2020**, *13*, x FOR PEER REVIEW 9 of 12 emissions compared to neat biodiesel and ABE-diesel blends. CO*2* emissions were slightly increased of ABE10Bd90 blends (Figure 14). It is clear that ABE0-D and ABE-Bd blends produced less CO emissions compared to diesel due to the lower carbon-to-hydrogen ratio [31,33,37]. *Energies* **2020**, *13*, x FOR PEER REVIEW 9 of 12 emissions compared to neat biodiesel and ABE-diesel blends. CO*2* emissions were slightly increased of ABE10Bd90 blends (Figure 14). It is clear that ABE0-D and ABE-Bd blends produced less CO emissions compared to diesel due to the lower carbon-to-hydrogen ratio [31,33,37].

**Figure 12.** Unburnt hydrocarbon (UHC) emissions of test fuels. **Figure 12.** Unburnt hydrocarbon (UHC) emissions of test fuels. **Figure 12.** Unburnt hydrocarbon (UHC) emissions of test fuels.

**Figure 13.** CO emissions of test fuels. **Figure 13.** CO emissions of test fuels. **Figure 13.** CO emissions of test fuels.

**Figure 14.** CO2 emissions of test fuels.

**Figure 14.** CO2 emissions of test fuels.

**Figure 13.** CO emissions of test fuels.

emissions compared to neat biodiesel and ABE-diesel blends. CO*2* emissions were slightly increased of ABE10Bd90 blends (Figure 14). It is clear that ABE0-D and ABE-Bd blends produced less CO

**Figure 12.** Unburnt hydrocarbon (UHC) emissions of test fuels.

emissions compared to diesel due to the lower carbon-to-hydrogen ratio [31,33,37].

**Figure 14.** CO2 emissions of test fuels. **Figure 14.** CO<sup>2</sup> emissions of test fuels.

### **4. Conclusions**

ABE mixture can increase the spray penetration of neat diesel and biodiesel resulting in improved atomization. The ABE10D90 blend gives a maximum in-cylinder pressure value 5 bars higher than neat diesel. The Bd and ABE-Bd blends gave a lower value of BP due to the low heating values. EGT, NO<sup>x</sup> and CO emissions were significantly reduced with the addition of the ABE blend to both neat diesel and biodiesel. UHC emissions were reduced when ABE was added to diesel. However, UHC emissions were increased when ABE was added to biodiesel. Thus, it can be concluded that the ABE mixture could be a good additive blend for neat diesel rather than neat biodiesel for improving diesel properties by using green energy for CI engines with no or minor modifications.

**Author Contributions:** S.J.M.A. and A.P.W. conceived and designed the experiments; S.J.M.A. performed the experiments; S.J.M.A. and A.P.W. analyzed the data. S.J.M.A. wrote the paper; A.P.W. provided editorial contribution and guidance. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The first author thanks the Iraqi Governments' Ministry of Higher Education and Scientific Research for supporting the research.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

**Michał Paduchowicz \* and Artur Górski**

Department of Machine and Vehicle Design and Research, Faculty of Mechanical Engineering, Wroclaw University of Science and Technology, 27 Wybrzeze Wyspia ´nskiego, 50-370 Wrocław, Poland; ˙ artur.gorski@pwr.edu.pl

**\*** Correspondence: michal.paduchowicz@pwr.edu.pl; Tel.: +48-71-320-39-08

**Abstract:** The article presents the results obtained during the inspection of the load-bearing structure of a power unit that suffered from fire. The inspection, consisting in the assessment of both the structure's technical condition and durability of welded joints, was performed on seven height levels of the power unit. The vibration spectrum of the unit's steel structure was analyzed, and frequency characteristics were, thus, obtained for individual measurement levels. Thermal vision measurements were also performed in the unit's all connection points to check for possible unsealing of some elements in the boiler island of the inspected power unit. The next stage consisted of performing strength calculations of the steel structure with a goal to estimate the structure's stress state. The conclusions contain suggestions for modernization of welded joints in order to maintain the power unit's design strength.

**Keywords:** power boilers; the load-bearing structures; damage assessment; fires; tanks

**Citation:** Paduchowicz, M.; Górski, A. Identification of the Effects of Fire-Wave Propagation through the Power Unit's Boiler Island. *Energies* **2021**, *14*, 1231. https://doi.org/ 10.3390/en14051231

Received: 2 December 2020 Accepted: 10 February 2021 Published: 24 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

## **1. Introduction**

The article addresses the problem presented in many other publications on the spread of fires in industrial installations [1]. As the results of the article [2] show, the main causes of fires in such places as coal-fired power plants, where hard coal are used, are explosive works, self-heating as well as friction and impact. In this study, an attempt was made to verify the causes of the fire and to verify the effects of the passage of the fire wave through the area of power units covered with brown coal together with biomass. The thesis was put forward that the combustion of these two fuels is dangerous at the same time due to the self-heating of biomass, which may lead to its spontaneous ignition. This, in turn, taking into account the ease and speed of coal dust (brown coal is crushed into a dust fraction and dried at the same time before being fed to the combustion chamber), may affect the intensity of combustion, which in turn may cause an explosion at high concentration of this element. If we take into account the fact that in the vicinity of fluidized bed power units, there are large amounts of coal dust (among others on communication elements such as platforms, barriers), this will largely help to spread the fire wave within the boiler installations.

Within a power unit's boiler island, a passing fire wave affected the unit's second pass (convection chamber), the cyclone, ash removal system, air-supply system, etc. The point of origin and the path of the fire wave are shown in Figure 1 [3]. References should be numbered in order of appearance and indicated by a numeral or numerals in square brackets, e.g., see the end of the document for further details on references [4,5].

The fire was initiated in the conveyor belt gallery situated in the coal and biomass supply line of power unit A. The fire wave later traveled through communication lines located between the power units and approached further generating units. The propagation of the fire wave necessitated evaluation of the degradation and damage caused to particular elements of the power units. The tests and measurements in this case were complex and consisted of [6]:


**Figure 1.** The fire's point of origin and way through the power units.

Determination of the actual state of effort of structural elements subjected to thermal loads is crucial in the assessment of the technical condition of power unit elements. The literature presents many practical methods; however, in the case of such a complex structure as a power unit with the entire installation of water, steam, coal, lime and air, its own methodology was proposed, consisting in the integration of numerical and experimental methods. The main assumption of the methodology is the dependence of the value of reduced stresses on temperature (1).

$$
\sigma\_i^{(T)} = -\frac{E\beta (T - T\_0)}{1 - 2\nu} \tag{1}
$$

where the temperature field is described (2):

$$\frac{\partial}{\partial \mathbf{x}\_i} \left[ \lambda\_{ij} \frac{\partial T}{\partial X\_j} \right] + q\_V = \rho c \frac{\partial T}{\partial t} \tag{2}$$


The research on the causes of the fire was based on simultaneous thermal and structural analyses, which are presented in the diagram below (Figure 2).

**Figure 2.** Scheme of tests carried out on the power unit where the fire occurred.

The research scheme presented in the figure above refers to the numerical analysis performed using the finite element method. First, the thermal analysis of the power device model is carried out, and then, using the obtained spatial temperature field and mechanical loads, the complex state of stress of this structure is determined.

#### **2. Operating Parameters of Fluidized Bed Power Unit**

For the past few years, electric energy producers have been obliged by EU legal regulations to cofire biomass with coal in conventional power units. This requirement compromises fire safety within the area of energy facilities in operation. One of the greatest threats is coal autoignition resulting from its coming into contact with biomass of typically high temperature. The above case was the hypothetical reason behind the ignition of fire in one of the biomass supply galleries in the analyzed power unit [8].

Apart from that, the use of biomass as one of the ingredients included in the fuel supplied to the boilers in conventional power units greatly influences operating conditions of the whole boiler system. The most frequent problem consists of lowered thermal efficiency, as biomass is less energetic than the fuel originally intended in the power unit's design. This is mainly due to the creation of residue on heat exchange surfaces and the start of corrosion and erosion processes on those elements. Fuel with biomass admixtures is usually burnt in fluidized bed boilers. Figure 3 shows a schematic diagram of such a type of power unit [9].

**Figure 3.** A scheme of fluidized bed boiler unit.

Table 1, on the other hand, shows technical parameters of a fluidized bed boiler unit.

**Table 1.** Operating parameters of CFB-670 boiler.


#### **3. Non-Destructive Tests**

The steel structure of the power units that experienced the fire was inspected for damage. The inspections included inter alia welded joints of the unit's main components, i.e., load-bearing structure, casing of both furnace and convection chambers, cyclone, ash removal system, air ducts, and auxiliary elements. The objective was to find potential material discontinuities (fractures) and to evaluate degradation level of selected areas on boiler elements against the required standards. Where necessary (and where possible), non-destructive tests were performed by means of magnetic particle inspection with permanent magnet magnetization. In order to detect magnetic stray field, during the magnetization process, the inspected element was sprayed with magnetic particles. In order to obtain clearer MPI magnetograms, the inspected areas were degreased and painted with undercoater. The first stage consisted of a visual inspection of those external surfaces that had close contact with the fire wave. Sample results are shown in Figures 4 and 5. Special scrutiny was given to welded joint and bolted joint areas in the main members of the boiler's support structure [10].

**Figure 4.** Diagram of the power unit showing the place and labels for inconsistencies found.

**Figure 5.** Weld cracks in the girders of the unit's load-bearing structure: (**a**) weld discontinuity between the face and the heat-affected zone (HAZ); (**b**) weld crack; (**c**) subsurface crack of the girder's weld; (**d**) weld crack at the connection point between load-bearing girders.

The inspected areas of the power unit were shown in Figure 4 and marked with schematic labels.

Examination of the power unit's load-bearing structure was mainly focused on the strength of load-bearing members, their connections, and joints. Inspections of load-bearing structure were performed on seven height levels (6, 11, 12, 32, 41, 47, and 52 m) indicated in Figure 4. The inspections covered welded joints and weld quality, as well as bolted joints and plastic strain of both beams and casing. The obtained results allowed to make the following observations [11]:


In order to estimate the effects of fire wave propagation from coal and biomass feeders to the boiler, control was needed of the degradation level of the boiler's loadbearing structure and of all machines within the analyzed power unit. To meet that goal, inspections of the boiler's systems and steel structure were performed. Inspection of the steel structure revealed faults that took the form of cracks in base material and between the weld and the heat affected zone, as shown in Figure 5. Much part of the structure's damage was caused by high temperature, which significantly lowered the strength of load-carrying elements to standard loads. That is when loss of stability took place. The inspection also revealed defects of the structure, caused by incorrectly performed assembly procedures and resulting in the low quality of welded joints, as well as in forbidden notches, cuts, and mount holes or process holes. The above-listed faults contributed to significantly decreased strength of the discussed elements. Additionally, geometrical notches cut in the structure (structural notches) significantly contributed to lowered durability of those elements. Apart from the above, the inspection also allowed to observe numerous signs of repairs, which had been performed during operation of the power unit and were never protected against corrosion, resulting in the appearance of numerous corrosion centers. During inspection, attention was also paid to inappropriate bracings, which may cause stress concentration in the areas of their construction. Incorrectly mounted, additional metal sheets create a structural notch, which means rapid change in stiffness, resulting in the interruption of proper load distribution. In addition, a great number of welded joints is accumulated on a small geometric area, which leads to the change in the material's structure due to great amount of heat supplied during welding. This causes the structural material to lose its strength [12,13].

#### **4. Measurement of Vibration Levels in the Load-Bearing Structure of the Power Units**

The planned scope of research also included measurements of vibrations that occur in the combustion chamber's load-bearing (support) structure, on its individual height levels, between 5 and 50 m. Figure 6 shows a general diagram of measurement points arrangement for the 32-m level [14].

Measurements of vibration levels were performed in two perpendicular directions on the load-carrying posts of the power unit's separate height levels. The diagram of separate measurement points, as well as the directions of the analyzed displacement and acceleration spectra are shown in Figure 7. The measurements allowed to obtain results in the form of vibration frequency characteristics for individual measurement points. Sample measurement results are shown in the following figures. Values of maximum vibration amplitudes (A) for the measured time range are given below the illustrations [14].

**Figure 6.** Diagram of measurement points on units 1, 2, and 3—the 32-m level.

**Figure 7.** Diagrams of accelerations and displacements in a sample point 2, where vibration amplitudes for the measuring range are inter alia (**a**) 0.815 ÷ 0.191 mmRMS, (**b**) −1.146 ÷ 1.961 mm.

#### **5. Measurements of Temperature Distribution on Particular Elements of the Power Unit**

The measurements covered heat loads of the whole power unit, including its loadbearing structure and other selected process equipment of the boiler island, and were performed using a thermal imaging camera with adequate measuring range and sensitivity. The temperature distribution, thus, obtained was used in further analysis, as one of the boundary conditions for structural strength calculations based on finite element method (FEM). Sample results of the measurements performed for the boiler's furnace chamber and auxiliary equipment are shown in Figures 8 and 9 [15].

**Figure 8.** Temperature distribution at the soot blower entrance point into the furnace chamber.

**Figure 9.** View of temperature distribution on the external casing of flue gas cyclone—clearly visible thermal bridges in the cyclone's upper part.

Thermal imaging inspection of temperature distribution was performed for the power unit's height levels in the range between 0 and 56 m. In addition, the obtained thermal images allowed us to estimate the quality of thermal insulation used in the furnace chamber's external casing and other boiler equipment. This allowed us to find areas with damaged thermal insulation and the resulting thermal bridges, especially those located on the connecting points of the machines working with furnace chamber and other elements. The damage of external insulation layers was closely connected to the propagation of the fire wave in the vicinity of the power units [16].

## **6. Strength Calculations for the Load-Bearing Structure of the Power Unit and Its Elements**

The measured temperature distribution was used to perform FEM strength calculations for some elements of the power units' parts, such as their load-bearing structure, cyclone, and fuel hopper. The above-listed elements were particularly subjected to heat loads resulting from the fire that occurred in the boiler hall, as they do not have insulated

external envelope. The calculations were performed using the finite element method in the NX-IDEAS system [17]. For the sake of the calculations, geometric models of the abovementioned elements were prepared (Figure 10), which were later digitized using appropriate finite elements. The next step consisted in strength calculations, with consideration paid to local temperature changes on external surfaces of the elements [18].

**Figure 10.** Load-bearing structure of the power unit—general view.

The calculations included both thermal and mechanical loads simultaneously. These are both gradients of temperature, which occur in steel structure due to its uneven heating resulting from the fire, and of structural loads, which occur due to the weight of subassemblies hung on the boiler island's steel structure. However, before the structure's effort was calculated, the spatial thermal field was found for the power unit's support structure. This kind of distribution is highly likely to occur in real structures. It is shown in Figure 11. The results of strength calculations obtained in the form of general displacement contour and Huber–Mises reduced stress contour are shown in Figure 12 [19].

**Figure 11.** Thermal field contour lines (◦C) in the whole load-bearing structure of the power unit.

**Figure 12.** Huber–Mises reduced stress contour lines for the load-bearing structure of the power unit [MPa]—view of the upper grate area.

FEM strength calculations of the load-bearing structure showed among other things that greatest reduced stress equal to 183 MPa could be observed in the area of the posts supporting return lines to the furnace chamber. These values are lower than the yield strength (235 MPa) of the construction material used to build the load-bearing structure of the power units. In the remaining parts of the units' load-bearing structures, the majority of reduced stresses were approximately 50 MPa. The above data prove that the fire in the boiler hall did not cause any damage to the support structure of the power unit. FEM strength calculations were also performed for the container, by building both geometric (Figure 13) and discrete models. The container is supported in four points located in its upper part, as can be seen from the model shown in Figure 13. The colors in the model correspond to metal sheets of different thicknesses [20].

**Figure 13.** View of the container's complete geometric model.

Digitization was performed using Thin Shell elements whose thickness corresponded to the thickness of metal sheets used to construct the container. The calculations included loads resulting from its normative service load and thermal gradients resulting from local exposure to heat due to the fire wave. The container was made of steel S235, which is why the following strength criteria were assumed:

$$
\sigma\_{\text{red}} < \sigma\_{\text{top}} = R\_{\text{\textdegree}} = \text{230 MPa} \tag{3}
$$

where:

*σred*—reduced stress according to Huber and Mises [MPa]; *σdop*—allowable stress assumed in the analysis [MPa]; *Re*—yield strength of the container structure's material [MPa].

In order to determine the reasons for the container's damage, a series of strength analyses was performed, with consideration paid to several factors that accompanied the fire and could cause a crack in the area of dumping hoppers and complete damage of the container's upper part. The rising temperature inside the boiler hall, which was the result of the fire, led to uneven heating of the boiler's surface. The thermal gradient of the boiler's casing that occurred in consequence, as well as the rise in air temperature inside the container—which translated into the rise in air pressure—both contributed to the emergence of cracks in the area of weld lines on the fuel hopper and sudden damage of its upper part that followed. Strength analysis was performed for the following air pressure values: 5 kPa, 10 kPa, 15 kPa, . . . , 50 kPa (values increased by 5 kPa). Impact strength of approximately 200 MPa was assumed for the observed low quality of weld lines on the fuel hoppers. The strength analysis showed that at air pressure inside the container equal to 50 kPa and at uneven thermal impact that occurred simultaneously, reduced stress in the area of hopper weld lines reached approximately 270 MPa (Figure 14). This contributed to the breaking of the discussed area of the container 8z.

**Figure 14.** Reduced stress contour lines in the lower part of the container, according to Huber and Mises hypothesis [MPa] (**a**) hopper area of the fluidized bed material tank; (**b**) the hopper weld area of the fluidized bed material reservoir where the crack has occurred

At the same time, reduced stress in the container's upper part was observed at approximately 870 MPa (Figure 15). The latter observation, along with the fact that the tests performed prove the ultimate strength of steel S235 to be at the level of 315 MPa, means that this area experienced a sudden discontinuity of the investigated object's casing [20].

The strength analysis also showed clearly that as the air pressure inside the container rose to approximately 50 kPa and the casing heated unevenly, the casing broke in its both upper and lower parts.

**Figure 15.** Reduced stress contour lines in the upper part of the container, according to Huber and Mises hypothesis [MPa]: (**a**) Calculation results of the upper cover of the bed material tank from the outside, (**b**) Calculation results of the upper cover of the bed material tank from the inside.

#### **7. Summary and Conclusions**

Measurement results allowed to estimate the effects of the fire wave that passed through a power station. The works were divided into several stages:


The passing fire wave caused greatest damage in front of the unit, at the coal and biomass supply galleries. Coal and biomass ignition in the space between the units generated soot, which combined with ash and water was used to extinguish the fire and produced compressed mass (Figure 14). The mass had the consistency of concrete and covered electric wires placed in conduits. The wires that were laid in conduits close to each other stopped the flowing mass, which set shortly. This phenomenon was observed only on the 6m level of the power unit [20].

**Author Contributions:** Conceptualization, M.P.; methodology, A.G.; software, M.P. validation, M.P.; formal analysis, A.G., M.P.; investigation, A.G.; resources, A.G.; data curation M.P.; writing—original draft preparation, A.G., M.P.; writing—review and editing, M.P.; visualization A.G., M.P.; supervision, A.G.; project administration, A.G.; funding acquisition, A.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to provisions in the contract with the company for which the research was carried out.

**Acknowledgments:** As part of the work carried out, no support was provided, either from external persons or institutions.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**


## **Investigation of Waste Biogas Flame Stability Under Oxygen or Hydrogen-Enriched Conditions**

## **Nerijus Striugas, Rolandas Paulauskas \*, Raminta Skvorˇcinskien ¯ e and Aurimas Lisauskas ˙**

Laboratory of Combustion Processes, Lithuanian Energy Institute, Breslaujos str. 3, LT-44403 Kaunas, Lithuania; Nerijus.Striugas@lei.lt (N.S.); Raminta.Skvorcinskiene@lei.lt (R.S.); Aurimas.Lisauskas@lei.lt (A.L.)

**\*** Correspondence: Rolandas.Paulauskas@lei.lt; Tel.: +370-37-401830

Received: 18 August 2020; Accepted: 9 September 2020; Published: 11 September 2020

**Abstract:** Increasing production rates of the biomethane lead to increased generation of waste biogases. These gases should be utilized on-site to avoid pollutant emissions to the atmosphere. This study presents a flexible swirl burner (~100 kW) with an adiabatic chamber capable of burning unstable composition waste biogases. The main combustion parameters and chemiluminescence emission spectrums were examined by burning waste biogases containing from 5 to 30 vol% of CH<sup>4</sup> in CO<sup>2</sup> under air, O2-enriched atmosphere, or with the addition of hydrogen. The tested burner ensured stable combustion of waste biogases with CH<sup>4</sup> content not less than 20 vol%. The addition of up to 5 vol% of H<sup>2</sup> expanded flammability limits, and stable combustion of the mixtures with CH<sup>4</sup> content of 15 vol% was achieved. The burner flexibility to work under O2-enriched air conditions showed more promising results, and the flammability limit was expanded up to 5 vol% of CH<sup>4</sup> in CO2. However, the combustion under O2-enriched conditions led to increased NOx emissions (up to 1100 ppm). Besides, based on chemiluminescence emission spectrums, a linear correlation between the spectral intensity ratio of OH\* and CH\* (IOH\*/ICH\*) and CH<sup>4</sup> content in CO<sup>2</sup> was presented, which predicts blow-off limits burning waste biogases under different H<sup>2</sup> or O<sup>2</sup> enrichments.

**Keywords:** swirl burner; waste biogas; hydrogen; oxygen; combustion; flame stability; blow-off limit

## **1. Introduction**

Due to concerns regarding global climate change, there has been a search for new ways to improve energy efficiency, reduce pollutant emission, and ensure sustainable management of resources. For these reasons, various alternative fuels and by-products (gas or waste fuel) with low calorific value are gaining more attention as an alternative energy source for energy production [1–3]. Basically, such fuels are gained in different fields with different compositions like steel production, refineries, landfills, biogas plants, gasification plants, and other industry sources. For example, a blast furnace gas or syngas consists of different concentrations of CO, H2, CO2, and/or N2, while biogas is mainly composed of CO<sup>2</sup> and CH4. Considering that the biogas and biomethane production is increasing over the years, rates of the waste biogas from biomethane production are also growing. These gases mainly consist of from few to 25 vol% of methane and should be utilized on-site to prevent pollutant emissions to the atmosphere. However, these gases are not well suitable to combust in existing natural gas-fired systems as the low concentrations of CH4, and the high amount of CO<sup>2</sup> could result in an unstable flame, blow-off, and a release of unburned gases into the environment. Mainly it caused by the high amount of CO2, which affects the flame temperature, thereby reducing the stability of the flame.

To avoid these problems and ensure complete and clean combustion of waste biogases, flame stability must be improved, and flammability limits expanded. It could be achieved using new concepts/modifications of burners [4–7]. For example, a new burner design based on modified geometry was proposed and tested for low calorific gas combustion by [8]. It was found that this burner ensures stable combustion of natural gas and syngas blends, but CO emissions exceed 200 mg/m<sup>3</sup> if thermal shares of syngas are below 70%. Another concept burner based on continuous air staging for low calorific gases was developed by a research group from Gaswärme-Institute. V. Essen (GWI) [9]. The burner (up to 200 kW) was tested on syngas, landfill gas, and mine gas. The obtained results showed that it is possible to achieve a stable combustion process with low NOx and CO emissions, but there is a need to preheat these gases up to 400 ◦C before supplying to the combustion zone. A similar study was performed by Mortberg et al. [10]. The authors preheated air up to 900 ◦C to enhance the low calorific value (LCV) gas combustion stability. During combustion tests, a cross-flow jet of LCV gas mixture (11.9 vol% CH<sup>4</sup> in N2) was introduced in the combustion airflow at oxygen-deficient conditions. The obtained results showed that this configuration using preheated combustion air is suitable to combust LCV gases, but results in the prolonged ignition delay, higher turbulence levels, and higher vorticity. Another work [11] showed that surface-stabilized combustion (SSC) technology burner is able to burn the LCV biogas with 20–65 vol% CO<sup>2</sup> in CH<sup>4</sup> or hydrogen-enriched CH<sup>4</sup> (15–100 vol% H<sup>2</sup> in CH4) with low pollutant emissions. However, the burner was more suitable for combustion of high calorific value gases. In the case of biogas with 65 vol% CO<sup>2</sup> in CH4, the flames were lifted, changing ϕ from 0.98 to 0.8, while further increase in ϕ led to blow-off. Another research [12] also focused on the combustion of LCV gases in burners with SiC and Al2O<sup>3</sup> porous structures. The authors were capable of combusting landfill gas with a methane content of 26 vol%. Though, at higher power (10 kW), the flammability limit decreased by 4 vol% of CH<sup>4</sup> in CO<sup>2</sup> due to increased flow speed. The authors also noticed that preheating the gas mixture, and the flammability limit could be extended by 2 vol%. Al-Attab et al. [13] investigated the combustion of producer gas from biomass gasification in a two-layer porous burner. Results showed that the burner is able to combust the producer gas with a lower calorific value of 5 MJ/m<sup>3</sup> changing equivalence ratios ϕ in the range of 0.33–0.71, but NOx emissions were in the range from 230 to 270 ppm. In order to increase the blow-off limit of LCV gases, Song et al. [14] investigated an improved preheating method of the gas mixture, based on an annular heat recirculation. The gas flow was preheated by a high-temperature wall near the gas inlet, which temperature increase, due to improved axial heat conduction and radiation heat transfer of porous media. It ensured stable combustion of the LCV gas of 1.4 MJ/m<sup>3</sup> . However, according to [15], higher volumetric capacity is needed to achieve higher power comparing to existing conventional burning systems. Besides, the use of porous media burners for the combustion of waste biogases with the varying composition is a bit complicated as a material for the porous media should be designed considering gases needed to burn; otherwise, a periodic replacement of the porous media could be needed.

Another option to improve flame stability and flammability is to use an addition of hydrogen/syngas or supply of oxygen-enriched air [16–22]. Chiu et al. [23] investigated the effect of H2/CO on a premixed methane flame. The research was performed in an impinging burner by changing an H2/CO concentration from 20/80 to 80/20 vol% and a methane concentration from 10 to 20 vol%. It was determined that the stable flame with H2/CO of 20/80 vol% is achieved at reach combustion conditions (ϕ ≥ 1.8) though the flame stability and flammability increased with increasing the hydrogen concentration in simulated syngas. It was also noted that with the increasing amount of H<sup>2</sup> in CO increases the flame temperature. A similar work focused on the hydrogen effect on the LCV gas combustion was performed by [24]. The authors performed numerical and experimental burner tests at a thermal power of 10 kW using syngas (~4.7 MJ/m<sup>3</sup> ) and blast-furnace gas (~3.7 MJ/m<sup>3</sup> ), including the addition of H<sup>2</sup> (26 and 52 vol%). It was determined that the burner properly burns the syngas, but the blast-furnace gas requires an additional supply of H<sup>2</sup> to avoid blow-off. Though, it was observed that the addition of H<sup>2</sup> negatively affects NOx emissions. To extend LCV gas combustion flammability and improve flame stability, this burner was also tested under oxy-fuel combustion conditions [25]. Results showed that NOx emissions decrease; however, to achieve a more stable flame and avoid overheating of the burner wall, a modification for the oxidizer distribution is needed. More detailed investigation of the oxygen enrichment was performed by Ylmaz et al. [26]. The authors investigated

the biogas flame stability and formation of emissions under different oxygen enrichment conditions in a pilot-scale model burner. It was pointed out that an increase of O<sup>2</sup> to 24% in the oxidizer leads to improved flame stability though CO emissions were equal to 372 ppm. A higher enrichment level (28 vol% of O2) caused decreased flame stability, but lower CO emissions, till 10 ppm. During the tests, NOx emissions increased with increased O<sup>2</sup> concentrations in the oxidizer. Ba et al. [20] tried to improve flame stability by coupling oxy-fuel combustion and fuel/air preheating. Firstly, this configuration was experimentally tested using a lab-scale tri-coaxial burner (25 kW), and later, the burner scaled-up to 180 kW was tested on a semi-industrial facility. In both cases, a stable flame of blast furnace gas was achieved with very low levels of CO and NOx emissions. However, this configuration was not tested on waste biogases, and the effect on flame stability is unknown.

Literature review reveals that there are many works proposing techniques to ensure stable combustion of syngas, furnace blast gases, natural gas blends with LCV gases and etc., but studies related to the combustion characteristics of waste biogas are still lacking. Taking into account that biogas production rates are increasing and to prevent waste biogas emissions to the atmosphere, these gases should be utilized safely on-site. The previous work [27] showed that the addition of up to 20 vol% hydrogen or oxygen-enriched air ensures stable combustion of waste biogases (CH<sup>4</sup> > 30% in CO2) with low emissions in a flat flame burner. Based on these findings, a flexible swirl burner (~100 kW) with an adiabatic chamber capable of burning unstable composition waste biogases was designed and developed. The developed combustion system provides a solution to recover primary energy and reduce pollutant emissions from biomethane production obtained waste biogases. This study presents the flexible combustion system in detail and performed tests in real conditions. During the experimental tests, flame stability and combustion limits were explored to ensure a wide range operation of the burner with different waste biogas containing from 5 to 30 vol% of CH<sup>4</sup> in CO<sup>2</sup> under air, oxygen-enriched atmosphere, or with hydrogen addition. For a better understanding of waste gas flame behavior under different conditions and to identify a burner performance via flame transition modes and blow-off limits, chemiluminescence emission spectrums (OH\* and CH\*) from flames of different waste biogas mixtures were obtained. During all tests, an online flue gas analysis was also performed to determine variations of NOx and CO emissions and compliance with emission standards.

#### **2. Materials and Methods**

#### *2.1. Experimental Setup*

The experiments were carried out in a 200 kWth semi-industrial scale test facility (Figure 1). In order to simulate waste biogases (LCV) with a low content of methane (CH4) in carbon dioxide (CO2) and its combustion with oxygen (O2) or hydrogen (H2) addition, a gas supply system was constructed (1–5). CO<sup>2</sup> was acquired in a liquid state (1) by a local supplier (JSC Gaschema), directed to a vaporizer (2) and supplied via the control unit to a gas-mixing system for preparation of LCV gases with pre-assigned compositions. Methane in the form of natural gas (NG) with the composition of 97.5 vol% CH4, 1.2 vol% C2H6, 0.27 vol% C3H<sup>8</sup> and C4H10, and 0.08 vol% CO<sup>2</sup> (NG composition is taken from the gas supplier's datasheet) was delivered from a distributed NG pipeline and supplied through a pressure regulating and metering station (3) to the gas-mixing system. For hydrogen or oxygen-enriched combustion, these gases were supplied from gas cylinders (4,5) to a novel swirl LCV gas burner (7). The required gas flow was ensured by a mass flow controller. The combustion air was supplied using an air blower (6).

*Energies* **2020**, *13*, x FOR PEER REVIEW 4 of 16

**Figure 1.** Scheme of the experimental burner testing setup: 1–A liquid CO2 tank; 2–a CO2 vaporizer; 3–a natural gas pressure regulating and metering station; 4–a hydrogen cylinder; 5–an O2 cylinder bundle with a manifold; 6–an air blower; 7–a novel low calorific value (LCV) gas burner; 8–an adiabatic combustion chamber; 9–a boiler; 10–optical fibers with collimating lens; 11–a spectrometer with an ICCD camera; 12–a data collecting system; 13-an operation panel; 14–a flue gas analysis system; 15–a flue gas blower. **Figure 1.** Scheme of the experimental burner testing setup: 1–A liquid CO<sup>2</sup> tank; 2–a CO<sup>2</sup> vaporizer; 3–a natural gas pressure regulating and metering station; 4–a hydrogen cylinder; 5–an O<sup>2</sup> cylinder bundle with a manifold; 6–an air blower; 7–a novel low calorific value (LCV) gas burner; 8–an adiabatic combustion chamber; 9–a boiler; 10–optical fibers with collimating lens; 11–a spectrometer with an ICCD camera; 12–a data collecting system; 13-an operation panel; 14–a flue gas analysis system; 15–a flue gas blower.

The combustion of the prepared LCV gas mixture was performed in an adiabatic chamber (8) with internal dimensions of 600 × 600 × 2500 mm. The temperatures in the chamber were measured with four K type thermocouples (T1–T4). The thermocouples were positioned at the 530, 900, 1690, and 2410 mm from the inner front wall, respectively. For spectral analysis of selected excited radicals from the flame, a spectrometer Andor Shamrock SR-303i coupled with an ICCD (Intensified Charge Coupled Device) camera Andor iStar DH734 (11) was used. Three optical fibers with collimating lenses (10) were installed to transmit the light to the spectrometer. The optical fibers were oriented horizontally to provide a radical distribution along with the flame and fitted at the distances of 170, 410, and 830 mm from the front wall. The combustion products are cooled down to appropriate temperature, which is measured with a K type thermocouple, by passing the boiler (9) mounted at the exit of the combustion chamber. A flow of cooling water is controlled, and water temperature is monitored before and after the boiler. The cooled flue gases are analyzed using a portable flue gas analyzer Testo 350XL (14) via a measurement point installed between the stack (15) and the boiler (9). The entire combustion process is automated and controlled from a control desk (13) installed outside the site. All operating signals from measuring equipment or sensors were collected and analyzed The combustion of the prepared LCV gas mixture was performed in an adiabatic chamber (8) with internal dimensions of 600 × 600 × 2500 mm. The temperatures in the chamber were measured with four K type thermocouples (T1–T4). The thermocouples were positioned at the 530, 900, 1690, and 2410 mm from the inner front wall, respectively. For spectral analysis of selected excited radicals from the flame, a spectrometer Andor Shamrock SR-303i coupled with an ICCD (Intensified Charge Coupled Device) camera Andor iStar DH734 (11) was used. Three optical fibers with collimating lenses (10) were installed to transmit the light to the spectrometer. The optical fibers were oriented horizontally to provide a radical distribution along with the flame and fitted at the distances of 170, 410, and 830 mm from the front wall. The combustion products are cooled down to appropriate temperature, which is measured with a K type thermocouple, by passing the boiler (9) mounted at the exit of the combustion chamber. A flow of cooling water is controlled, and water temperature is monitored before and after the boiler. The cooled flue gases are analyzed using a portable flue gas analyzer Testo 350XL (14) via a measurement point installed between the stack (15) and the boiler (9). The entire combustion process is automated and controlled from a control desk (13) installed outside the site. All operating signals from measuring equipment or sensors were collected and analyzed further.

#### further. *2.2. Waste Biogas Burner*

*2.2. Waste Biogas Burner*  A novel swirl burner was developed for the waste biogas (LCV) combustion (Figure 2). The burner is designed to operate in a flexible mode for a variety of waste biogas compositions, including the possibility of hydrogen or oxygen addition. The burner has multiple inlets for different gases and operational regimes. The ignition of the main fuel (LCV gas) is executed with a pilot natural gas flame. After the ignition of the main fuel, a supply of pilot gas is stopped. The LCV gas is supplied A novel swirl burner was developed for the waste biogas (LCV) combustion (Figure 2). The burner is designed to operate in a flexible mode for a variety of waste biogas compositions, including the possibility of hydrogen or oxygen addition. The burner has multiple inlets for different gases and operational regimes. The ignition of the main fuel (LCV gas) is executed with a pilot natural gas flame. After the ignition of the main fuel, a supply of pilot gas is stopped. The LCV gas is supplied via an annular channel installed in the center of the burner. Air was used as the main oxidizer, which enters

via an annular channel installed in the center of the burner. Air was used as the main oxidizer, which

the burner through a peripheral annulus. When a calorific value of waste biogas becomes very low, a stable operation is ensured by the oxygen or hydrogen addition. The supply of these gases is designed in the center of the burner. low, a stable operation is ensured by the oxygen or hydrogen addition. The supply of these gases is designed in the center of the burner.

*Energies* **2020**, *13*, x FOR PEER REVIEW 5 of 16

**Figure 2.** Schematic view of the experimental LCV gas burner:1–A burner embrasure; 2–a burner quarl; 3–an axial airflow swirler; 4–an axial LCV gas swirler; 5–a pilot gas lance; 6–an air annulus; 7– an LCV gas annulus; 8–an oxygen or hydrogen orifice; 9–a pilot gas annulus. **Figure 2.** Schematic view of the experimental LCV gas burner:1–A burner embrasure; 2–a burner quarl; 3–an axial airflow swirler; 4–an axial LCV gas swirler; 5–a pilot gas lance; 6–an air annulus; 7–an LCV gas annulus; 8–an oxygen or hydrogen orifice; 9–a pilot gas annulus.

The flame stabilization is achieved by swirling both, the combustion air and LCV gas flow with an axial vane swirl angle of 30°. Both swirlers has a swirl number of Sw = 0.5, which was defined by Equation (1) [28]: The flame stabilization is achieved by swirling both, the combustion air and LCV gas flow with an axial vane swirl angle of 30◦ . Both swirlers has a swirl number of S<sup>w</sup> = 0.5, which was defined by Equation (1) [28]:

$$\mathbf{S}\_{\rm W} = \frac{2}{3} \tan(\alpha\_{\rm sw}) \left( \frac{1 - \mathbf{R}^3}{1 - \mathbf{R}^2} \right) \tag{1}$$

where αsw is the axial vane swirl angle, R is the ratio of the center-body diameter and outer inlet tube where αsw is the axial vane swirl angle, R is the ratio of the center-body diameter and outer inlet tube diameter.

#### diameter. *2.3. Experimental Procedure*

*2.3. Experimental Procedure*  The stability limits of waste biogas (LCV) combustion were determined by running the combustor at a fixed heat load of 80 kWth. The heat load was calculated using the lower heating values (LHV) and measuring flows of combustible gases, namely, CH4, and H2 when it was used. Thus, the The stability limits of waste biogas (LCV) combustion were determined by running the combustor at a fixed heat load of 80 kWth. The heat load was calculated using the lower heating values (LHV) and measuring flows of combustible gases, namely, CH4, and H<sup>2</sup> when it was used. Thus, the main variables were CO<sup>2</sup> content in LCV gas, an amount of the oxidizer, and the fuel to air ratio (ϕ).

main variables were CO2 content in LCV gas, an amount of the oxidizer, and the fuel to air ratio (*φ*). Three sets of experiments were performed to determine the limits of stable waste biogas (LCV) combustion. The first trial was associated with a determination of the lowest possible, stable combustion regime by decreasing LHV of LCV gas, which was changed by increasing the volumetric fraction of CO2. The air was used as the oxidizer. At each LCV gas composition, the burner performance was tested for different fuel to air ratios, which varied in the range from 0.7 to 1.0. The Three sets of experiments were performed to determine the limits of stable waste biogas (LCV) combustion. The first trial was associated with a determination of the lowest possible, stable combustion regime by decreasing LHV of LCV gas, which was changed by increasing the volumetric fraction of CO2. The air was used as the oxidizer. At each LCV gas composition, the burner performance was tested for different fuel to air ratios, which varied in the range from 0.7 to 1.0. The fuel to air ratio was determined by measuring the O<sup>2</sup> concentration in flue gases.

fuel to air ratio was determined by measuring the O2 concentration in flue gases. The second round of experiments was conducted with the H2 addition to extend LCV gas combustion limits. In a previous work [29], the possibility to produce H2 on-site from LCV gas was explored. The idea was based on the power of syngas (P2SG) technology. Therein to produce H2 rich syngas, the thermal plasma was used to run a dry reforming of gases containing a low concentration of CH4 in CO2. Finally, it was concluded that from an economic point of view, this technology is cost competitive when a produced H2 content for 1 m3 LCV gas combustion will constitute up to 5 vol%. To simulate this situation in the present work, the hydrogen addition of 2.5 and 5 vol% was The second round of experiments was conducted with the H<sup>2</sup> addition to extend LCV gas combustion limits. In a previous work [29], the possibility to produce H<sup>2</sup> on-site from LCV gas was explored. The idea was based on the power of syngas (P2SG) technology. Therein to produce H<sup>2</sup> rich syngas, the thermal plasma was used to run a dry reforming of gases containing a low concentration of CH<sup>4</sup> in CO2. Finally, it was concluded that from an economic point of view, this technology is cost competitive when a produced H<sup>2</sup> content for 1 m<sup>3</sup> LCV gas combustion will constitute up to 5 vol%. To simulate this situation in the present work, the hydrogen addition of 2.5 and 5 vol% was investigated. As in the previous case, the air was used as the oxidizer with φ ranging from 0.7 to 1.0.

investigated. As in the previous case, the air was used as the oxidizer with ϕ ranging from 0.7 to 1.0. During the third set of experiments, the stability limits were determined using O2-enriched air instead of the H2 addition. The amount of oxygen was gradually increased until the flame becomes During the third set of experiments, the stability limits were determined using O2-enriched air instead of the H<sup>2</sup> addition. The amount of oxygen was gradually increased until the flame becomes stable to avoid possible problems described in some works [21,26] like increased CO emissions, high flame temperature, or even unstable combustion. When the flame became stable, the burner performance for different LCV gases against the various fuel to air ratios from 0.7 to 1.0 was tested. The expression of the volumetric oxygen content in this work corresponds to the O<sup>2</sup> concentration in air, which was defined as follows:

$$\text{O}\_2 \text{ in air} = \frac{\text{V}\_{\text{air}}\text{0.21} + \text{V}\_{\text{O}\_2}}{\text{V}\_{\text{air}} + \text{V}\_{\text{O}\_2}} \tag{2}$$

where Vair and VO<sup>2</sup> are the volume flow rates of air and O2, respectively. The upper value of O<sup>2</sup> in the air was pre-selected to not exceed 30 vol%.

During all experimental runs, emissions of CO and NOx in flue gases were measured continuously. CO is considered as the main indicator for incomplete combustion, thus despite the existence of flame stability, it demonstrates the inappropriate burner operation regime. NOx is one of the main regulated gaseous pollutants, which should be controlled. It is well known, that the main source of NOx formation is attributed to the thermal pathway. Thus, increasing oxygen content in the oxidizer during LCV gas combustion could lead to noticeable higher NOx emissions [3]. In order to know the level of this pollutant during combustion and compliance with emission standards (EU directive 2015/2193 [30]), it was measured and analyzed regarding the LCV combustion mode.

In addition, the chemiluminescence based optical diagnostic tool was used to identify the possible flame transition mode. The main chemiluminescent species of interest were OH\* and CH\* with a wavelength of 308.9 ± 2.0 nm and 431.4 ± 2.0 nm, respectively. According to [31–34], CH\* formation is related to high temperature, and the main formation pathways are:

$$\rm C\_2 + OH \rightarrow CH^\* + CO,\tag{3}$$

$$\rm C\_2H + O \rightarrow CO + CH^\* , \tag{4}$$

$$\rm C\_2H + O\_2 \rightarrow CO\_2 + CH^\* \tag{5}$$

The formation and excitation of the OH\* radical are also attributed to the thermal excitation, and the main formation reactions are [31–34]:

$$\rm H + O + M \to OH^\* + M,\tag{6}$$

$$\rm{CH} + \rm{O}\_2 \rightarrow \rm{OH}^\* + \rm{CO}\_2 \tag{7}$$

The optical data were collected from three-measurement points using a spectrometry system (Figure 1). The light coming from the flame is dispersed, and entire spectra is collected on the split area of the ICCD camera sensor. The camera exposure duration was set to 3 s, a side input slit to 100 µm, and grating to 300 L/mm. Each spectrum consisted of 10 acquisitions.

#### **3. Results and Discussion**

#### *3.1. Stability Maps of Waste Biogas (LCV) Combustion*

This section presents combustion stability maps for developed LCV gas burner under different tested combustion modes. Firstly, LCV gases with different concentrations of CH<sup>4</sup> and CO<sup>2</sup> were burned in the air over a wide range of fuel to air ratios to determine the limit of stable combustion under ordinary conditions. As can be seen from Figure 3, the stable combustion of the waste biogas (LCV) with CH<sup>4</sup> ≥ 25 vol% in CO<sup>2</sup> can be reached in the air under all tested range of φ. Compared to previous results [27], the combustion of waste biogas in a flat flame burner under the air atmosphere was only able with a mixture of 30 vol% CH<sup>4</sup> in CO2. In that case, the developed burner shows wider flammability ranges under air conditions. A reduced CH<sup>4</sup> content to 20 vol% in the waste biogas (LCV) leads to flame stability problems, and blow-off occurs at a richer combustion mode, φ ≥ 0.87. According to obtained results (Figure 3), the stable combustion of LCV gases with lower CH<sup>4</sup> content (>20 vol%) is only possible with H<sup>2</sup> or O<sup>2</sup> addition. The addition of H<sup>2</sup> to the biogas flame was studied in detail by others [16,35,36]. Zhen et al. [35] noted that the H<sup>2</sup> addition of 5 vol% improves biogas flame stability, and CO emissions decrease, but further addition increase to 10 vol% of H<sup>2</sup> is insignificant on flame stability. Similar findings were also reported by Leung and Wierzba [17]. The authors investigated the effect of 10 vol%, 20 vol%, and 30 vol% H<sup>2</sup> addition on the biogas flame stability and determined that the most significant results are achieved by supplying a small amount of H2. *Energies* **2020**, *13*, x FOR PEER REVIEW 7 of 16 in detail by others [16,35,36]. Zhen et al. [35] noted that the H2 addition of 5 vol% improves biogas flame stability, and CO emissions decrease, but further addition increase to 10 vol% of H2 is insignificant on flame stability. Similar findings were also reported by Leung and Wierzba [17]. The authors investigated the effect of 10 vol%, 20 vol%, and 30 vol% H2 addition on the biogas flame stability and determined that the most significant results are achieved by supplying a small amount of H2.

**Figure 3.** Stability maps of waste biogas (LCV) combustion with different content of CH4 in CO2 for a range of (**a**)H2 addition and (**b**) O2 air enrichment. **Figure 3.** Stability maps of waste biogas (LCV) combustion with different content of CH<sup>4</sup> in CO<sup>2</sup> for arange of (**a**) H<sup>2</sup> addition and (**b**) O<sup>2</sup> air enrichment.

Even though the concentration of CH4 in biogas is higher and varies from 40% to 60% compared to the studied range of CH4 concentrations (5–30 vol%), the provided findings are in good agreement with the obtained results. The stable combustion of LCV gas containing 15 vol% of CH4 is obtained within the range of *ϕ* from 0.7 to 0.85 by adding H2 of 2.5vol%, while an increased addition of H2 to 5 vol% has an insignificant effect, and the flammability limit is extended only to a richer region, up to *ϕ* of 0.95. However, further decrease of CH4 content in LCV gases results in flame blow-off even in both cases of H2 addition (Figure 3a). In order to extend the flammability limit and ensure stable combustion of LCV gases with very low CH4 content (5–15 vol%), O2-enriched air was introduced. Considering that a high oxygen enrichment level could lead to the unstable flame and increased CO emissions [26], two cases with 25 vol% and 30 vol% of O2 were tested (Figure 3b). As was expected, using O2-enriched air with O2 content of 25 vol%, the stable combustion of waste biogases with CH4 content from 30 to 5 vol% was ensured. However, in this case, the stable combustion of LCV gas containing 5 vol% of CH4 is achieved only in the narrow range of *ϕ*, from 0.79 to 0.86. Meanwhile, at a higher enrichment level (30 vol% of O2), the stable combustion of LCV gases is acquired in all tested points (Figure 3b). Similar results were also acquired numerically studying methane combustion under CO2/O2 atmosphere, and researchers pointed out that the most effective and stable combustion is achieved at O2 concentrations of 28–32 vol% [37]. Though another work [21] showed a bit higher level of O2 (31–35 vol%) is needed to achieve a stable flame. Considering the obtained results, it could be assumed that the designed burned with the adiabatic chamber ensure stable combustion of waste biogases (5–30 vol% of CH4 in CO2) at lower concentrations of O2 (25–30 vol%), due to a special design Even though the concentration of CH<sup>4</sup> in biogas is higher and varies from 40% to 60% compared to the studied range of CH<sup>4</sup> concentrations (5–30 vol%), the provided findings are in good agreement with the obtained results. The stable combustion of LCV gas containing 15 vol% of CH<sup>4</sup> is obtained within the range of φ from 0.7 to 0.85 by adding H<sup>2</sup> of 2.5vol%, while an increased addition of H<sup>2</sup> to 5 vol% has an insignificant effect, and the flammability limit is extended only to a richer region, up to φ of 0.95. However, further decrease of CH<sup>4</sup> content in LCV gases results in flame blow-off even in both cases of H<sup>2</sup> addition (Figure 3a). In order to extend the flammability limit and ensure stable combustion of LCV gases with very low CH<sup>4</sup> content (5–15 vol%), O2-enriched air was introduced. Considering that a high oxygen enrichment level could lead to the unstable flame and increased CO emissions [26], two cases with 25 vol% and 30 vol% of O<sup>2</sup> were tested (Figure 3b). As was expected, using O2-enriched air with O<sup>2</sup> content of 25 vol%, the stable combustion of waste biogases with CH<sup>4</sup> content from 30 to 5 vol% was ensured. However, in this case, the stable combustion of LCV gas containing 5 vol% of CH<sup>4</sup> is achieved only in the narrow range of φ, from 0.79 to 0.86. Meanwhile, at a higher enrichment level (30 vol% of O2), the stable combustion of LCV gases is acquired in all tested points (Figure 3b). Similar results were also acquired numerically studying methane combustion under CO2/O<sup>2</sup> atmosphere, and researchers pointed out that the most effective and stable combustion is achieved at O<sup>2</sup> concentrations of 28–32 vol% [37]. Though another work [21] showed a bit higher level of O<sup>2</sup> (31–35 vol%) is needed to achieve a stable flame. Considering the obtained results, it could be assumed that the designed burned with the adiabatic chamber ensure stable combustion of waste biogases (5–30 vol% of CH<sup>4</sup> in CO2) at lower concentrations of O<sup>2</sup> (25–30 vol%), due to a special design

higher fuel to air ratios than at lower ones, and this was found for all studied cases (Figure 3).

of the burner allowing the faster mixing of fuel oxidizer, faster reaction kinetic and minimal heat loss via the adiabatic combustion chamber walls. However, the presented design of the LCV gas burner shows an opposite relation in non-premixed combustion. The flame extinguish is more likely at higher fuel to air ratios than at lower ones, and this was found for all studied cases (Figure 3). Typically, the flame is more stable, or a flash-back phenomenon is more likely at lower fuel to air ratios, even in premixed flames [25]. The reasons for an instability arose might be related to ongoing several complex processes: The aerodynamic of the flame shape, which allows a proper internal hot gas recirculation zone, and an amount of heat provided from both recirculating gases and re-radiation from furnace walls [37]. *Energies* **2020**, *13*, x FOR PEER REVIEW 8 of 16 Typically, the flame is more stable, or a flash-back phenomenon is more likely at lower fuel to air ratios, even in premixed flames [25]. The reasons for an instability arose might be related to ongoing several complex processes: The aerodynamic of the flame shape, which allows a proper internal hot gas recirculation zone, and an amount of heat provided from both recirculating gases and re-radiation from furnace walls [37].

In order to study the impact of those two parameters, the flame core temperature and averaged axial flow speed were analyzed. Figure 4 represents data obtained by T<sup>1</sup> thermocouple and shows temperature ranges during the combustion of waste biogases (LCV) with different content of CH<sup>4</sup> in CO<sup>2</sup> for all three tested cases: Under air only combustion, under H<sup>2</sup> addition, and under O2-enriched air. The upper and bottom limits stand for the maximum (φ*max*) and minimum (φ*min*) fuel to air ratio φ at which stable combustion still occurs. As can be seen from the temperature map (Figure 4a,b), the temperature in the core of the flame also rises with an increase of φ. While in opposite the temperature decreases with a decrease of CH<sup>4</sup> content in CO2. The latter trends are well known and logical. However, the experiments showed that the temperature increase not always plays a key role and is sufficient to maintain a stable flame. Meanwhile, considering the axial velocity ranges (Figure 5) can be assumed that a change in velocities is contrary to the temperature and a decrease at higher φ ratios occurs. Based on this can be stated that a reduction in flow velocity affects an inner recirculation zone, and thus, prevents passing a sufficient amount of high temperature gas, which in turn negatively affects the combustion stability. This is well proven comparing the cases with H<sup>2</sup> addition and O<sup>2</sup> enrichment at 15 vol% of CH<sup>4</sup> in CO<sup>2</sup> as the flame temperature was higher almost by 100 ◦C under the oxygen-enriched conditions than in the case of H<sup>2</sup> addition (Figure 4a,b). Though, the flow velocities were in the same order of magnitude. Taking into account these findings, it can be concluded that a certain amount of heat is critical and should be ensured and maintained to ignite the incoming cold air-fuel mixture. In order to study the impact of those two parameters, the flame core temperature and averaged axial flow speed were analyzed. Figure 4 represents data obtained by T1 thermocouple and shows temperature ranges during the combustion of waste biogases (LCV) with different content of CH4 in CO2 for all three tested cases: Under air only combustion, under H2 addition, and under O2-enriched air. The upper and bottom limits stand for the maximum (*ϕmax*) and minimum (*ϕmin*) fuel to air ratio *ϕ* at which stable combustion still occurs. As can be seen from the temperature map (Figure 4a,b), the temperature in the core of the flame also rises with an increase of *ϕ*. While in opposite the temperature decreases with a decrease of CH4 content in CO2. The latter trends are well known and logical. However, the experiments showed that the temperature increase not always plays a key role and is sufficient to maintain a stable flame. Meanwhile, considering the axial velocity ranges (Figure 5) can be assumed that a change in velocities is contrary to the temperature and a decrease at higher *ϕ* ratios occurs. Based on this can be stated that a reduction in flow velocity affects an inner recirculation zone, and thus, prevents passing a sufficient amount of high temperature gas, which in turn negatively affects the combustion stability. This is well proven comparing the cases with H2 addition and O2 enrichment at 15 vol% of CH4 in CO2 as the flame temperature was higher almost by 100 °C under the oxygen-enriched conditions than in the case of H2 addition (Figure 4a,b). Though, the flow velocities were in the same order of magnitude. Taking into account these findings, it can be concluded that a certain amount of heat is critical and should be ensured and maintained to ignite the incoming cold air-fuel mixture.

**Figure 4.** Temperature ranges during waste biogas (LCV) combustion with different content of CH4 in CO2 for a range of (**a**) H2 addition and (**b**) O2 air enrichment. The upper and bottom limit stands for a *ϕmax* and *ϕmin,* respectively. **Figure 4.** Temperature ranges during waste biogas (LCV) combustion with different content of CH<sup>4</sup> in CO<sup>2</sup> for a range of (**a**) H<sup>2</sup> addition and (**b**) O<sup>2</sup> air enrichment. The upper and bottom limit stands for a φ*max* and φ*min,* respectively.

*Energies* **2020**, *13*, x FOR PEER REVIEW 9 of 16

**Figure 5.** Axial velocity ranges during waste biogas (LCV) combustion with different content of CH4 in CO2 for a range of (**a**) H2 addition and (**b**) O2 air enrichment. The upper and bottom limit stands for a *ϕmax* and *ϕmin*, respectively. **Figure 5.** Axial velocity ranges during waste biogas (LCV) combustion with different content of CH<sup>4</sup> in CO<sup>2</sup> for a range of (**a**) H<sup>2</sup> addition and (**b**) O<sup>2</sup> air enrichment. The upper and bottom limit stands for a φ*max* and φ*min*, respectively.

#### *3.2. Changes in Flame Chemiluminescence 3.2. Changes in Flame Chemiluminescence*

For a better understanding of waste biogas (LCV) flame behavior under H2 or O2 addition and identify a possible performance for flame transition modes, chemiluminescence emission spectrums from flames of different LCV gas mixtures were obtained. The obtained OH\* emission intensities distributed per the chamber length at *ϕ* = 0.85 are presented in Figure 6. The highest OH\* intensities were observed at the front of the flame (17 cm from the burner) where the main reactions take place. Comparing cases with and without H2 enrichment, it was observed that the H2 addition of 2.5 vol% has a negligible effect on the LCV flame enhancement, especially for the mixture with 25 vol% CH4 in CO2 as the combustion chamber temperatures are near identical to ones without the H2 addition (Figure 4a). Similar results were also determined analyzing the obtained OH\* emission intensities of hydrogen-enriched LCV flames (Figure 6a). Though, the increase of hydrogen addition to 5 vol% led to higher OH\* emission intensities by ~1.6 times and a higher chamber temperature at point T1 by ~30 °C burning LCV gas mixtures with CH4 content of 25 vol% and 20 vol%. According to previous work [27], the increased OH\* intensities and chamber temperatures could be attributed to an improved flame stability. However, the waste biogas (LCV) flammability limit using the H2 enrichment was achieved burning mixtures with CH4 content of 15 vol%, and a further decrease in CH4 content led to the flame blow-off. In the case of the mixture with 15 vol% CH4, lowermost OH\* emission intensities, and chamber temperatures were determined using both H2 additions (2.5 and 5 vol%). This could be related to a high level of the diluent (CO2), which in turn lowers the flame temperature, and thus, OH\* emission. The latter findings are in agreement with those found in the work of Guiberi et al. [38], where pointed out that OH\* intensity decrease of CO2-diluted flame is more intense with the increasing diluent concentration. For a better understanding of waste biogas (LCV) flame behavior under H<sup>2</sup> or O<sup>2</sup> addition and identify a possible performance for flame transition modes, chemiluminescence emission spectrums from flames of different LCV gas mixtures were obtained. The obtained OH\* emission intensities distributed per the chamber length at φ = 0.85 are presented in Figure 6. The highest OH\* intensities were observed at the front of the flame (17 cm from the burner) where the main reactions take place. Comparing cases with and without H<sup>2</sup> enrichment, it was observed that the H<sup>2</sup> addition of 2.5 vol% has a negligible effect on the LCV flame enhancement, especially for the mixture with 25 vol% CH<sup>4</sup> in CO<sup>2</sup> as the combustion chamber temperatures are near identical to ones without the H<sup>2</sup> addition (Figure 4a). Similar results were also determined analyzing the obtained OH\* emission intensities of hydrogen-enriched LCV flames (Figure 6a). Though, the increase of hydrogen addition to 5 vol% led to higher OH\* emission intensities by ~1.6 times and a higher chamber temperature at point T<sup>1</sup> by ~30 ◦C burning LCV gas mixtures with CH<sup>4</sup> content of 25 vol% and 20 vol%. According to previous work [27], the increased OH\* intensities and chamber temperatures could be attributed to an improved flame stability. However, the waste biogas (LCV) flammability limit using the H<sup>2</sup> enrichment was achieved burning mixtures with CH<sup>4</sup> content of 15 vol%, and a further decrease in CH<sup>4</sup> content led to the flame blow-off. In the case of the mixture with 15 vol% CH4, lowermost OH\* emission intensities, and chamber temperatures were determined using both H<sup>2</sup> additions (2.5 and 5 vol%). This could be related to a high level of the diluent (CO2), which in turn lowers the flame temperature, and thus, OH\* emission. The latter findings are in agreement with those found in the work of Guiberi et al. [38], where pointed out that OH\* intensity decrease of CO2-diluted flame is more intense with the increasing diluent concentration.

*Energies* **2020**, *13*, x FOR PEER REVIEW 10 of 16

**Figure 6.** OH\* spectral intensity distribution per distance at *ϕ* = 0.85 under (**a**) hydrogen-enrichment and (**b**) oxygen-enriched air. **Figure 6.** OH\* spectral intensity distribution per distance at φ = 0.85 under (**a**) hydrogen-enrichment and (**b**) oxygen-enriched air.

A bit different results were obtained using oxygen-enriched air (Figure 6b). As an example, an increased O2 content by 4 vol% in the air affected the mixture with 25 vol% of CH4 combustion process A bit different results were obtained using oxygen-enriched air (Figure 6b). As an example, an increased O<sup>2</sup> content by 4 vol% in the air affected the mixture with 25 vol% of CH<sup>4</sup> combustion

process significantly, and the OH\* emission intensity at the flame front was almost four times higher compared to the intensity obtained under ordinary combustion in air. Besides, higher chamber temperatures were also observed (Figure 6b). Further increase in O<sup>2</sup> content from 25 to 30 vol% also led to increased OH\* emission intensities and the chamber temperatures. A similar trend of OH\* emission intensity increase, due to a higher O<sup>2</sup> content in the oxidizer was also determined in previous works [27,39]. According to He et al. [18], the OH\* formation reaction (7) dominates under oxygen-enriched conditions as the intensity of reaction (7) is enhanced when the concentration of O<sup>2</sup> in the oxidizer increases and vice versa. However, with increasing concentration of diluent or decreasing concentration of methane, the OH\* emissions weaken gradually, resulting in decreased spectrum intensity. This tendency was determined by burning LCV gases with lower content of CH<sup>4</sup> (20–5 vol%). Even though the supply of oxygen-enriched air expanded LCV gas flammability and was able to burn the mixture with CH<sup>4</sup> content of 10 and 5 vol%.

At the lowest CH<sup>4</sup> concentration (5 vol%) under oxygen-enriched air (25 and 30 vol% of O2), OH\* emission intensities were lower than that in the case of the mixture with 20 vol% CH<sup>4</sup> burned under air combustion mode. Though the chamber temperature at point T<sup>1</sup> was about 900 ◦C, the combustion process was stable using both oxygen-enrichments, possibly due to enhanced fuel oxidation. According to obtained results under oxygen-enriched conditions, the OH\* intensity decrease coincides with the decrease of chamber temperatures, and the OH\* emission intensities are related to the flame temperature. In overall, this tendency is in good agreement with spectral intensity ratios of IOH\*/ICH\* changing CH<sup>4</sup> content in CO<sup>2</sup> (Figure 7) even though the ratio of IOH\*/ICH\* is mostly used to determine the global equivalence ratio [40,41]. Increasing the CO<sup>2</sup> dilution level leads to decreased flame temperature as the LHV decreases, and the ratio of IOH\*/ICH\* also decreases. The supply of O2-enriched air results in rapidly increased ratios of IOH\*/ICH\*, but it also depends on the O<sup>2</sup> level in the air and CO<sup>2</sup> levels in LCV gases (Figure 7). Considering the stability of waste biogas (LCV) combustion, the ratios of IOH\*/ICH\* determined burning LCV gases under air combustion conditions could be assumed as indicators for a threshold of the stable combustion as a further decrease in the ratios of IOH\*/ICH\* lead to blow-off. Under oxygen-enriched conditions, the lowest ratio of IOH\*/ICH\* is achieved by burning LCV gas with the lowest CH<sup>4</sup> content (5 vol%), and a further decrease in CH<sup>4</sup> content or O<sup>2</sup> enrichment level also leads to the flame extinguish and blow-off. Taking into account these observations, a dashed line in black was introduced in Figure 7 to represent the stability threshold of LCV gas combustion. The ratios of IOH\*/ICH\* on or above the dashed line show the stable combustion, and the combustion stability the increases with increase of CH<sup>4</sup> content in CO<sup>2</sup> and O<sup>2</sup> level in the air. Meanwhile, the flame blow-off occurs below the stability line. Besides, the proposed indication of the stable combustion is in close agreement with the determined stability maps (Figure 4b). However, this tendency is not fully valid for the cases with H<sup>2</sup> addition. This could be related to low content of H<sup>2</sup> addition, as it might not be sufficient to represent the flame enhancement via flame chemiluminescence. The previous research showed similar results with the same level of H<sup>2</sup> addition, and the significant combustion improvement and higher intensities of OH\* and CH\* were achieved only at higher levels of H<sup>2</sup> addition [27].

*Energies* **2020**, *13*, x FOR PEER REVIEW 12 of 16

**Figure 7.** The tendency of the spectral intensity ratio of IOH\*/ICH\* with the increase of CH4 content in CO2. **Figure 7.** The tendency of the spectral intensity ratio of IOH\*/ICH\* with the increase of CH<sup>4</sup> content in CO<sup>2</sup> .

#### *3.3. Addictive Influence on NOx and CO Emission 3.3. Addictive Influence on NOx and CO Emission*

During the waste biogas (LCV) combustion tests, an online flue gas analysis was performed to determine the behavior of NOx and CO emissions under all tested cases. The obtained concentrations of NOx and CO are presented graphically in Figure 8. The burner configuration ensured low NOx (up to 25 ppm) and zero CO emissions burning LCV gases with CH4 content of 25 and 20 vol% with air in the range of *ϕ* from 0.7 to 0.9. However, as noted before, the combustion at higher *ϕ* values (>0.9) becomes unstable, and CO is formed, in which concentration in the exhaust gases rises up to 45 ppm. This trend was determined in all tested conditions. At *ϕ* values below 0.9, the stable combustion is achieved, and variations in CO and NOx emissions mainly depend on the CH4 content in LCV gases and the level of O2 or H2 addition. In the case of H2 addition (Figure 8a), the NOx concentrations increased with increasing addition of H2 compared to emissions obtained at air combustion mode. For example, in the case of the mixture with 20 vol% CH4, the NOx concentrations increased by 2-5 ppm and by 5–7 ppm using the H2 addition of 2.5 vol% and 5 vol%, respectively. According to Figure 4a, it could be considered that NOx concentrations increase due to increased flame temperature, which in turn is affected by enhanced flame speed, and thermal NOx formation intensifies with higher levels of H2 addition. Meanwhile, the CO concentrations stand at zero, due to enhanced combustion. It is important to point out, that the lowest NOx (up to 14 ppm) and zero CO concentrations using the H2 addition are obtained burning the LCV gases with CH4 content of 15 vol%. In comparison, this is not able to combust under normal conditions (Figure 8a). Besides, the obtained emissions of NOx during the waste biogas combustion with hydrogen addition does not exceed the emission limit value (97 ppm) according to EU directive 2015/2193 [30]. Considering the oxygen-enriched conditions, the stable combustion of the 15 vol% CH4 in CO2 is also achieved, but NOx concentrations increase up to 78 ppm (Figure 8b). Moreover, with the increase of O2 enrichment in air, NOx increases gradually, even an N2 concentration in air decreases. This tendency was also determined with LCV gases containing higher LHV values, and the highest NOx concentrations were determined (up to 1100 ppm) burning the LCV gas with CH4-25%. However, minor emissions of NOx (up to 20 and 40 ppm) were observed burning LCV gases with CH4 content of 5 vol% and 10 vol%, respectively. Besides, the results show that during the combustion of waste biogas (up to 20 vol% CH4) under oxygen-enriched conditions (25 and 30 vol% of O2), the NOx and CO emissions do not exceed the established emission limits [30]. But the obtained results during the combustion of During the waste biogas (LCV) combustion tests, an online flue gas analysis was performed to determine the behavior of NOx and CO emissions under all tested cases. The obtained concentrations of NOx and CO are presented graphically in Figure 8. The burner configuration ensured low NOx (up to 25 ppm) and zero CO emissions burning LCV gases with CH<sup>4</sup> content of 25 and 20 vol% with air in the range of φ from 0.7 to 0.9. However, as noted before, the combustion at higher φ values (>0.9) becomes unstable, and CO is formed, in which concentration in the exhaust gases rises up to 45 ppm. This trend was determined in all tested conditions. At φ values below 0.9, the stable combustion is achieved, and variations in CO and NOx emissions mainly depend on the CH<sup>4</sup> content in LCV gases and the level of O<sup>2</sup> or H<sup>2</sup> addition. In the case of H<sup>2</sup> addition (Figure 8a), the NOx concentrations increased with increasing addition of H<sup>2</sup> compared to emissions obtained at air combustion mode. For example, in the case of the mixture with 20 vol% CH4, the NOx concentrations increased by 2–5 ppm and by 5–7 ppm using the H<sup>2</sup> addition of 2.5 vol% and 5 vol%, respectively. According to Figure 4a, it could be considered that NOx concentrations increase due to increased flame temperature, which in turn is affected by enhanced flame speed, and thermal NOx formation intensifies with higher levels of H<sup>2</sup> addition. Meanwhile, the CO concentrations stand at zero, due to enhanced combustion. It is important to point out, that the lowest NOx (up to 14 ppm) and zero CO concentrations using the H<sup>2</sup> addition are obtained burning the LCV gases with CH<sup>4</sup> content of 15 vol%. In comparison, this is not able to combust under normal conditions (Figure 8a). Besides, the obtained emissions of NOx during the waste biogas combustion with hydrogen addition does not exceed the emission limit value (97 ppm) according to EU directive 2015/2193 [30]. Considering the oxygen-enriched conditions, the stable combustion of the 15 vol% CH<sup>4</sup> in CO<sup>2</sup> is also achieved, but NOx concentrations increase up to 78 ppm (Figure 8b). Moreover, with the increase of O<sup>2</sup> enrichment in air, NOx increases gradually, even an N<sup>2</sup> concentration in air decreases. This tendency was also determined with LCV gases containing higher LHV values, and the highest NOx concentrations were determined (up to 1100 ppm) burning the LCV gas with CH4-25%. However, minor emissions of NOx (up to 20 and 40 ppm) were observed burning LCV gases with CH<sup>4</sup> content of 5 vol% and 10 vol%, respectively. Besides, the results show that during the combustion of waste biogas (up to 20 vol% CH4) under oxygen-enriched conditions (25 and 30 vol% of O2), the NOx and CO emissions do not exceed the established emission limits [30]. But the obtained results during the combustion of mixtures with a higher amount of CH<sup>4</sup> in CO<sup>2</sup> confirm that the NOx

formation strongly belongs to the thermal NO formation route under oxygen-enriched conditions. As it mainly depends on the flame temperature, the O<sup>2</sup> enrichment level should be controlled based on the LHV value to prevent unwanted NOx formation. flame temperature, the O2 enrichment level should be controlled based on the LHV value to prevent unwanted NOx formation.

*Energies* **2020**, *13*, x FOR PEER REVIEW 13 of 16

**Figure 8.** NOx and CO emissions of all tested waste biogas (LCV) mixture combustion under hydrogen-enrichment (**a**) and oxygen-enriched air (**b**) versus fuel-to-air ratio. **Figure 8.** NOx and CO emissions of all tested waste biogas (LCV) mixture combustion under hydrogen-enrichment (**a**) and oxygen-enriched air (**b**) versus fuel-to-air ratio.

#### **4. Conclusions 4. Conclusions**

This study describes a flexible swirl burner (~100 kW) with an adiabatic chamber, which was designed and developed to burn unstable composition waste biogases (5–30 vol% of CH4 in CO2) and performed tests at real conditions. During the experimental tests, the flame stability and combustion limits were explored under air, oxygen-enriched atmosphere, or with hydrogen addition. The stable operation of non-premixed swirl combustion of waste biogas in the air was achieved for CH4 content in CO2 with not less than 20 vol%. Further decrease of CH4 content in CO2 requires H2 or O2 enrichment to extend flammability and improve flame stability. It was determined that by adding up to 5 vol% of H2, the stable combustion of waste biogas with CH4 content of 15 vol% is obtained. A greater improvement in flame stability of waste biogas was established using O2-enrichment. The supply of O2 by 25 and 30 vol% resulted in the stable combustion of waste biogases gases containing 10 vol% This study describes a flexible swirl burner (~100 kW) with an adiabatic chamber, which was designed and developed to burn unstable composition waste biogases (5–30 vol% of CH<sup>4</sup> in CO2) and performed tests at real conditions. During the experimental tests, the flame stability and combustion limits were explored under air, oxygen-enriched atmosphere, or with hydrogen addition. The stable operation of non-premixed swirl combustion of waste biogas in the air was achieved for CH<sup>4</sup> content in CO<sup>2</sup> with not less than 20 vol%. Further decrease of CH<sup>4</sup> content in CO<sup>2</sup> requires H<sup>2</sup> or O<sup>2</sup> enrichment to extend flammability and improve flame stability. It was determined that by adding up to 5 vol% of H2, the stable combustion of waste biogas with CH<sup>4</sup> content of 15 vol% is obtained. A greater improvement in flame stability of waste biogas was established using O2-enrichment. The supply of O<sup>2</sup> by 25 and 30 vol% resulted in the stable combustion of waste biogases gases containing 10 vol% and

5 vol% of CH4, respectively. Besides, the tests of the developed swirl burner at real conditions revealed that the flame blow-off is more likely at higher fuel to air ratios (φ > 0.95), and this was established for all studied cases.

This work shows the different stabilization modes and the changes in intensities of OH\*chemiluminescence spectra, which enables to determine the possible transition between regimes and when flammability limit might occur. The linear correlation between the spectral intensity ratio of OH\* and CH\* (IOH\*/ICH\* ) and CH<sup>4</sup> content in LCV gas is determined, which clearly demonstrates the change in intensity for the tested modes and can be used for the prediction of blow-off limits burning different LCV gases under different H<sup>2</sup> or O<sup>2</sup> enrichments.

During all tests, the online flue gas analysis was performed to determine the behavior of NOx and CO emissions and explore its relation to flame stability. Minor emissions of NOx (up to 20 and 40 ppm) are observed only burning waste biogas (LCV) in air and with H<sup>2</sup> addition. During combustion under oxygen-enriched conditions, such concentrations can be obtained only by combusting waste biogases containing the CH<sup>4</sup> content of 5 vol% and 10 vol%, respectively. The CO emissions in most cases, were determined to be at zero levels and formed only by approaching the blow-off limit. In that case, the proposed combustion system meets the EU established emission standards for gas-burning systems.

However, if waste biogas calorific value increase, O<sup>2</sup> enrichment should be minimized to prevent unwanted high NOx formation. Besides, thermal damage to the combustion system and high NOx emissions are possible if the oxygen-enriched air is supplied with higher O<sup>2</sup> concentrations than 30 vol%.

Overall, the developed swirl burner with the adiabatic combustion chamber is a flexible combustion system that could be installed in the biogas plant to produce heat and electricity and also provides a solution to recover primary energy and reduce pollutant emissions from biomethane production obtained waste biogases.

**Author Contributions:** Conceptualization, N.S. and R.P.; methodology, N.S.; software, R.S.; validation, N.S. and R.S.; formal analysis, N.S. and R.P.; investigation, N.S., R.P., R.S. and A.L.; resources, N.S.; data curation, N.S. and R.P.; writing—original draft preparation, N.S. and R.P.; writing—review and editing, R.S. and A.L.; visualization, A.L.; supervision, N.S.; project administration, N.S.; funding acquisition, R.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This project has received funding from the Research Council of Lithuania (LMTLT), agreement No P-MIP-17-26.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **E**ffi**ciency Analysis and Integrated Design of Rocket-Augmented Turbine-Based Combined Cycle Engines with Trajectory Optimization**

#### **Feng Guo <sup>1</sup> , Wenguo Luo <sup>1</sup> , Feng Gui <sup>2</sup> , Jianfeng Zhu 1,\*, Yancheng You <sup>1</sup> and Fei Xing <sup>1</sup>**


Received: 9 May 2020; Accepted: 2 June 2020; Published: 5 June 2020

**Abstract:** An integrated analysis method for a rocket-augmented turbine-based combined cycle (TBCC) engine is proposed based on the trajectory optimization method of the Gauss pseudospectral. The efficiency and energy of the vehicles with and without the rocket are analyzed. Introducing an appropriate rocket to assist the TBCC-powered vehicle will reduce the total energy consumption of drag, and increase the vehicle efficiency in the transonic and the mode transition. It results in an increase in the total efficiency despite a reduction in engine efficiency. Therefore, introducing a rocket as the auxiliary power is not only a practical solution to enable flight over a wide-speed range when the TBCC is incapable but also probably an economical scheme when the the TBCC meets the requirements of thrust. When the vehicle drag is low, the rocket works for a short time and its optimal relative thrust is small. Thus, the TBCC combined with a booster rocket will be a more simple and suitable scheme. When the vehicle drag is high, the operating time of the rocket is long and the optimal relative thrust is large. The specific impulse has a significant impact on the flight time and the total fuel consumption. Accordingly, the combination form for the rocket-based combined cycle (RBCC) engines and the turbine will be more appropriate to obtain higher economic performance.

**Keywords:** turbine-based combined cycle engine; TBCC; rocket-augmented; trajectory optimization; Gauss pseudospectral method; efficiency analysis; combined design; integrated design

## **1. Introduction**

Turbine-based combined cycle (TBCC) engines are believed to be a promising means of power for wide-speed range hypersonic vehicles [1]. Due to combining turbine engines with ramjet engines or dual-mode scramjet engines, typical TBCC engines have the advantages of low fuel consumption, the high reliability of turbine engines and the high-speed cruise ability of ramjet/dual mode ramjet (DMRJ) engines [2]. However, several technical issues with TBCC engines still exist, such as "Thrust Pinch", so the engines are incapable of providing sufficient thrust for acceleration during transonic and mode transition [3]. To resolve this problem, the "rocket-augmented TBCC" concept has been proposed as a near-term solution, where the rockets are used as auxiliary power for thrust augmentation when extra thrust is needed.

In terms of combination form, "rocket-augmented TBCC" engines are classified into two broad categories: "T+RBCC" and "R+TBCC". For "T+RBCC" engines, the rockets are typically mounted within the high-speed ducts and integrated as ejector-ramjets (ERJs). To date, some schemes have been published, such as TriJet [3,4], TRRE [5,6], and XTER [7]. All these engines utilize the thrust and specific impulse potential of ERJs [8–10], but the rockets should be synergistically integrated into the thermodynamic cycle at the cost of complexity of the rocket-matching design. By comparison, the rockets in "R+TBCC" engines are installed as additional boosters, which are generally isolated from the typical TBCC engines. Many researchers have discussed these kinds of engine and vehicle as well, for example, the advanced Rocket/Dual-mode ramjet propulsion of Lapcat II [11] and the high-speed near space flight vehicle proposed by the Beijing Institute of Mechanical and Electrical Engineering [12]. Although the integration design method for the "R+TBCC" engines is relatively simple, this combination may increase extra vehicle drag and inert weight, which may raise the fuel consumption and reduce the payload weight, respectively. From the above, in order to achieve synergistic benefits from the collaboration of TBCCs and rockets excellently, the performance of the vehicle, TBCC, and rockets should be taken into account when designing the combination form of the rocket-augmented TBCC.

Generally, the rockets are introduced to meet the vehicle thrust requirement when the TBCCs are incapable of providing sufficient thrust. Naturally, due to the low specific impulse (Isp) of rockets, the economic performance of the vehicle seems to be reduced. However, in flight, the acceleration increase by the rockets reduces the flight time and results in fuel saving for the TBCC engine. Therefore, the rockets used for the thrust augmentation have a mixed effect on the total economic performance. In addition, the acceleration characteristics are related to trajectory features; for example, the trajectory of the "climb-dive" (or gravity-assist) is used to overcome the transonic thrust pinch in the acceleration phase. Thus, the tradeoff should take into consideration the trajectory features in the thrust pinch regions. The rocket-matching analysis based on trajectory optimization is necessary for the integrated design of the rocket-augmented TBCC. Trajectory optimization methods have been studied for decades [13–15]. The Gauss pseudospectral method as a direct method has the ability to obtain accurate estimates of the state, costate, and control for continuous time optimal control problems [16] and has been successfully implemented in trajectory optimization problems of supersonic or hypersonic vehicles [17,18]. Thus, the Gauss pseudospectral method could be directly used to evaluate the integrated design of a rocket-augmented TBCC under different rocket schemes, and the results from trajectory optimization could be used for choosing the appropriate combination form for rockets and TBCCs for vehicle systems.

The paper is organized as follows: Section 2 will give a problem statement of the rocket-augmented strategy from the aspect of efficiency. In Section 3, an integrated analysis model for a rocket-augmented TBCC-powered vehicle with the trajectory optimization method of the Gaussian pseudospectral is established. Section 4 analyzes the performance of the vehicle from the aspect of efficiency and energy, and the rocket thrust is optimized in a specified vehicle case. Based on the above understanding, the parameter study of different vehicles and rockets is discussed in Section 5, and the combination form for rockets and TBCCs is given. Finally, Section 6 summarizes this paper.

#### **2. Problem Statement**

TBCC vehicles suffer from two main thrust pinches [1], as shown in Figure 1a. One is the transonic thrust pinch, and the other is the transition thrust pinch. The thrust pinches are limited by the current performance level of vehicle and TBCC engines. The thrust pinch issues can be described as flight with small acceleration and long acceleration times, which can be measured by vehicle efficiency, η*veh* [18]:

$$
\eta\_{\text{level}} = E\_{\text{veh}} / (TV \cos \alpha) \tag{1}
$$

*Eveh* represents the vehicle work:

$$E\_{\text{vel}} = (T \cos \alpha - D - mg \sin r)V \tag{2}$$

where *T* and *D* denote the thrust and drag, respectively. α and *r* denote angle of attack and the flight path angle, respectively. *m* denotes the mass of the vehicle, and *V* denotes the flight velocity.

thrust pinch region is very low.

thrust pinch region is very low.

*Energies* **2020**, *13*, 2911 3 of 17

As shown in Figure 1b, the reference ℎ curve could effectively reflect the acceleration efficiency characteristics corresponding to Figure 1a. It can be seen that the vehicle efficiency in the

**Figure 1.** (**a**) Thrust pinch in transonic and mode transition; (**b**) Vehicle efficiency. **Figure 1.** (**a**) Thrust pinch in transonic and mode transition; (**b**) Vehicle efficiency. of carrying it onboard, like a rocket [19]. This results in the higher Isp of TBCC engines compared to

For wide-speed flight, the TBCC has the distinct advantage of utilizing oxygen in the air instead of carrying it onboard, like a rocket [19]. This results in the higher Isp of TBCC engines compared to rockets, as shown in Figure 2a [20]. Similarly, the engine efficiency can be formulated with As shown in Figure 1b, the reference η*veh* curve could effectively reflect the acceleration efficiency characteristics corresponding to Figure 1a. It can be seen that the vehicle efficiency in the thrust pinch region is very low. rockets, as shown in Figure 2a [20]. Similarly, the engine efficiency can be formulated with the Isp [21]: = ⋅ (3)

the Isp [21]: = ℎ ⋅ (3) For wide-speed flight, the TBCC has the distinct advantage of utilizing oxygen in the air instead of carrying it onboard, like a rocket [19]. This results in the higher Isp of TBCC engines compared to rockets, as shown in Figure 2a [20]. Similarly, the engine efficiency η*enigne* can be formulated with the Isp [21]: ℎ where ℎ denotes the heating value of the fuel. As shown in Figure 2b, the rockets' engine efficiency is obviously lower than that of the airbreathing engines as well, such as turbojets, ramjets, and scramjets. However, rockets have

significantly higher unit frontal area thrust (or unit weight thrust). Besides, the thrust of rockets

accelerate the vehicle when going through the thrust pinch region, especially at high altitudes.

$$
\eta\_{\text{engine}} = \frac{\text{g}V}{h\_{\text{PR}}} \cdot \text{Isp} \tag{3}
$$

breathing engines as well, such as turbojets, ramjets, and scramjets. However, rockets have where *hPR* denotes the heating value of the fuel. trend due to the reduction in air density. Therefore, a smaller rocket may generate enough thrust to

**Figure 2.** (**a**) Reference Isp of engines; (**b**) Reference efficiency of engines. **Figure 2.** (**a**) Reference Isp of engines; (**b**) Reference efficiency of engines.

(**a**) (**b**) **Figure 2.** (**a**) Reference Isp of engines; (**b**) Reference efficiency of engines. As shown in Figure 2b, the rockets' engine efficiency is obviously lower than that of the air-breathing engines as well, such as turbojets, ramjets, and scramjets. However, rockets have significantly higher unit frontal area thrust (or unit weight thrust). Besides, the thrust of rockets increases gradually as the flight altitude increases, while the air-breathing engines have the opposite trend due to the reduction

in air density. Therefore, a smaller rocket may generate enough thrust to accelerate the vehicle when going through the thrust pinch region, especially at high altitudes. According to the description of Figures 1 and 2, introducing a small rocket to assist the TBCC could significantly increase the vehicle efficiency, ℎ , in the transonic and mode transition. Although the engine efficiency, η , may be reduced when the rocket works, the rocket-

According to the description of Figures 1 and 2, introducing a small rocket to assist the TBCC could significantly increase the vehicle efficiency, η*veh*, in the transonic and mode transition. Although the engine efficiency, η*enigne*, may be reduced when the rocket works, the rocket-augmented strategy is still possible to improve the local total efficiency η*tot*, as shown in Figure 3. It is to be noted that η*tot* is defined as follows: augmented strategy is still possible to improve the local total efficiency , as shown in Figure 3. It is to be noted that is defined as follows: = ℎ ⋅ (4) This analysis shows that introducing a rocket as the auxiliary power is not only a practical

economical scheme when the TBCC meets the requirements of thrust.

$$
\eta\_{\text{tot}} = \eta\_{\text{teth}} \cdot \eta\_{\text{engine}} \tag{4}
$$

**Figure 3.** Total efficiency might be improved by rockets. **Figure 3.** Total efficiency might be improved by rockets.

**3. Integrated Analysis Method** *3.1. Description of the Vehicle Model* This analysis shows that introducing a rocket as the auxiliary power is not only a practical solution to enable flight over a wide-speed range when the TBCC is incapable but also probably an economical scheme when the TBCC meets the requirements of thrust.

#### The baseline vehicle is assumed to have a take-off weight of 19,050 kg and a wing loading of 460 kg/m<sup>2</sup> . As a result of a lack of performance data for currently available TBCC-powered vehicles, the **3. Integrated Analysis Method**

#### aerodynamic coefficients in high Mach (>1.5) refer to X − 43 [22], and the data for low Mach (≤1.5) are *3.1. Description of the Vehicle Model*

obtained from the reference [23]. It is noted that the low-Mach data in the case of X − 43 are not suitable for showing the transonic thrust pinch. For the trajectory optimization in the conceptual design, a point mass model for motion in a vertical plane is usually quite adequate [23]. The equations of motion for this model [24] are given by: The baseline vehicle is assumed to have a take-off weight of 19,050 kg and a wing loading of 460 kg/m<sup>2</sup> . As a result of a lack of performance data for currently available TBCC-powered vehicles, the aerodynamic coefficients in high Mach (>1.5) refer to X − 43 [22], and the data for low Mach (≤1.5) are obtained from the reference [23]. It is noted that the low-Mach data in the case of X − 43 are not suitable for showing the transonic thrust pinch.

ℎ̇ = (5) For the trajectory optimization in the conceptual design, a point mass model for motion in a vertical plane is usually quite adequate [23].

̇ = − − (6) The equations of motion for this model [24] are given by:

$$
\dot{h} = V \sin \gamma \tag{5}
$$

$$
\dot{V} = \frac{T \cos \alpha - D}{m} - g \sin \gamma \tag{6}
$$

$$\dot{\gamma} = \frac{\text{T}\text{sinc}\, + \, L}{mV} + \left(\frac{V}{r} - \frac{\text{g}}{V}\right) \text{cosy} \tag{7}$$

$$
\dot{m} = -\frac{T}{g \cdot lsp} \tag{8}
$$

= (, ) (10)

The lift *L* and drag *D* are defined as:

$$L = q \mathbf{S} \mathbf{C}\_L(\mathbf{a}, \mathbf{M} \mathbf{a}) \tag{9}$$

$$D = q \pounds \mathcal{C}\_D(\alpha, \mathcal{M}a) \tag{10}$$

where *q* is the flight dynamic pressure and *S* is the reference area of the vehicle. The lift coefficient *C<sup>L</sup>* and the drag coefficient *C<sup>D</sup>* are the interpolation functions of the attack angle α and the Mach number.

#### *3.2. Description of the Engines Model*

The TBCC engine model adopts the lapping combination of the Ma 4.4 turbine and the HRE scramjet from the Air Force Research Laboratory (AFRL) [22]. The ratio of the number of turbine engines to ramjet engines is 20:1, and the takeoff thrust/weight ratio is 1.0. The thrust and the Isp are obtained by interpolating the altitude and the velocity data table. To highlight the "thrust trap" feature of the mode transition, the thrust and the Isp are calculated by Equations (11) and (12) [25], respectively:

$$T(\lambda, M) = T\_{\text{turb}} \left( \frac{M\_2 - M}{M\_2 - M\_1} \right)^{\lambda} + T\_{\text{num}} \left( \frac{M - M\_1}{M\_2 - M\_1} \right)^{\lambda}, M \in [M\_1, M\_2] \tag{11}$$

$$\text{Isp}(\lambda, M) = \frac{T(\lambda, M)}{\left(\frac{T\_{\text{turb}}}{\text{Isp}\_{\text{turb}}} \left(\frac{M\_2 - M}{M\_2 - M\_1}\right)^{\lambda} + \frac{T\_{\text{Irm}}}{\text{Isp}\_{\text{lim}}} \left(\frac{M - M\_1}{M\_2 - M\_1}\right)^{\lambda}\right)}, M \in [M\_1, M\_2] \tag{12}$$

where *M*<sup>1</sup> (4.0) and *M*<sup>2</sup> (4.4) are the start and the end Mach numbers of the mode transition, and the parameter λ is the pinch coefficient and can measure the minimum thrust in mode transition. In this paper, the *T* λ, *M*1+*M*<sup>2</sup> 2 is equal to the 2/3 of *T* 1, *<sup>M</sup>*1+*M*<sup>2</sup> 2 , in which the λ is correspondingly given as 1.585.

The Isp of the baseline rocket is 300 s. The thrust of the baseline rocket is adjustable, and the adjustable thrust is the product of the rocket throttle, *Thr*, and the maximum of thrust, *TRocket*\_*max*:

$$T\_{\text{Rocket}} = T\_{\text{Rocket\\_max}} \cdot Tlr\tag{13}$$

where *TRocket*\_*max* is defined as the product of the vehicle takeoff weight, *mto*, and the maximum relative thrust of the rocket, *RTmax*:

$$T\_{\text{Rocket\\_max}} = m\_{\text{to}} \cdot \text{g} \cdot \text{RT}\_{\text{max}} \tag{14}$$

In addition, introducing the rocket increases the frontal area of the vehicle and the weight of the propulsion system, which influences the performance of the vehicle. In this study, the rocket weight is estimated at 1/50 of the maximum thrust. The drag coefficient is used to estimate the performance affected by the frontal area [12].

#### *3.3. Integrated Trajectory Optimization Method*

#### 3.3.1. Optimization Problem

According to both the vehicle model and the engine model, the flight altitude *h*, speed *V*, flight path angle γ, vehicle mass *m*, and attack angle α are set as state parameters. Meanwhile, the gradient of the attack angle . <sup>α</sup> and rocket throttle ratio *thr* are set as control variables. It is noted that the attack angle . α is aimed at smoothing flight control. Based on the vehicle control equations, the optimization problem of the ascent trajectory is formulated with the corresponding cost function and parameter constraints.

The objective is to determine the minimum-fuel (minF) trajectory and control from takeoff to a specified speed and altitude. Besides, the minimum-time (minT) trajectory optimization problem is analyzed as a comparative case. The cost functions are formulated as

$$J\_{\rm min,fuel} = -m \text{(t}\_f\text{)}\tag{15}$$

$$\mathbf{J}\_{\text{min,time}} = \mathbf{t}\_f \tag{16}$$

The boundary conditions and constraints are shown in Table 1.

**Table 1.** Boundary conditions and constraints.


Additionally, the operating range of dynamic pressures, *q*, is also restrained in different flight phases.

$$\begin{cases} 10kPa \le q \le 75kPa, \quad Ma \le M\_1, Ma > M\_2\\ 40kPa \le q \le 75kPa, \quad M\_1 < Ma \le M\_2 \end{cases} \tag{17}$$

#### 3.3.2. Gauss Pseudospectral Method

The optimization problem above can be converted to a continuous Bolza problem by using the Gauss pseudospectral method (GPM). The GPM constructs the Lagrange interpolation polynomials on a set of Legendre points to approximate the state variables and control variables of the system, so the continuous-optimal control problem can be transcribed into a nonlinear planning problem (NLP). Studies have shown that Karush–Kuhn–Tucker (KKT) multipliers of the NLP can be utilized to estimate the costate at both the Legendre-Gauss points and the boundary points accurately, which is due to the equivalence between the KKT conditions and the discretized first-order necessary conditions [12]. In this paper, the trajectory optimization problem is solved by the general software package GPOPS; more details of this algorithm are discussed in [26,27].

#### **4. E**ffi**ciency Analysis and Rocket Optimization**

According to the predictions in the problem statement, the engine efficiency is depressed when the rocket is introduced, yet the vehicle efficiency in thrust pinches may be improved significantly. With overall consideration, introducing a rocket is still possible to increase the total efficiency. In this section, the efficiency analysis of the trajectory with the rocket is performed on the baseline TBCC-powered vehicle by using the integrated analysis method. However, introducing a rocket increases the frontal area of the vehicle and the weight of the propulsion system. Taking the above factor into account, the rocket with an appropriate thrust will yield the optimal performance. Therefore, the rocket optimization is also studied in this section.

Figure 4 shows the trajectories of the baseline vehicle with minF and minT as the optimization objectives, and of the rocket-augmented TBCC-powered vehicle with minF as the optimization objective. Simulations are performed from Mach 0.38 to 5 with the constraints given in Table 1. For the baseline vehicle, the minF path shows that the vehicle goes through transonic and mode transition in the way of rapid climb and dive, and flies at the constant dynamic pressure of 75 kPa in most of the remaining voyage. By contrast, the minT path shows a relatively lower climb altitude during the transonic and without "climb-dive" during mode transition. It is indicated that the "climb-dive" trajectory strategy

is an efficient means of fuel saving in the thrust pinch regions. It is also necessary when going through *Energies* transonic in the path of minT, which has been discussed in J. Zheng et al.'s work [ **2020**, *13*, 2911 25]. 7 of 17

*Energies* **2020**, *13*, 2911 7 of 17

**Figure 4.** Flight envelopes. **Figure 4.** Flight envelopes. Figure 5 shows a detailed comparison of the flight time of the three trajectories above. For the

Figure 5 shows a detailed comparison of the flight time of the three trajectories above. For the TBCC-powered vehicle, the times taken to reach Mach 5 of minF and minT are 719.2 s and 614.5 s, respectively. The time difference between the minF and minT closely depends on the different trajectory strategy in the transonic. The "climb-dive" trajectory strategy for fuel saving takes a great amount of time. The total time of the rocket-augmented minF is 648.4 s, which is 9.8% less than that Based on the TBCC model, the rocket with an *RTmax* of 10% is used to optimize the trajectory of minF, shown as the dashed line in Figure 4. Interestingly, the rocket does not work in the transonic region, but only turns on during the mode transition. Moreover, the vehicle firstly climbs to the limit boundary of the minimum dynamic pressure of 40 kPa and then takes advantage of the gravity-assist to accelerate during the mode transition. Due to the change in trajectory when the rocket works, the local performance of vehicle might be changed. TBCC-powered vehicle, the times taken to reach Mach 5 of minF and minT are 719.2 s and 614.5 s, respectively. The time difference between the minF and minT closely depends on the different trajectory strategy in the transonic. The "climb-dive" trajectory strategy for fuel saving takes a great amount of time. The total time of the rocket-augmented minF is 648.4 s, which is 9.8% less than that of the baseline minF. The time difference mainly depends on the different accelerations in the mode transition. In addition, as a result of the flight drag caused by the rocket, the trajectory of the minF with rocket takes slightly more flight time than the baseline minF to accelerate to the same Mach

of the baseline minF. The time difference mainly depends on the different accelerations in the mode transition. In addition, as a result of the flight drag caused by the rocket, the trajectory of the minF with rocket takes slightly more flight time than the baseline minF to accelerate to the same Mach number. The throttle control of the rocket is also shown in Figure 5. The rocket works at 495.8 s and lasts for 34.3 s. The corresponding Mach number ranges from 4.12 to 4.39. By contrast, the TBCC-powered vehicle without a rocket takes 119.5 s to flight over the same Mach range. In addition, the rocket throttle is allowed to vary freely between 0% and 100% in the trajectory optimization; however, the results show that the throttle mostly remains in the two states of 0% and 100%. This indicates that the Figure 5 shows a detailed comparison of the flight time of the three trajectories above. For the TBCC-powered vehicle, the times taken to reach Mach 5 of minF and minT are 719.2 s and 614.5 s, respectively. The time difference between the minF and minT closely depends on the different trajectory strategy in the transonic. The "climb-dive" trajectory strategy for fuel saving takes a great amount of time. The total time of the rocket-augmented minF is 648.4 s, which is 9.8% less than that of the baseline minF. The time difference mainly depends on the different accelerations in the mode transition. In addition, as a result of the flight drag caused by the rocket, the trajectory of the minF with rocket takes slightly more flight time than the baseline minF to accelerate to the same Mach number. number. The throttle control of the rocket is also shown in Figure 5. The rocket works at 495.8 s and lasts for 34.3 s. The corresponding Mach number ranges from 4.12 to 4.39. By contrast, the TBCC-powered vehicle without a rocket takes 119.5 s to flight over the same Mach range. In addition, the rocket throttle is allowed to vary freely between 0% and 100% in the trajectory optimization; however, the results show that the throttle mostly remains in the two states of 0% and 100%. This indicates that the adjustment of thrust is not necessary. Therefore, the fixed-thrust design could satisfy the thrust requirement and simplify the structure of the rocket.

adjustment of thrust is not necessary. Therefore, the fixed-thrust design could satisfy the thrust

**Figure 5.** Flight time profiles. **Figure 5.** Flight time profiles.

4444.2 kg, respectively. The fuel saving is 424.4 kg (8.5%) from the minT to the minF, and a further

**Figure 5.** Flight time profiles.

Figure 6 shows the vehicle mass variations with the flight Mach. The total fuel consumption of

The throttle control of the rocket is also shown in Figure 5. The rocket works at 495.8 s and lasts for 34.3 s. The corresponding Mach number ranges from 4.12 to 4.39. By contrast, the TBCC-powered vehicle without a rocket takes 119.5 s to flight over the same Mach range. In addition, the rocket throttle is allowed to vary freely between 0% and 100% in the trajectory optimization; however, the results show that the throttle mostly remains in the two states of 0% and 100%. This indicates that the adjustment of thrust is not necessary. Therefore, the fixed-thrust design could satisfy the thrust requirement and simplify the structure of the rocket.

Figure 6 shows the vehicle mass variations with the flight Mach. The total fuel consumption of the minT, the minF, and the rocket-augmented minF to reach Mach 5 are 4992.8 kg, 4568.4 kg, and 4444.2 kg, respectively. The fuel saving is 424.4 kg (8.5%) from the minT to the minF, and a further 124.2 kg (2.7%) from the minF to the rocket-augmented minF. Due to the gravity-assist, compared with the minT, the minF saves 415.3 kg of fuel in the process of transonic, contributing 97.9% of the total fuel saving. With the assistance of the rocket, an additional 201.6 kg of fuel is saved during the mode transition, which improves the economic performance of the TBCC vehicle even with the additional fuel consumption resulting from the drag of the rocket. If the weight of the rocket is considered as a part of the fuel consumption, the fuel saving is 86.1 kg (1.9%). Based on these results, it appears that the rocket-augmented TBCC is a better option to promote economic performance than the TBCC without a rocket for wide-speed range vehicles. *Energies* **2020**, *13*, 2911 8 of 17 124.2 kg (2.7%) from the minF to the rocket-augmented minF. Due to the gravity-assist, compared with the minT, the minF saves 415.3 kg of fuel in the process of transonic, contributing 97.9% of the total fuel saving. With the assistance of the rocket, an additional 201.6 kg of fuel is saved during the mode transition, which improves the economic performance of the TBCC vehicle even with the additional fuel consumption resulting from the drag of the rocket. If the weight of the rocket is considered as a part of the fuel consumption, the fuel saving is 86.1 kg (1.9%). Based on these results, it appears that the rocket-augmented TBCC is a better option to promote economic performance than the TBCC without a rocket for wide-speed range vehicles.

**Figure 6.** Vehicle mass profiles. **Figure 6.** Vehicle mass profiles.

The TBCC-powered vehicle assisted by a rocket could improve both the flight timeliness and the economic performance, which could be explained in terms of the energy and efficiency. The TBCC-powered vehicle assisted by a rocket could improve both the flight timeliness and the economic performance, which could be explained in terms of the energy and efficiency.

From the perspective of energy, the vehicle climbs from a low-potential and low-kinetic state to a high-potential and high-kinetic state. In the process of ascending, besides increasing kinetic energy and potential energy, the thrust provides the most of the energy to overcome the drag. For instance, the drag dissipates 35.45 GJ of energy in the trajectory of the minT, while the increases in the kinetic energy and the potential energy are 15.56 GJ and 3.42 GJ, respectively. By contrast, the drag dissipates 31.31 GJ of energy in the trajectory of the minF, which is 11.7% less than that of the minT. Under the assistance of the rocket in the minF, the energy consumption of drag is reduced from 31.31 GJ to 26.58 GJ. The difference in the drag dissipation between the minF and the minT mainly lies in the transonic region, as shown in Figure 7a. The energy of the drag has a significant reduction from the minT to the minF, which results from the gravity-assist strategy. After climbing to a higher altitude, the vehicle flies with a smaller drag and lower energy consumption. Meanwhile, the thrust is reduced and the Isp of the engine remains almost constant in the process of ascending, which is one of the reasons for the fuel saving. Since the gravitational potential energy is comparable to the kinetic energy in the transonic region, the gravity-assist strategy could achieve an efficient increase in the kinetic energy from the conversion of the potential energy. However, in the mode transition, the potential energy is much less than the kinetic energy and the gravity-assist strategy is inefficient, as From the perspective of energy, the vehicle climbs from a low-potential and low-kinetic state to a high-potential and high-kinetic state. In the process of ascending, besides increasing kinetic energy and potential energy, the thrust provides the most of the energy to overcome the drag. For instance, the drag dissipates 35.45 GJ of energy in the trajectory of the minT, while the increases in the kinetic energy and the potential energy are 15.56 GJ and 3.42 GJ, respectively. By contrast, the drag dissipates 31.31 GJ of energy in the trajectory of the minF, which is 11.7% less than that of the minT. Under the assistance of the rocket in the minF, the energy consumption of drag is reduced from 31.31 GJ to 26.58 GJ. The difference in the drag dissipation between the minF and the minT mainly lies in the transonic region, as shown in Figure 7a. The energy of the drag has a significant reduction from the minT to the minF, which results from the gravity-assist strategy. After climbing to a higher altitude, the vehicle flies with a smaller drag and lower energy consumption. Meanwhile, the thrust is reduced and the Isp of the engine remains almost constant in the process of ascending, which is one of the reasons for the fuel saving. Since the gravitational potential energy is comparable to the kinetic energy in the transonic region, the gravity-assist strategy could achieve an efficient increase in the kinetic energy from the conversion of the potential energy. However, in the mode transition, the potential energy is

drag dissipation in mode transition.

much less than the kinetic energy and the gravity-assist strategy is inefficient, as shown in Figure 7b. With the assistance of the rocket, the time spent in the mode transition is reduced from 164.5 s to 73.2 s, and the energy consumption of drag is reduced from 9.73 GJ to 3.47 GJ. The results indicate that the rocket-augmented scheme seems to be a better choice for the reduction of drag dissipation in mode transition. *Energies* **2020**, *13*, 2911 9 of 17

**Figure 7.** (**a**) Energy profiles in low Mach; (**b**) Energy profiles in high Mach. **Figure 7.** (**a**) Energy profiles in low Mach; (**b**) Energy profiles in high Mach.

Figure 8 shows the efficiency profiles of the two trajectories of the minF and the rocketaugmented minF. As shown in Figure 8a, the vehicle efficiency, ℎ, has a sharp decrease before Mach 0.9 due to the rapid dropping of acceleration in the phase of the climb. In the final stage of climb, the thrust is not able to accelerate the vehicle, so the value of ℎ is negative. With the rapid dive in transonic, the ℎ increases first and then decreases, under the combined effect of drag, thrust, and gravity. By contrast, the variation of the is slight, because the Isp is insensitive to altitude. Figure 8 shows the efficiency profiles of the two trajectories of the minF and the rocket-augmented minF. As shown in Figure 8a, the vehicle efficiency, η*veh*, has a sharp decrease before Mach 0.9 due to the rapid dropping of acceleration in the phase of the climb. In the final stage of climb, the thrust is not able to accelerate the vehicle, so the value of η*veh* is negative. With the rapid dive in transonic, the η*veh* increases first and then decreases, under the combined effect of drag, thrust, and gravity. By contrast, the variation of the η*enigne* is slight, because the Isp is insensitive to altitude.

*Energies* **2020**, *13*, 2911 10 of 17

However, when the rocket works, the ℎ is improved and the is reduced during the

**Figure 8.** (**a**) Vehicle efficiency profiles; (**b**) Engine efficiency profiles; (**c**) Total efficiency profiles. **Figure 8.** (**a**) Vehicle efficiency profiles; (**b**) Engine efficiency profiles; (**c**) Total efficiency profiles.

**Figure 9.** Rocket *RTmax* vs. weight consumption.

In Figure 9, when the rocket thrust is relatively small (*RTmax* < 6%), as the thrust of the rocket increases, the fuel consumption of the TBCC reduces rapidly. Even if the fuel consumption and the weight of the rocket are increased, the total weight of consumption is still reduced. It is noted that the total weight in Figure 9 is the sum of the fuel consumption of the TBCC, the fuel consumption of the rocket, and the weight of the rocket. As the thrust increases (*RTmax* > 6%), the fuel consumption of the TBCC increases slowly, which is due to the increase in the drag. In the meantime, the weight and the fuel consumption of the rocket increase almost linearly. Consequently, the total weight of consumption is at a minimum value when the relative thrust of the rocket is about 6%, which is 2.2%

should be considered in the thrust design for the rocket for optimal performance.

less than the total weight consumed by the vehicle without the rocket.

However, when the rocket works, the η*veh* is improved and the η*enigne* is reduced during the mode transition. Additionally, the η*tot* is optimized, especially in the excessively inefficient region around Ma 4.3. In order to evaluate the effectiveness of the performance optimization, the average vehicle efficiency, engine efficiency, and total efficiency are respectively defined as:

*Energies* **2020**, *13*, 2911 10 of 17

$$\overline{\eta}\_{\text{velt}} = \frac{\int \Delta E\_{\text{veh}} dt}{\int TV cosat} \tag{18}$$

$$\overline{\eta}\_{engine} = \frac{\int TVcosat}{\int h\_{PR}\dot{m}\_f dt} \tag{19}$$

$$\overline{\eta}\_{\text{tot}} = \frac{\int \Delta E\_{\text{cell}} dt}{\int h\_{\text{PR}} \dot{m}\_f dt} \tag{20}$$

where . *m<sup>f</sup>* denotes the fuel mass flow rate.

These variables in Equations (18)–(20) are defined in Section 2. The η*veh* increases by 9.9% (from 0.415 to 0.456) after the introduction of the rocket, and the η*engine* is reduced by 6.8%, resulting in an increase of 2.6% in the η*tot*. Therefore, introducing a suitable rocket to assist the TBCC increases the vehicle efficiency in the mode transition, which results in an increase in the total efficiency despite the reduction in engine efficiency. **Figure 8.** (**a**) Vehicle efficiency profiles; (**b**) Engine efficiency profiles; (**c**) Total efficiency profiles.

It seems that the greater the thrust of the rocket, the better the vehicle efficiency. However, the increase in rocket thrust will inevitably bring increased drag and additional weight. The two factors should be considered in the thrust design for the rocket for optimal performance. It seems that the greater the thrust of the rocket, the better the vehicle efficiency. However, the increase in rocket thrust will inevitably bring increased drag and additional weight. The two factors

In Figure 9, when the rocket thrust is relatively small (*RTmax* < 6%), as the thrust of the rocket increases, the fuel consumption of the TBCC reduces rapidly. Even if the fuel consumption and the weight of the rocket are increased, the total weight of consumption is still reduced. It is noted that the total weight in Figure 9 is the sum of the fuel consumption of the TBCC, the fuel consumption of the rocket, and the weight of the rocket. As the thrust increases (*RTmax* > 6%), the fuel consumption of the TBCC increases slowly, which is due to the increase in the drag. In the meantime, the weight and the fuel consumption of the rocket increase almost linearly. Consequently, the total weight of consumption is at a minimum value when the relative thrust of the rocket is about 6%, which is 2.2% less than the total weight consumed by the vehicle without the rocket. should be considered in the thrust design for the rocket for optimal performance. In Figure 9, when the rocket thrust is relatively small (*RTmax* < 6%), as the thrust of the rocket increases, the fuel consumption of the TBCC reduces rapidly. Even if the fuel consumption and the weight of the rocket are increased, the total weight of consumption is still reduced. It is noted that the total weight in Figure 9 is the sum of the fuel consumption of the TBCC, the fuel consumption of the rocket, and the weight of the rocket. As the thrust increases (*RTmax* > 6%), the fuel consumption of the TBCC increases slowly, which is due to the increase in the drag. In the meantime, the weight and the fuel consumption of the rocket increase almost linearly. Consequently, the total weight of consumption is at a minimum value when the relative thrust of the rocket is about 6%, which is 2.2% less than the total weight consumed by the vehicle without the rocket.

**Figure 9.** Rocket *RTmax* vs. weight consumption. **Figure 9.** Rocket *RTmax* vs. weight consumption.

According to Equations (18) and (19), Figure 10 shows the variations in the average efficiencies with the rocket thrust. With the rocket *RTmax* increasing from 0% to 8%, the η*veh* has a sharp increase (9.6%) while the η*engine* drops by about 6%, which results in a 3% increase in the η*tot*. With the rocket *RTmax* further increasing (*RTmax* ≥ 8%), the η*veh* shows a slightly decreasing trend and the η*engine* keeps decreasing, so the η*tot* starts to decrease. According to Equations (18) and (19), Figure 10 shows the variations in the average efficiencies with the rocket thrust. With the rocket *RTmax* increasing from 0% to 8%, the ̅ℎ has a sharp increase (9.6%) while the ̅ drops by about 6%, which results in a 3% increase in the ̅. With the rocket *RTmax* further increasing (*RTmax* ≥ 8%), the ̅ℎ shows a slightly decreasing trend and the ̅ keeps decreasing, so the ̅ starts to decrease.

*Energies* **2020**, *13*, 2911 11 of 17

**Figure 10.** Rocket *RTmax* vs. average efficiencies. **Figure 10.** Rocket *RTmax* vs. average efficiencies.

With the assistance of the rocket, the time spent and the energy consumption of drag in the mode transition could be reduced significantly. Therefore, introducing an appropriate rocket to assist the TBCC-powered vehicle could increase vehicle efficiency in the mode transition, which results in an increase in total efficiency despite the reduction of engine efficiency. For a specified TBCC-powered vehicle, the thrust of the rocket has an optimal value to maximize the average total efficiency. The rocket with an overly small thrust could not improve the average vehicle efficiency sufficiently, but an overly large thrust might lead to excessive drag. With the assistance of the rocket, the time spent and the energy consumption of drag in the mode transition could be reduced significantly. Therefore, introducing an appropriate rocket to assist the TBCC-powered vehicle could increase vehicle efficiency in the mode transition, which results in an increase in total efficiency despite the reduction of engine efficiency. For a specified TBCC-powered vehicle, the thrust of the rocket has an optimal value to maximize the average total efficiency. The rocket with an overly small thrust could not improve the average vehicle efficiency sufficiently, but an overly large thrust might lead to excessive drag.

#### **5. Parameter Study and Integrated Design 5. Parameter Study and Integrated Design**

As mentioned previously, the performance of a specified TBCC-powered vehicle could be optimized by introducing the rocket. The essence of the optimization is the tradeoff between the efficiency and engine efficiency. Considering that vehicle drag and rocket Isp are respectively related to vehicle efficiency and engine efficiency, the optimization of the thrust and operating time of the rocket should involve the influence of vehicle drag and rocket Isp. The thrust, Isp, and operating time of the rocket would determine the combination form ("T+RBCC" or "R+TBCC") of the rocket and TBCC. In this section, the study of the vehicle drag and the rocket Isp is conducted, and the combination form for the rocket and TBCC is also discussed. As mentioned previously, the performance of a specified TBCC-powered vehicle could be optimized by introducing the rocket. The essence of the optimization is the tradeoff between the efficiency and engine efficiency. Considering that vehicle drag and rocket Isp are respectively related to vehicle efficiency and engine efficiency, the optimization of the thrust and operating time of the rocket should involve the influence of vehicle drag and rocket Isp. The thrust, Isp, and operating time of the rocket would determine the combination form ("T+RBCC" or "R+TBCC") of the rocket and TBCC. In this section, the study of the vehicle drag and the rocket Isp is conducted, and the combination form for the rocket and TBCC is also discussed.

Figure 11 shows the total weight consumption as a function of rocket thrust under different drag coefficients. To represent vehicles with different drag coefficients, AF<sup>D</sup> is introduced in the present study and defined as the amplification factor of the drag coefficient of a specified vehicle to that of the baseline vehicle. Under an AF<sup>D</sup> of 1.0 and 1.1, the TBCC engines could propel the vehicle to accelerate in the wide-speed range without a rocket. As the AF<sup>D</sup> is increased, the TBCC thrust is unable to overcome the drag in the thrust pinch regions. Consequently, the rocket is introduced to augment the thrust. For each TBCC-powered vehicle, the thrust of the rocket has an optimal value to minimize the total weight consumption. With the increase in the AFD, the minimum value and the optimal value of the relative thrust both increase. Although the introduction of rockets could propel the vehicle to accelerate in the wide-speed range, the total weight consumption in the process of acceleration increases with the increase in drag. Figure 11 shows the total weight consumption as a function of rocket thrust under different drag coefficients. To represent vehicles with different drag coefficients, AF<sup>D</sup> is introduced in the present study and defined as the amplification factor of the drag coefficient of a specified vehicle to that of the baseline vehicle. Under an AF<sup>D</sup> of 1.0 and 1.1, the TBCC engines could propel the vehicle to accelerate in the wide-speed range without a rocket. As the AF<sup>D</sup> is increased, the TBCC thrust is unable to overcome the drag in the thrust pinch regions. Consequently, the rocket is introduced to augment the thrust. For each TBCC-powered vehicle, the thrust of the rocket has an optimal value to minimize the total weight consumption. With the increase in the AFD, the minimum value and the optimal value of the relative thrust both increase. Although the introduction of rockets could propel the vehicle to accelerate in the wide-speed range, the total weight consumption in the process of acceleration increases with the increase in drag.

*Energies* **2020**, *13*, 2911 12 of 17

*Energies* **2020**, *13*, 2911 12 of 17

**Figure 11.** Effect of rocket *RTmax* on total weight consumption under different AFD. **Figure 11.** Effect of rocket *RTmax* on total weight consumption under different AFD. **Figure 11.** Effect of rocket *RTmax* on total weight consumption under different AFD.

The optimal relative thrust of the rocket is presented in Figure 12 under various AF<sup>D</sup> and Isp. In the low-AF<sup>D</sup> cases (such as AF<sup>D</sup> = 1, 1.1), the optimal relative thrust is insensitive to the Isp. A possible reason is that the rocket works for a short time and the consumption is relatively small. Compared to the total fuel consumption, the fuel reduction caused by the increase in the rocket Isp could be negligible. In the high-AF<sup>D</sup> cases (such as AF<sup>D</sup> = 1.3, 1.4), the changes in the optimal relative thrust with Isp are obvious, because the operating time of the rocket is too long to ignore the impact of Isp. The higher the Isp, the less the reduction in the engine efficiency caused by the rocket, and the more the increase in vehicle efficiency. Thus, the optimal relative thrust increases to satisfy the requirement of higher vehicle efficiency. The results suggest that it is not necessary to pay much attention to the Isp of the rocket in the case of low drag while the Isp is important in the case of high drag. The optimal relative thrust of the rocket is presented in Figure 12 under various AF<sup>D</sup> and Isp. In the low-AF<sup>D</sup> cases (such as AF<sup>D</sup> =1, 1.1), the optimal relative thrust is insensitive to the Isp. A possible reason is that the rocket works for a short time and the consumption is relatively small. Compared to the total fuel consumption, the fuel reduction caused by the increase in the rocket Isp could be negligible. In the high-AF<sup>D</sup> cases (such as AF<sup>D</sup> = 1.3, 1.4), the changes in the optimal relative thrust with Isp are obvious, because the operating time of the rocket is too long to ignore the impact of Isp. The higher the Isp, the less the reduction in the engine efficiency caused by the rocket, and the more the increase in vehicle efficiency. Thus, the optimal relative thrust increases to satisfy the requirement of higher vehicle efficiency. The results suggest that it is not necessary to pay much attention to the Isp of the rocket in the case of low drag while the Isp is important in the case of high drag. The optimal relative thrust of the rocket is presented in Figure 12 under various AF<sup>D</sup> and Isp. In the low-AF<sup>D</sup> cases (such as AF<sup>D</sup> = 1, 1.1), the optimal relative thrust is insensitive to the Isp. A possible reason is that the rocket works for a short time and the consumption is relatively small. Compared to the total fuel consumption, the fuel reduction caused by the increase in the rocket Isp could be negligible. In the high-AF<sup>D</sup> cases (such as AF<sup>D</sup> = 1.3, 1.4), the changes in the optimal relative thrust with Isp are obvious, because the operating time of the rocket is too long to ignore the impact of Isp. The higher the Isp, the less the reduction in the engine efficiency caused by the rocket, and the more the increase in vehicle efficiency. Thus, the optimal relative thrust increases to satisfy the requirement of higher vehicle efficiency. The results suggest that it is not necessary to pay much attention to the Isp of the rocket in the case of low drag while the Isp is important in the case of high drag.

**Figure 12.** Optimal rocket *RTmax* under various AF<sup>D</sup> and rocket Isp. **Figure 12.** Optimal rocket *RTmax* under various AF<sup>D</sup> and rocket Isp. **Figure 12.** Optimal rocket *RTmax* under various AF<sup>D</sup> and rocket Isp.

Figure 13 shows the thrust profiles of the rocket under different AF<sup>D</sup> over the entire Mach range. To compare the thrust, the combined variable of AF<sup>D</sup> and rocket relative thrust *RT* is on the horizontal axis. It is to be noted that the *RT* is the product of the maximum relative thrust, *RTmax*, and the throttle, *Thr*. The distance from the solid line to the dotted line represents the relative thrust of the rocket. With the increase in AFD, the operating range of the rocket expands in transonic and mode transition, and the value of the optimal thrust increases. This could be explained in that the expansion and deepening of the thrust gaps with the increase in the AF<sup>D</sup> call for a bigger thrust and a longer Figure 13 shows the thrust profiles of the rocket under different AF<sup>D</sup> over the entire Mach range. To compare the thrust, the combined variable of AF<sup>D</sup> and rocket relative thrust *RT* is on the horizontal axis. It is to be noted that the *RT* is the product of the maximum relative thrust, *RTmax*, and the throttle, *Thr*. The distance from the solid line to the dotted line represents the relative thrust of the rocket. With the increase in AFD, the operating range of the rocket expands in transonic and mode transition, and the value of the optimal thrust increases. This could be explained in that the expansion and deepening of the thrust gaps with the increase in the AF<sup>D</sup> call for a bigger thrust and a longer Figure 13 shows the thrust profiles of the rocket under different AF<sup>D</sup> over the entire Mach range. To compare the thrust, the combined variable of AF<sup>D</sup> and rocket relative thrust *RT* is on the horizontal axis. It is to be noted that the *RT* is the product of the maximum relative thrust, *RTmax*, and the throttle, *Thr*. The distance from the solid line to the dotted line represents the relative thrust of the rocket. With the increase in AFD, the operating range of the rocket expands in transonic and mode transition, and the value of the optimal thrust increases. This could be explained in that the expansion and deepening of the thrust gaps with the increase in the AF<sup>D</sup> call for a bigger thrust and a longer operation

of the rocket could satisfy the thrust requirement.

1.56.

time of the rocket. In addition, under each AFD, the rocket almost operates at the corresponding maximum thrust constantly in the operating range. It means the fixed-thrust design of the rocket could satisfy the thrust requirement. *Energies* **2020**, *13*, 2911 13 of 17 corresponding maximum thrust constantly in the operating range. It means the fixed-thrust design of the rocket could satisfy the thrust requirement.

*Energies* **2020**, *13*, 2911 13 of 17

**Figure 13.** Rocket *RT* profiles at different AF<sup>D</sup> (Isp = 300 s). **Figure 13.** Rocket *RT* profiles at different AF<sup>D</sup> (Isp = 300 s). Mach range. Under the AF<sup>D</sup> of 1.1 but different Isp, the maximum values of the thrust are approximately equal. The higher the Isp, the less the reduction in the engine efficiency caused by the

Similarly, Figure 14 shows the thrust profiles of the rocket under different Isp over the entire Mach range. Under the AF<sup>D</sup> of 1.1 but different Isp, the maximum values of the thrust are approximately equal. The higher the Isp, the less the reduction in the engine efficiency caused by the rocket. Therefore, the operating ranges of the rocket are widened. In the case of Isp = 240 s, the rocket hardly works in the transonic, while in the case of Isp = 360 s, it operates in the range of Mach 1.18 to Similarly, Figure 14 shows the thrust profiles of the rocket under different Isp over the entire Mach range. Under the AF<sup>D</sup> of 1.1 but different Isp, the maximum values of the thrust are approximately equal. The higher the Isp, the less the reduction in the engine efficiency caused by the rocket. Therefore, the operating ranges of the rocket are widened. In the case of Isp = 240 s, the rocket hardly works in the transonic, while in the case of Isp = 360 s, it operates in the range of Mach 1.18 to 1.56. rocket. Therefore, the operating ranges of the rocket are widened. In the case of Isp = 240 s, the rocket hardly works in the transonic, while in the case of Isp = 360 s, it operates in the range of Mach 1.18 to 1.56. No matter what AF<sup>D</sup> and Isp are, the thrust is basically equal to the two values of 0 and maximum. This indicates that the thrust adjustment of the rocket is not necessary, and the fixedthrust design could satisfy the thrust requirement and simplify the structure of the rocket.

**Figure 14.** Rocket *RT* profiles under different rocket Isp (AF<sup>D</sup> = 1.1). **Figure 14.** Rocket *RT* profiles under different rocket Isp (AF<sup>D</sup> = 1.1).

**Figure 14.** Rocket *RT* profiles under different rocket Isp (AF<sup>D</sup> = 1.1). Figure 15 shows the average vehicle efficiency, ̅ℎ, and the average engine efficiency, ̅, corresponding to different Isp and AFD. With the increase in Isp, the ̅ℎ increases due to the expansion of the rocket operating range, while the ̅ is roughly constant after including the No matter what AF<sup>D</sup> and Isp are, the thrust is basically equal to the two values of 0 and maximum. This indicates that the thrust adjustment of the rocket is not necessary, and the fixed-thrust design could satisfy the thrust requirement and simplify the structure of the rocket.

Figure 15 shows the average vehicle efficiency, ̅ℎ, and the average engine efficiency, ̅, corresponding to different Isp and AFD. With the increase in Isp, the ̅ℎ increases due to the expansion of the rocket operating range, while the ̅ is roughly constant after including the effect of the improved engine efficiency of the rocket. When AF<sup>D</sup> = 1.4, as the Isp increases from 240 s to 360 s, the ̅ℎ is improved from 0.368 to 0.453, an increase of 23.1%. By comparison, when AF<sup>D</sup> = 1.0, the Isp has little effect on the ̅ and the ̅ℎ, because the operating range of the rocket is very small. With the increase in AFD, the ̅ℎ decreases first and then increases. In low-AF<sup>D</sup> cases, effect of the improved engine efficiency of the rocket. When AF<sup>D</sup> = 1.4, as the Isp increases from 240 s to 360 s, the ̅ℎ is improved from 0.368 to 0.453, an increase of 23.1%. By comparison, when AF<sup>D</sup> = 1.0, the Isp has little effect on the ̅ and the ̅ℎ, because the operating range of the rocket is very small. With the increase in AFD, the ̅ℎ decreases first and then increases. In low-AF<sup>D</sup> cases, the TBCC provides most of the thrust for acceleration, and the rocket works within a narrow range, serving as auxiliary power. The increase in drag decreases the ̅ℎ. By comparison, in high-AF<sup>D</sup> cases, the TBCC works independently in a narrow range. The rocket widens the operating range and Figure 15 shows the average vehicle efficiency, η*veh*, and the average engine efficiency, η*engine*, corresponding to different Isp and AFD. With the increase in Isp, the η*veh* increases due to the expansion of the rocket operating range, while the η*engine* is roughly constant after including the effect of the improved engine efficiency of the rocket. When AF<sup>D</sup> = 1.4, as the Isp increases from 240 s to 360 s, the η*veh* is improved from 0.368 to 0.453, an increase of 23.1%. By comparison, when AF<sup>D</sup> = 1.0, the Isp has little effect on the η*engine* and the η*veh*, because the operating range of the rocket is very small. With the increase in AFD, the η*veh* decreases first and then increases. In low-AF<sup>D</sup> cases, the TBCC provides most of the thrust for acceleration, and the rocket works within a narrow range, serving as

auxiliary power. The increase in drag decreases the η*veh*. By comparison, in high-AF<sup>D</sup> cases, the TBCC works independently in a narrow range. The rocket widens the operating range and increase the thrust of itself, and becomes a great part of the propulsion system. Due to the increase in rocket thrust, the η*veh* increases. *Energies* **2020**, *13*, 2911 14 of 17 increase the thrust of itself, and becomes a great part of the propulsion system. Due to the increase in rocket thrust, the ̅ℎ increases. increase the thrust of itself, and becomes a great part of the propulsion system. Due to the increase in rocket thrust, the ̅ℎ increases.

**Figure 15.** Effect of rocket Isp and vehicle drag on ̅ and ̅. **Figure 15.** Effect of rocket Isp and vehicle drag on η*engine* and η*veh*. is widened and the EDrag is reduced. When AF<sup>D</sup> = 1.4, as the Isp increases from 240 s to 360 s, the ̅ is improved from 0.066 to 0.079, namely an increase of 19.7%, and the EDrag is decreased from 35.37

Figure 16 shows the average total efficiency, η̅,and the total energy consumption of drag, EDrag, corresponding to different Isp and AFD. With the increase in Isp, the ̅ increases, because the ̅ℎ increases, and the ̅ remains almost unchanged. Meanwhile, the operating range of the rocket is widened and the EDrag is reduced. When AF<sup>D</sup> = 1.4, as the Isp increases from 240 s to 360 s, the ̅ is improved from 0.066 to 0.079, namely an increase of 19.7%, and the EDrag is decreased from 35.37 GJ to 24.71 GJ, namely a decrease of 30%. The increase in AF<sup>D</sup> decreases the ̅. Moreover, with the increase in AFD, the EDrag increases Figure 16 shows the average total efficiency, η*tot*, and the total energy consumption of drag, *EDrag*, corresponding to different Isp and AFD. With the increase in Isp, the η*tot* increases, because the η*veh* increases, and the η*engine* remains almost unchanged. Meanwhile, the operating range of the rocket is widened and the *EDrag* is reduced. When AF<sup>D</sup> = 1.4, as the Isp increases from 240 s to 360 s, the η*tot* is improved from 0.066 to 0.079, namely an increase of 19.7%, and the *EDrag* is decreased from 35.37 GJ to 24.71 GJ, namely a decrease of 30%. GJ to 24.71 GJ, namely a decrease of 30%. The increase in AF<sup>D</sup> decreases the ̅. Moreover, with the increase in AFD, the EDrag increases first and then decreases. The role played by the rocket accounts for this trend of EDrag*.* When the rocket serves as an auxiliary power at low AFD, the increase in AF<sup>D</sup> decreases the acceleration, resulting in an increase in the *EDrag*. When the rocket serves as a great part of the propulsion system at high AFD, the increase in AF<sup>D</sup> widens the operating range of the rocket and increases the thrust of the rocket, which results in a decrease in the *EDrag*.

**Figure 16.** Effect of rocket Isp and vehicle drag on ̅ and . **Figure 16.** <sup>E</sup>ffect of rocket Isp and vehicle drag on <sup>η</sup>*tot* and *<sup>E</sup>Drag*.

The increase in AF<sup>D</sup> decreases the η*tot*. Moreover, with the increase in AFD, the *EDrag* increases first and then decreases. The role played by the rocket accounts for this trend of EDrag. When the rocket serves as an auxiliary power at low AFD, the increase in AF<sup>D</sup> decreases the acceleration, resulting in an increase in the *EDrag*. When the rocket serves as a great part of the propulsion system at high AFD,

**Figure 16.** Effect of rocket Isp and vehicle drag on ̅ and .

is increased.

the increase in AF<sup>D</sup> widens the operating range of the rocket and increases the thrust of the rocket, which results in a decrease in the *EDrag*. rocket could serve as an additional booster, whose structure is simple. The combination of the rocket and TBCC engine is similar to the "R+TBCC". By contrast, in the high-drag area, the operating time of the rocket is long and the optimal relative thrust is large. The Isp has a significant impact on the

concluded that the improvement of the Isp of the small rocket is not necessary. In this case, the small

*Energies* **2020**, *13*, 2911 15 of 17

Figure 17 presents the total weight consumption ratio and flight time under different Isp and AFD. With the increase in Isp, the total weight consumption and the flight time are gradually reduced, because the ̅ increases and the *EDrag* decreases. As the AF<sup>D</sup> increases, the total weight consumption

The effect of Isp on weight consumption and flight time varies considerably under different AFD. Therefore, this network map can be roughly divided into two areas: low-drag area and high-drag area. In the low-drag area, the rocket works for a short time and the optimal relative thrust is small.

Figure 17 presents the total weight consumption ratio and flight time under different Isp and AFD. With the increase in Isp, the total weight consumption and the flight time are gradually reduced, because the η*tot* increases and the *EDrag* decreases. As the AF<sup>D</sup> increases, the total weight consumption is increased. flight time and the total fuel consumption. It is necessary to pay attention to the Isp of the big rocket. Introducing a big rocket as a booster might bring a large base drag. Instead, the rocket should be integrated into the TBCC engine, such as an RBCC or ERJs, which could deliver augmented thrust at an Isp performance of up to twice or triple that of a rocket through proper design. The combination of the rocket and TBCC engine is similar to the "T+RBCC".

**Figure 17.** Total weight consumption ratio and flight time under different rocket Isp and AFD. **Figure 17.** Total weight consumption ratio and flight time under different rocket Isp and AFD.

According to the parameter study of the vehicle drag and rocket Isp, two design recommendations for rocket-augmented TBCC engines corresponding to the low-drag area and the high-drag area are as follows. In the low-drag area, the rocket works for a short time and its optimal relative thrust is small, so optimizing the Isp of the small rocket is not necessary. The TBCC combined with a booster rocket could be a more simple and suitable scheme. By comparison, in the high-drag area, the operating time of the rocket is long and the optimal relative thrust is large. The Isp has a significant impact on the flight time and the total fuel consumption. Therefore, combining an RBCC (or ERJs) and a turbine could produce higher economic performance. **6. Conclusions** In this study, an integrated analysis method for rocket-augmented TBCC-powered vehicles was proposed based on the trajectory optimization method of the Gauss pseudospectral. The trajectories of the baseline vehicle with and without a rocket were analyzed from the aspect of efficiency and energy. The combination form for rockets and TBCC was discussed with the parameter study of The effect of Isp on weight consumption and flight time varies considerably under different AFD. Therefore, this network map can be roughly divided into two areas: low-drag area and high-drag area. In the low-drag area, the rocket works for a short time and the optimal relative thrust is small. The consumption of the rocket is relatively small compared to the total fuel consumption. It is concluded that the improvement of the Isp of the small rocket is not necessary. In this case, the small rocket could serve as an additional booster, whose structure is simple. The combination of the rocket and TBCC engine is similar to the "R+TBCC". By contrast, in the high-drag area, the operating time of the rocket is long and the optimal relative thrust is large. The Isp has a significant impact on the flight time and the total fuel consumption. It is necessary to pay attention to the Isp of the big rocket. Introducing a big rocket as a booster might bring a large base drag. Instead, the rocket should be integrated into the TBCC engine, such as an RBCC or ERJs, which could deliver augmented thrust at an Isp performance of up to twice or triple that of a rocket through proper design. The combination of the rocket and TBCC engine is similar to the "T+RBCC".

vehicle drag and rocket Isp. Accordingly, the following conclusions were obtained. According to the parameter study of the vehicle drag and rocket Isp, two design recommendations for rocket-augmented TBCC engines corresponding to the low-drag area and the high-drag area are as follows. In the low-drag area, the rocket works for a short time and its optimal relative thrust is small, so optimizing the Isp of the small rocket is not necessary. The TBCC combined with a booster rocket could be a more simple and suitable scheme. By comparison, in the high-drag area, the operating time of the rocket is long and the optimal relative thrust is large. The Isp has a significant impact on the flight time and the total fuel consumption. Therefore, combining an RBCC (or ERJs) and a turbine could produce higher economic performance.

#### **6. Conclusions**

In this study, an integrated analysis method for rocket-augmented TBCC-powered vehicles was proposed based on the trajectory optimization method of the Gauss pseudospectral. The trajectories of the baseline vehicle with and without a rocket were analyzed from the aspect of efficiency and energy. The combination form for rockets and TBCC was discussed with the parameter study of vehicle drag and rocket Isp. Accordingly, the following conclusions were obtained.


**Author Contributions:** Conceptualization, F.G. and J.Z.; methodology, F.G. and J.Z.; software, F.G.; validation, F.G. and W.L.; formal analysis, F.G. and J.Z.; investigation, F.G.; data curation, F.G.; writing—original draft preparation, F.G.; writing—review and editing, F.G., W.L. and J.Z.; visualization, F.G.; supervision, F.X.; project administration, J.Z. and Y.Y.; resources, Y.Y.; funding acquisition, Y.Y. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was funded by the National Natural Science Foundation of China (No. 51906208 and No. 91941103), the Equipment Exploratory Research Fund (No. 61402060301), the Aeronautics Power Foundation (No. 6141B090325) and the Weapon Exploratory Research Fund (No. 6141B010266).

**Acknowledgments:** The team members of School of Aerospace Engineering of Xiamen University are gratefully acknowledged. In addition, Feng Guo wants to thank the care and support from Shengnan Dai over the passed years.

**Conflicts of Interest:** The authors declare no conflicts of interest.

## **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Review*

## **A Review on Liquefied Natural Gas as Fuels for Dual Fuel Engines: Opportunities, Challenges and Responses**

#### **Md Arman Arefin <sup>1</sup> , Md Nurun Nabi 2,\*, Md Washim Akram <sup>3</sup> , Mohammad Towhidul Islam <sup>1</sup> and Md Wahid Chowdhury <sup>1</sup>**


Received: 22 October 2020; Accepted: 20 November 2020; Published: 23 November 2020

**Abstract:** Climate change and severe emission regulations in many countries demand fuel and engine researchers to explore sustainable fuels for internal combustion engines. Natural gas could be a source of sustainable fuels, which can be produced from renewable sources. This article presents a complete overview of the liquefied natural gas (LNG) as a potential fuel for diesel engines. An interesting finding from this review is that engine modification and proper utilization of LNG significantly improve system efficiency and reduce greenhouse gas (GHG) emissions, which is extremely helpful to sustainable development. Moreover, some major recent researches are also analyzed to find out drawbacks, advancement and future research potential of the technology. One of the major challenges of LNG is its higher flammability that causes different fatal hazards and when using in dual-fuel engine causes knock. Though researchers have been successful to find out some ways to overcome some challenges, further research is necessary to reduce the hazards and make the fuel more effective and environment-friendly when using as a fuel for a diesel engine.

**Keywords:** liquefied natural gas; diesel engine; greenhouse gas emissions; sustainable development

## **1. Introduction**

Recently, it is projected that global energy demand has been increasing day by day [1,2]. The increased amount of energy demand produces a large amount of greenhouse gases (GHGs) by the burning of fossil fuels, which ultimately causes global warming. Currently, in the industrial and transportation sectors, diesel is mainly used as fossil fuel. The consumption of diesel fuel is increasing day by day unexpectedly because of the dramatic increase in vehicles, mostly in Asian countries like Bangladesh, Korea, India and China [3]. Currently, researchers all over the world are concerned about the way of mitigating this large amount of energy demand and at the same time, carbon dioxide (CO2) emission reduction, which is one of the major components of GHGs [4,5]. In this regards, other sources of fuel may be a feasible alternative to conventional fuels. Natural gas would be a promising alternative fuel source in the transportation sector because of some of its remarkable advantages. Though natural gas is also derived from fissile resources, it can be converted from renewable sources, i.e., it can be produced through the biomass conversion process (biomethane, which is also known as biogas, is a pipeline-quality gas made from organic matter), attractive cost, better combustion efficiency and greenhouse gas reduction, which are the significant advantages as alternative fuel [3].

Besides, the interest in alternative fuels is rapidly growing because of the energy security concern all over the world. Among the candidates of alternative fuel, biofuels, liquefied petroleum gas (LPG) and liquefied natural gas (LNG) are the potential ones. Nevertheless, the economic aspects and availability make LPG and LNG more realistic solutions compared to biofuels. Additionally, natural gas can be considered as one of the solutions to control engine emissions at present as compared to traditional fuels. In a homogenous charge compression ignition engine (HCCI), the after-treatment technique combined with natural gas present an outstanding potential to meet strict requirements for reducing the emissions [6]. In the transportation sector, both forms of natural gas, i.e., liquefied natural gas (LNG) and compressed natural gas (CNG) can be used as an alternative fuel. However, CNG has already gained popularity in the automobile sector due to its lower cost. Besides, the interest in alternative fuels is rapidly growing because of the energy security concern all over the world. Among the candidates of alternative fuel, biofuels, liquefied petroleum gas (LPG) and liquefied natural gas (LNG) are the potential ones. Nevertheless, the economic aspects and availability make LPG and LNG more realistic solutions compared to biofuels. Additionally, natural gas can be considered as one of the solutions to control engine emissions at present as compared to traditional fuels. In a homogenous charge compression ignition engine (HCCI), the after-treatment technique combined with natural gas present an outstanding potential to meet strict requirements for reducing the emissions [6]. In the transportation sector, both forms of natural gas, i.e., liquefied natural gas (LNG) and compressed natural gas (CNG) can be used as an alternative fuel. However, CNG has already gained popularity in the automobile sector due to its lower cost.

LNG is mainly used for electricity production and transportation. The main advantages offered by LNG are higher safety, easier transportation and storage capacity compared to CNG [3]. LNG is cleaner than coal and oil, therefore, it has got a plethora of recognition in the global market [7]. Since LNG is clean, it is a promising alternative to diesel vehicles and is capable of compensating some of the severe drawbacks of natural gas vehicles; for instance, LNG fuelled trucks have a higher range (up to 700–1000 km) due to higher energy density [8]. However, while considering LNG as an alternative, the economic viability should also be taken into consideration. The cost of an LNG fuel tank and the engine is approximately twice as high as a diesel engine and tank [9]. Consequently, during this decade, the main contributors to LNG supply growth will be Australia and USA, though 18 other countries have already joined and others have a short and long term goal to join the industry [10]. Besides, among various alternatives of natural gas, LNG is the most preferred mode for long distance transportation as because of its liquefaction attribute, its volume reduces by a factor of 600 [11]. Compared to piped natural gas (PNG), the annual growth of LNG trade is two times higher, which accounts for 10% and 31% of global natural gas consumption and trade respectively [8]. According to recent study reports, from 2005 to 2015, Europe, Asia and Oceania were the primary recipients of LNG imports consisted of 90% of global imports [12]. LNG is mainly used for electricity production and transportation. The main advantages offered by LNG are higher safety, easier transportation and storage capacity compared to CNG [3]. LNG is cleaner than coal and oil, therefore, it has got a plethora of recognition in the global market [7]. Since LNG is clean, it is a promising alternative to diesel vehicles and is capable of compensating some of the severe drawbacks of natural gas vehicles; for instance, LNG fuelled trucks have a higher range (up to 700–1000 km) due to higher energy density [8]. However, while considering LNG as an alternative, the economic viability should also be taken into consideration. The cost of an LNG fuel tank and the engine is approximately twice as high as a diesel engine and tank [9]. Consequently, during this decade, the main contributors to LNG supply growth will be Australia and USA, though 18 other countries have already joined and others have a short and long term goal to join the industry [10]. Besides, among various alternatives of natural gas, LNG is the most preferred mode for long distance transportation as because of its liquefaction attribute, its volume reduces by a factor of 600 [11]. Compared to piped natural gas (PNG), the annual growth of LNG trade is two times higher, which accounts for 10% and 31% of global natural gas consumption and trade respectively [8]. According to recent study reports, from 2005 to 2015, Europe, Asia and Oceania were the primary recipients of LNG imports consisted of 90% of global imports [12].

#### *1.1. Economics and Life Cycle of LNG 1.1. Economics and Life Cycle of LNG*

Figure 1 shows the worldwide LNG price as of June 2019. South Korea and China had the highest landed price of LNG in the world. The price received at the regasification plant is referred to as the landed price. Netback price is taken into consideration for the determination of these prices, which is based on the effective price for a seller and producer at a definite location. Figure 1 shows the worldwide LNG price as of June 2019. South Korea and China had the highest landed price of LNG in the world. The price received at the regasification plant is referred to as the landed price. Netback price is taken into consideration for the determination of these prices, which is based on the effective price for a seller and producer at a definite location.

**Figure 1.** Worldwide landed price of liquefied natural gas (LNG) [13]. **Figure 1.** Worldwide landed price of liquefied natural gas (LNG) [13].

It is also required to make an economic calculation of the LNG liquefaction plant because fuel imposes the second highest operational cost after labor. There are some reasons behind the cost of It is also required to make an economic calculation of the LNG liquefaction plant because fuel imposes the second highest operational cost after labor. There are some reasons behind the cost of

**Fuel** 

**Type** 

LNG not being as straight forward as diesel, which are: (1) distribution costs are not included in LNG fuel price, which is dissimilar to diesel fuel price; (2) generally state tax rates for LNG is different from other fuels, and it also varies state-wise; (3) federal tax rates for LNG fuel are also different and (4) energy content of different fuels are different [14]. Table 1 shows the national average cost for LNG and diesel fuel. Equivalent fuel cost is calculated in dollars per diesel equivalent liter by using the energy content of the fuel. Diesel equivalent liter refers to the amount of energy, which is equivalent to one-liter diesel fuel. Since the distribution cost is not included in LNG fuel price, thus, it is more competitive. Based on the distribution costs and state taxes, the costs of LNG fuel will vary radically by location to location. According to the Taxpayer Relief Act of 1997, based on the energy content of transportation fuel it has been taxed, such as the federal tax rate on LNG is changed from \$0.0503 (unit conversion) per liter to USD 0.0317 (unit conversion) per liter [14]. content of transportation fuel it has been taxed, such as the federal tax rate on LNG is changed from \$0.0503 (unit conversion) per liter to USD 0.0317 (unit conversion) per liter [14]. **Table 1.** Average LNG as well as diesel fuel cost worldwide (unit conversion) [14]. **Fuel Cost (\$/L) Distribution Cost (\$/L) State Tax (\$/L) Federal Tax (\$/L) Total Fuel Cost (\$/L) Equivalent Fuel Cost (\$/Diesel Equivalent**  Diesel 0.1717 0.0634 0.0555 0.2906 0.2906

*Energies* **2020**, *13*, x FOR PEER REVIEW 3 of 22

fuel price, which is dissimilar to diesel fuel price; (2) generally state tax rates for LNG is different

from other fuels, and it also varies state-wise; (3) federal tax rates for LNG fuel are also different and

(4) energy content of different fuels are different [14]. Table 1 shows the national average cost for

LNG and diesel fuel. Equivalent fuel cost is calculated in dollars per diesel equivalent liter by using

the energy content of the fuel. Diesel equivalent liter refers to the amount of energy, which is

equivalent to one-liter diesel fuel. Since the distribution cost is not included in LNG fuel price, thus,

it is more competitive. Based on the distribution costs and state taxes, the costs of LNG fuel will vary


**Table 1.** Average LNG as well as diesel fuel cost worldwide (unit conversion) [14]. LNG 0.0925 0.0264 0.0476 0.0660 0.2325 0.3936

**Liter)** 

Note: Permission granted from National Renewable Energy Laboratory. extracted raw natural gas is sent to the plant for liquefaction, where LNG is obtained as a byproduct.

Figure 2 depicts a schematic diagram of the LNG life cycle. With the help of the pipeline extracted raw natural gas is sent to the plant for liquefaction, where LNG is obtained as a byproduct. Mainly, the ship is used to carry LNG to the regasification terminal. Meanwhile, a small amount of liquid is drawn off and carried by small tanker trucks to the service station for fuelling LNG trucks. On the other hand, in each transport and storage, phase boil-off gas is produced, which can be fully recovered for the purpose of using as a gaseous fuel [15]. Mainly, the ship is used to carry LNG to the regasification terminal. Meanwhile, a small amount of liquid is drawn off and carried by small tanker trucks to the service station for fuelling LNG trucks. On the other hand, in each transport and storage, phase boil-off gas is produced, which can be fully recovered for the purpose of using as a gaseous fuel [15].

**Figure 2.** LNG life cycle (adapted from [15]). Note: Permission granted from the publisher (Elsevier-**Figure 2.** LNG life cycle (adapted from [15]). Note: Permission granted from the publisher (Elsevier-Applied Energy).

Applied Energy). Although natural gas is benign and available in nature, it is hardly used in a large diesel engine and still limited to a small engine i.e., spark ignition (SI) engine. Natural gas is mainly used in heavy

diesel operated transportation fleets [3]. It is reported that fuelling natural gas with diesel fuel as

compared to fuelling diesel only causes a less lean mixture and a lower the efficiency. This is because

when natural gas is fuelled it tends to mix with air and could move into the cylinder. Hence, this

induced less air leads to an increase in equivalence ratio. Again, when the load is decreased, it causes

lesser inducing of natural gas, which results in reducing the efficiency in fixed load speed condition

and, thus, more natural gas is needed to be aspirated [16]. Besides, knocking is one of the severe

problems of dual-fuel engines [17–19]. According to previous study reports, though advancing the

injection timing in the dual-fuel engine reduces emissions [20], may increase the tendency of noise

diesel operated transportation fleets [3]. It is reported that fuelling natural gas with diesel fuel as compared to fuelling diesel only causes a less lean mixture and a lower the efficiency. This is because when natural gas is fuelled it tends to mix with air and could move into the cylinder. Hence, this induced less air leads to an increase in equivalence ratio. Again, when the load is decreased, it causes lesser inducing of natural gas, which results in reducing the efficiency in fixed load speed condition and, thus, more natural gas is needed to be aspirated [16]. Besides, knocking is one of the severe problems of dual-fuel engines [17–19]. According to previous study reports, though advancing the injection timing in the dual-fuel engine reduces emissions [20], may increase the tendency of noise due to fast pressure rise [21,22]. Thus, natural gas is always used as minor fuel in most of the dual-fuel engine operations [3]. One of the key initiatives to utilize LNG as a fuel in a diesel engine is to alleviate the knocking problem and take adequate precautions when knocking resistance is low [23]. In terms of emissions, LNG provides far better results compared to conventional fuels. For instance, an experiment in the Netherlands showed that for trucks GHG emission will be reduced by 10–15% if LNG is used [24]. With the increase in power, the smoke density of diesel increases sharply while LNG-duel fuel increases slightly. Under high loads, the smoke density of LNG duel fuel decreases by a huge amount when compared with diesel [25].

From the above discussion, it can be said that with the increasing demand for natural gas vehicles, the demand for higher performance, environment-friendly and more efficient engines have also been increased. The perspective of this review is to bring attention to the current position of LNG and the use of it in the diesel engine. Engine performances and emission characteristics by using LNG are also discussed in this study. The consciousness of using environmentally viable and economical energy source is attracting researchers and environmental scientists to concentrate on LNG as fuels to fulfill both energy crisis and to control environmental abnormalities.

#### *1.2. Properties of LNG*

LNG is a mixture of gases, and its liquefaction is done by reducing the temperature below the boiling point. The amount of methane in LNG is about 87–99 mole%, and the remaining portion is ethane, propane and other heavier hydrocarbons depending on different LNG sources [26,27]. For instance, the LNG imported from Belgium contains 90% (by mass) methane and 10% (by mass) ethane [28]. The lower calorific value of LNG is 21 MJ/L, and the higher calorific value is 24 MJ/L at −164 ◦C [29]. To produce LNG, natural gas is refrigerated at −162 ◦C at atmospheric pressure; thus, LNG is known as a cryogenic liquid [30,31]. During liquefaction of natural gas, the primary component, i.e., methane, is cooled below its boiling point. At the same time, the concentrations of oxygen, carbon dioxide, water, hydrocarbons and some sulphur compounds are either removed or reduced in some small extent [32]. Normally, LNG is stored and handled below 1.586 MPa (unit conversion) pressure in extremely cold temperature in the tank [33]. The density of LNG is between 410 and 500 kg/m<sup>3</sup> [29], hence, it is lighter than water [28] and will float if it spilled on water. Per unit volume combustion heat production of LNG is much higher compared to natural gas. At atmospheric conditions, to produce equal energy, natural gas requires 600 times larger volume compared to LNG [28]. Besides, both LNG and its vapor are not explosive when exposed to the unconfined environment [32]. Moreover, LNG is more efficient and feasible for transportation compared to pipeline gases [7,29,34]. Well purified and condensed LNG can be easily transported over the sea [35]. While transporting, to handle the low temperature of LNG, specifically designed double-hulled ships are used [32].

LNG is a non-toxic, non-corrosive, colorless, odorless, safe and clean form of natural gas [7,36]. LNG is non-flammable; therefore, the liquid itself will not burn. However, the vapor of LNG is highly flammable with air, which causes flash fire. The flashpoint of LNG is −187.8 ◦C, but the autoignition temperature is 537 ◦C [37]. The specific gravity of LNG is 0.45 and 0.6 of gas [37] and 2.7488 liters of LNG has 100% of the energy of 3.7854 liters (unit conversion) of diesel [38]. The burning speed of LNG is 0.38 m/s in the stoichiometric mixture. This burning speed is comparatively lower for using only LNG in a diesel engine [37]. Due to the requirement of thermal shield and the lower density of

LNG compared to heavy fuel oil (HFO), the fuel tank required for LNG is 2.5–3 times bigger than HFO tank [39]. During combustion, LNG has almost no SO<sup>2</sup> and particulate matter emission [40]. Furthermore, the life cycle CO<sup>2</sup> emissions of LNG are 18% less than its counterpart gasoline vehicle model [41]. These advantages enable LNG as a potential fuel for the transportation sector. Depending on the liquefaction process and the plant where LNG is produced, the chemical components of LNG vary slightly, i.e., Europe, America and Asia, shown in Table 2 [42]. During designing LNG power plant, the regasification and liquefaction processes are controlled by the phase behavior and thermodynamic properties of LNG. Mokhatab et al. [26] have summarized the multiphase equilibrium of LNG using numerical methods.


**Table 2.** Chemical compositions of LNG imported from various countries all over the world [42].

Note: Reprinted with permission from the publisher (Elsevier-Renewable and Sustainable Energy Reviews).

LNG is not explosive; therefore, to be ignited, first, it must be vaporized and then mixed with air at a proper portion [43]. Nevertheless, in the LNG transportation sector, there are some possibilities of accident, for example, the leakage of LNG into ground and sea, and rollover of the LNG tank [35]. However, LNG vapor only burns when it is mixed in a concentration of 5%–15% with air [7,32]. All these factors make LNG a safe alternative fuel for the transportation sector.

LNG makes two distinct types of gases, depending on the extraction procedure. They are natural boil-off gas and forced boil-off gas. From the top of the fuel tank, natural boil-off gas is collected, which contains a higher amount of methane and a lower amount of nitrogen; thus, its knocking resistance is very high. Investigation predicts the value of methane number (MN) is around 100 and the lower calorific value between 33 and 35 MJ/nm<sup>3</sup> [44].

On the other hand, from the down in the tank, forced boil-off gas is extracted and evaporated separately. Forced boil-off gas contains the mixture of all hydrocarbons of the liquid. The knocking resistance of the derived gas varies from load to load and even from origin to origin with the methane number ranging between 70 and 80. The calorific value of forced boil-off gas is about 38–39 MJ/nm<sup>3</sup> , which is quite high compared to the natural boil-off gas [44]. Besides, the forced boil-off gas is quite stable than the natural boil-off gas. Thus, for general shipping forced boil-off gas is becoming very popular [44]. After LNG being heated by ambient air, boil-off gas is generated from the tank at the lowest boiling point of the constituents. Thus, it is very difficult to keep the LNG compositions unchanged for a long time once it started to boil-off. Since heavy trucks are operated in the same routes thus, LNG is more suitable as the fuel of those engines. Furthermore, the cruising distance will be tripled if LNG is used as a fuel in the vehicle instead of CNG because the energy density of CNG is three times lower than LNG [45].

Table 3 depicts the comparison of the physical properties of LNG, CNG, diesel and gasoline. With the change of chemical compositions, the physical properties of LNG vary like diesel and gasoline. In some cases, like the pump octane number and lower heating value, CNG and LNG show somewhat similar results. It is important to understand the cryogenic property of LNG during using it as a vehicle fuel. The phenomenon "weathering" or "enrichment" arises because natural gas is a chemical mixture

of ethane, nitrogen and methane. LNG contents vary in the percentage of methane as well as other hydrocarbons. Generally, methane content may vary from 92% to 99% [14]. Besides, other hydrocarbon contents in natural gas are ethane 1%–6%, butane 0% to 2%, propane 1%–4% and other compounds [14]. The boiling point of each of the chemical compounds is methane, ethane, propane and butane is −162, −88, −42 and −0.5 ◦C respectively [45]. This is the reason why the chemical compounds in the liquid vaporize at its unique boiling point. Since the boiling point of LNG components are different and it vaporizes at its unique boiling point, the concentration of heavier hydrocarbons of LNG is increased. Therefore, premature ignition occurs, which tends to "knock" due to the higher concentrations of heavier hydrocarbons. To relieve this uncontrolled knocking tendency, it is required to use LNG before it becomes weathered. To distinguish this potential difficulty, during the manufacture of LNG, it should be ensured that the amount of methane is 99.4% or higher [14]. The higher amount of methane reduces the detrimental constituents thus weathering cannot create harmful fuel mixtures in LNG [14].


**Table 3.** Comparison of different properties of LNG with CNG, diesel and gasoline.

## **2. LNG Fuel System**

There are mainly three types of the engine that burns LNG as combustion fuel in the marine industry. First one is the "spark ignited" engine, which uses only LNG as fuel with major requirements, simplicity and good overall performance at the lowest total emissions. This engine is generally used in the marine industry for short-distance ferries [44]. The second one has the dual-fuel capability and mainly used for land power plant because of higher specific power production capacity. The third one is the "direct gas injection diesel engine", which was firstly used in the offshore industry because of its high fuel flexibility and power density [44]. The updated concept of this type of engine in the marine industry has made it too flexible but the use of this type is still limited. The multifuel capability, flexibility in fuel mixing and high-pressure gas injection (300–350 bar) made this engine very promising. This type of engine is unique in terms of having no distinct requirements for stabilization of self-ignition of fuel gas. In addition, its operating principle for diesel provides a complete combustion of the gas fuel, though the NOx emission occurred higher than other gas engine types.

### *2.1. Descriptions of the Additional Components*

Generally, diesel fuel containers are a low-pressure tank and kept uninsulated. However, LNG must be kept in a very well insulated and pressurized tank to minimize vapor loss and weathering because of its cryogenic nature [49]. To utilize LNG in a vehicle, a fuel control system is required, which consists of a pressure management system, vaporizer, fill and vent connections and a cryogenic tank. These devices are pre-plumbed and often integral to the tank. Tank pressure depends on fuel temperature and corresponding vapor pressure. Primarily, the saturated vapor pressure is lower than the fill pressure pumped. However, for a given fuel temperature, the saturation pressure will rapidly drop, and condensation of vapor takes place into the liquid. Most of the systems entirely depend on fuel saturation pressure and temperature, but there are some special devices with systems, which can add heat during lowering the tank pressure [50]. Between the inner and outer shell of the fuel tank, there is a perlite insulation to reduce the heat leakages in the vacuum and to maintain the heat in-between the tanks [51].

All the fuel lines, fittings and valves between LNG fuel tank and the vaporizer must be capable of providing cryogenic services. The connecting lines from the vaporizer to the engine are required only to carry the fuel vapor to the engine inlet. It is recommended to use corrosion-resistant fittings to avoid contamination. Total pressure drops to deliver fuel between the fuel tank and engine inlet must be adequate. The pressure drop testing is necessary to measure the performance and isolate the components if excessive system pressure drop occurs and then, system modification is also necessary [50]. Among various types of vaporizers, the intermediate fluid vaporizer (IFV) shows several advantages over others. For designing the IFV, the heat source must be determined considering the climatic conditions and environmental issues [52].

Besides the cryogenic tank, liquid level and pressure transmitters, venting system, cryogenic filling connector and pressure relief valve are other important components. The main feature of the pressure relief valve is to control the boil-off of the tank. To take the liquid from the tank and then pass it through a pressure regulator and vaporizer to the engine feeding system and injector a withdrawal system is used, which is associated with the cryogenic tank. However, there are some modern engines that are designed to directly inject liquid natural gas in the combustion chamber. Due to having a boil-off valve with a cryogenic container, LNG vehicle should be operated continuously without long parking [53].

#### *2.2. Possible Outcome*

From a long time, natural gas has been used as fuel for private cars. Heavy vehicles such as heavy trucks and buses are also using natural gas recently. Pressure vessel stores natural gas at very high pressure of about 20 MPa. In the transport sectors, natural gas shows numerous benefits, especially where heavy fuels are utilized [28]. Almost 25% CO<sup>2</sup> [54], 85% NOx [28], 98% SOx [28] and 90% particulate matters [28] are reduced during combustion of the engine when natural gas is used in place of diesel fuel. Shortly, heavy fuels containing sulphur will be severely restricted; thus, Western Europe has a considerable interest to use LNG for ship propulsion [28]. Worldwide share of ship emissions to total emissions is increasing day by day. Ship emission produces SO2, CO<sup>2</sup> and NOX, which contributes 4–8%, 2–4% and 10–20% of global emissions, respectively [55]. In the gas-burning mode, LNG vessels result in the elimination of all SO<sup>2</sup> emission and the production of NOX, CO<sup>2</sup> and particulate matter are less compared to a vessel driven by marine diesel [55]. Generally, the greenhouse gas emission of a diesel truck is higher than the LNG truck in the wheel to wheel (WTW) analysis. Figure 3 shows a particularity of source of origin of tank to wheel (TTW) emission factors, where DENA extricated the information from German Consultancy-Ludwig Bölkow Systemtechnik (LBST) and the rest of the sources are from North American experiments [34]. From Figure 4, it can be seen that the total WTW emissions of all gas heavy ground vehicles (HGVs) are significantly lower than diesel HGVs and LNG shows way better results in terms of emissions than CNG and diesel. Besides, the reduction of emission, LNG provides other competitive advantages in night-time services through inner-city by reducing the noise level when using LNG Otto-cycle engines [56].

*Energies* **2020**, *13*, x FOR PEER REVIEW 8 of 22

*Energies* **2020**, *13*, x FOR PEER REVIEW 8 of 22

**Figure 3.** Life cycle greenhouse gas comparison of diesel and LNG heavy-duty vehicle (adapted from [40]). Note: Permission granted from the publisher (Elsevier-Renewable and Sustainable Energy Reviews). **Figure 3.** Life cycle greenhouse gas comparison of diesel and LNG heavy-duty vehicle (adapted from [40]). Note: Permission granted from the publisher (Elsevier-Renewable and Sustainable Energy Reviews). **Figure 3.** Life cycle greenhouse gas comparison of diesel and LNG heavy-duty vehicle (adapted from [40]). Note: Permission granted from the publisher (Elsevier-Renewable and Sustainable Energy Reviews).

**Figure 4.** Wheel to Wheel emissions of heavy goods vehicles, adapted from [57]. **Figure 4.** Wheel to Wheel emissions of heavy goods vehicles, adapted from [57]. **Figure 4.** Wheel to Wheel emissions of heavy goods vehicles, adapted from [57].

For many years, the price of LNG depends on the HFO price though often LNG is cheaper. By considering the lower heating value and price of LNG, it can be said that LNG cost is about 60% of HFO [37]. Due to the savings of the reliquefaction process, the cost of produced boil-off gas is decreasing on gas carriers. Due to introducing shale gas in the US market, the natural gas price, including LNG, decreased in 2009 and 2010. Therefore, LNG has become a strong competitive of HFO. The price of HFOIF-0380 in the middle of 2008 was over \$1000 per metric ton. In the middle of 2011, For many years, the price of LNG depends on the HFO price though often LNG is cheaper. By considering the lower heating value and price of LNG, it can be said that LNG cost is about 60% of HFO [37]. Due to the savings of the reliquefaction process, the cost of produced boil-off gas is decreasing on gas carriers. Due to introducing shale gas in the US market, the natural gas price, including LNG, decreased in 2009 and 2010. Therefore, LNG has become a strong competitive of HFO. The price of HFOIF-0380 in the middle of 2008 was over \$1000 per metric ton. In the middle of 2011, For many years, the price of LNG depends on the HFO price though often LNG is cheaper. By considering the lower heating value and price of LNG, it can be said that LNG cost is about 60% of HFO [37]. Due to the savings of the reliquefaction process, the cost of produced boil-off gas is decreasing on gas carriers. Due to introducing shale gas in the US market, the natural gas price, including LNG, decreased in 2009 and 2010. Therefore, LNG has become a strong competitive of HFO. The price of HFOIF-0380 in the middle of 2008 was over \$1000 per metric ton. In the middle of 2011, it was about \$650 and later, the price of HFO further increased. Therefore, it can be said that the price of LNG is more stable compared to the HFO depending on the industry price [37].

Moreover, LNG is a very pure fuel; therefore, the operational costs of the engine are decreasing, the technical states of the engines are better, and the number of emergencies and failures are dropping. The price of LNG in the next 20 years will be equivalent to HFO, but later the LNG price will fall [37].

To make a proper comparison, different performance parameters of LNG-diesel dual-fuel engine such as torque, fuel consumption, smoke density, brake power and different emissions are needed to be measured and compared with a single diesel fuel engine (Table 4). For LNG-diesel dual-fuel engine, the brake specific fuel consumption (BSFC) is lower under light loads. Compared to the diesel engine, the torque and brake power of LNG-diesel dual-fuel engine remain unchanged [58]. Some researchers estimated brake thermal efficiency of the liquid natural gas engine has the same level of the corresponding conventional diesel engine [45]. Moreover, LNG is more preferable for heavy-duty vehicles compared to CNG as the density of LNG is 435 kg/m<sup>3</sup> as compared to 175 kg/m<sup>3</sup> for CNG at 20 MPa [58].


**Table 4.** Recent research on LNG as a fuel for the diesel engine.


#### **Table 4.** *Cont.*

## *2.3. Challenges*

Though there are many merits of LNG, it also has some challenges to overcome. Frost burn occurs when LNG comes into contact with human skin because of being a cryogenic fluid. This phenomenon is also known as frostbite or cold burn [27,28]. Structural failure may occur due to metal cracks and metal embrittlement of the LNG storage tank, which rises by LNG spill [28]. Besides, different common usable materials such as plastic, carbon steel and rubber become brittle at an enormously cold temperature [27]. Since methane does not sustain breathing, thus, due to a large concentration of methane dispersed in air and vapor accumulated near the ground may rise to asphyxiation [28]. When LNG is released on water, it becomes less dense, which gives rise to rapid vaporization of liquefied natural gas and known as rapid phase transition (RPT) [37,71–73]. Consequently, due to the accidental release of LNG, a strong pressure wave is developed and causes heat radiation and burning clouds. Besides, higher flammability of LNG causes different fatal hazards [28]. LNG vaporizes readily when it is released from containment. Then, the liquid will be heated by the surroundings, thus, causing it to vaporize. The vapor will be mixed with the surrounding air and carried out downward causing a cloud. Eventually, it will be mixed with additional air and will be further diluted. The vapor cloud may ignite if the flammable portion comes in contact with a fire source. Secondary fires can be generated by the burn-backs of the primary vapor cloud and cause severe damage to the persons caught within the cloud [74].

According to a safety study regarding LNG, the maximum effective distance of LNG transportation with trucks is 230 m [28]. Flash fire of LNG causes truck tank rupture, which is related to the distance effect. Boiling liquid expanding vapor explosion (BLEVE) may occur with lethal effects during rupturing an LNG truck tank with elevated pressure [28]. Sometimes because of BLEVE, the released liquid immediately flashes and atomizes the resulting fireball. Though fireball lasts for only a couple of seconds, its effects can be very dangerous [74]. If LNG truck tank rupture occurs during engulfing fuel firing, then the maximum effect distance would be 190 m [28]. Pool fires occur due to immediate ignition of LNG, which can cause several damages to the surrounding equipment and burns to people caught within the cloud. The surface emissive power of LNG pool fires generally lies in the range of <sup>220</sup> <sup>±</sup> 50 kW/m<sup>2</sup> . However, a huge amount of smoke is produced during a large fire, which is very dangerous [27].

Due to having different density, LNG has the potential to layer in unstable strata within the tank. These strata have the potential of rolling over to stabilize the liquid in the tank. Due to normal heat leak, the longer LNG layer gets heated and changes density until it becomes lighter than the upper layer. As a result, a liquid rollover may occur with sudden vaporization of LNG. However, if the design of pressure relief systems is not adequate the excess pressure can result in cracks or other structural failures in the tank [71]. Since LNG is extremely cold, it can cause damage to eyes and tissue. Oxygen deficient hazard (ODH) results from the displace of air caused by the release of LNG [75]. Due to the accidental release of flammable liquid from pressurized containment, the leak takes the form of a spray of liquid droplets and vapors and if it is ignited the resulting fire is called torch fire. Torch fire possesses similar types of hazards as a pool fire. In some cases, for similar size pool and torch fire, the radiant heating power of a torch fire is frequently greater than that of a pool fire [74].

In a high pressure dual fuel engine, the combustion is nearly complete when LNG is injected at a pressure of 30–35 MPa with a small amount of diesel. However, when using in low pressure dual-fuel LNG is injected at a low pressure, which is comparable to the Otto cycle. A major disadvantage of the dual fuel diesel engine is the high amount of methane slip compared to the diesel engine. At high and medium loads the air fuel mixture enters the crevices and cylinder wall causes methane slip [76]. Moreover, at low speed the emission problem is more severe because of a low air-fuel ratio. This phenomenon therefore causes high methane slip and fuel consumption as a result of bulk quenching in the coldest areas of the combustion chamber [77]. In an Otto based dual fuel engine, at high power the methane slip is low, however, with the decrease of power, methane slip increases significantly [78].

#### **3. Possible Solutions**

Researchers all over the world are trying to find out some ways to face the challenges to use LNG and have already found some techniques as discussed below. The safety barriers, i.e., layers of protections, may have been used for preventing, controlling or mitigating undesired accidents or events. Similarly, for preventing the depth of catastrophic accident, different types of safeguards and protection layers have been used. Currently, the LNG industry uses several layers of protection for minimizing and controlling the consequences connected with vapor dispersions, LNG spills and subsequent fires and explosions [27]. To minimize cryogenic embrittlement, different types of special steels have been developed [28]. Moreover, currently, operators are using LNG rollover models, which can optimize boil-off costs by inducing density stratification and by doing so, converting a dangerous configuration into a potentially operational asset [71].

Basic responses to accidental LNG spills include detecting the spill, securing the origin and taking measures to prevent it from worsening. It is very important to secure the leak and the area, move the people away from the spill, prevent ignition sources and monitoring should be performed carefully until no vapor remains in the flammable limits [79]. A potential technique to reduce the size of the flammable vapor cloud is to increase the vapor dispersion generated by the liquefied gas spill [27]. The cloud generated by the spill of unconfined LNG on water travels at the speed of the wind before dispersion. Since it is denser than air, for land-based facilities, it has some advantages i.e., it can be easily controlled, though sometimes it can be a disadvantage if it takes longer to disperse [80]. Some experimental conditions of LNG spills on the water are summarized in Table 5. To determine low flammability limit (LFL), spill rate, volume, vaporization rate and atmospheric condition are taken into account [80]. However, reducing the pool surface area would be an effective solution to reduce LNG vaporization [27]. Water spray curtains may be used as a promising technique to mitigate many toxic and flammable LNG vapors by reducing the concentration of LNG vapor clouds [80]. A properly

designed water curtain is capable of enhancing the dispersion of LNG vapor cloud and reducing the vapor cloud exclusion zone through mechanical effects, dilution and thermal effects. It is considered one of the most economical and efficient cloud control techniques [81]. However, the effectiveness of different water curtains is still widely unknown because of the temperature increase and LNG concentration reduction [80].


**Table 5.** LNG dispersion test on water [80].

LLNL—Lawrence Livermore National Laboratory; note: reprinted with permission from the publisher (American Chemical Society—Energy and Fuels).

Three types of thermal coatings are used for exposure protection. For many years, as a thermal protective coating in refineries and petrochemical industries, concrete has been used [82]. To mitigate LNG vapor ignition and pool fire, high expansion foam application would be one of the effective solutions [82]. Additionally, to extinguish LNG fire dry chemicals such as sodium bicarbonate, potassium bicarbonate and urea potassium bicarbonate are a suitable accomplishment [83]. In most cases, for high expansion foam-controlled LNG pool fire dry chemical is useful, where the magnitude of the fire is small, the production of heat is reduced and fire fighters can apply dry chemical competently. To create greater firefighting capabilities, potassium bicarbonate is usually mixed with an aqueous film forming foam (AFFF) to create a dual agent system. However, inducing corrosions of exposed metals, low visibility after discharge, inducing breathing hazards, clogging ventilation filters, etc., are some limitations of dry chemicals [27].

A very effective method in managing LNG fires is the use of the medium (20:1 up to 200:1) and high (200:1 up to 1000:1) expansion foams. For LNG fires the effective ratio for high expansion foam should be at least 500:1. Gaz de France experience has demonstrated that the utilization of high expansion foam on LNG flames can diminish the fire stature by 60% and also the radiant heat can be decreased by around 90% [79]. Moreover, the results of a recent experiment show that exfoliated ZrP nanoplates can stabilize high expansion foam and this foam can increase the duration over which it needs to be replenished, while by exchanging heat with the vapor and foam and makes the vapor lighter, which ensures effective vapor dispersion [84]. Nevertheless, it is very important here to mention that applying water on LNG fire is not effective as it will not extinguish it; rather, the liquid will vaporize more quickly if the water is applied. Moreover, the fire will get hotter and burn faster if the water is applied to it [79]. Moreover, in recent years, some improvements have been done on LNG dual-fuel engine to improve efficiency, reduce pollution and improve sustainability as discussed in Table 6.


#### **Table 6.** Recent advancements on LNG dual-fuel engine.


**Table 6.** *Cont.*

#### *3.1. Engine Knock Reduction*

Knocking is one of the severe problems of the LNG dual-fuel engine. One of the key initiatives to utilize LNG as a fuel in a diesel engine is to alleviate the knocking problem and take adequate precautions when knocking resistance is low. However, there is no specific technique to reduce knocking completely. One of the solutions to improve the knock resistance of LNG is to keep higher methane in LNG. During the manufacture of LNG, it should be ensured that the amount of methane is 99.4% or higher [14]. The knock resistance of LNG is strongly dependent upon the boil-off rate of the cargo container system used and the initial composition of the fuel. The reduction of methane and nitrogen in LNG increases the fraction of butane and propane and results in decrease in the PKI-MN (propane knock index-methane number), which is responsible for knocking [23]. Further, to reduce the knock combustion, leaning combustible mixture and exhaust gas recirculation strategy can be applied as an efficient way [98]. In the case of an SI engine, higher turbulent kinetic energy in the centre of the combustion chamber, lower mean flow velocity in the spark plug region and near the spark plug can produce knock [99]. Therefore, it is necessary to develop combustion chamber, which can provide higher mean flow velocity.

#### *3.2. Future Research Perspective*

With the increasing demand and potential of LNG, numerous research has been conducted regarding the use of it in a diesel engine. According to previous discussion and literature, researchers and engineers have successfully used LNG in a diesel engine, where the emissions are less, and the power output is similar to conventional engines. However, more research is necessary regarding fuel consumption, fuel atomization, engine equipment and safety of using LNG. According to research [19], below the power of 45.63 kW, the BSFC of LNG dual-fuel engine is higher, though after this certain limit BSFC tends to reduce. The reason can be that the LNG–air mixture may be too lean to burn under light loads. However, as the power increases, the mixture of fuel in the cylinder enriches, eventually the conditions of combustion are improved, which reduces BSFC. Further research is necessary on how the BSFC of LNG dual-fuel engine can be reduced under light loads. Literature suggest that

under a certain load and power output conditions, formation of hydrocarbon is almost double in LNG-dual fuel engine than a conventional on board diesel engine, which also should be a concern for the researchers. Research related to different powertrain such as series and parallel hybrids associated with the use of LNG are very rare. Hengbing Zhao et al. [99] did research, where the authors developed and analyzed some hybrid configurations and according to the simulation results, LNG-CI hybrids showed better results in terms of emissions for Class 8 Hybrid-Electric Truck. Experiments should be conducted in this area to verify the results. Moreover, the hazards and dynamics of LNG spill require more research. Regarding vaporizers, the IFV shows various advantages over other types. In the above discussed literature, finding related to design, configurational variations, working fluids, etc., are discussed. However, it is recommended to develop a three-dimensional unsteady CFD model to evaluate instantaneous flow and heat transfer. As discussed above, in the LNG dual-fuel engine, at low and medium speeds, the production of NOx is higher. Advancing the diesel injection timing, increases the proportion of compression and eventually increases the peak in-cylinder temperature. At medium loads with the increase of temperature, energy in the exhaust gas increases, which eventually increases the intake pressure. This also should be a concern of researchers and further research is necessary to develop a solution to mitigate this problem. For determining hazards, experimental data from large scale experiments are not sufficient. Therefore, it is recommended to consider small scale experiments where large scale is not feasible.

### **4. Conclusions**

All over the world, liquefied natural gas (LNG) would be a clean primary energy source shortly. LNG provides a plethora of potential point of interests as a fuel for a dual fuel engine. It is capable of compensating some of the major drawbacks of natural gas and diesel vehicles. LNG is not only clean, has relatively higher density, has easier transportation and is safe but also the price is reasonable. The indicated thermal efficiency and brake thermal efficiency of an engine using LNG are equivalent to that of diesel and the production of emission is less when using LNG. This paper also reviews successfully the challenges, finds out some responses to overcome those challenges and provides some recommendations for proper utilization of LNG as a fuel for the dual fuel engine. The major challenges of LNG are its higher flammability causes different fatal hazards and when using in dual-fuel engine causes knock. Though researchers have been successful to find out some ways to overcome some challenges, further research is necessary to reduce the hazards and make the fuel more effective and environment-friendly when using as a fuel for a diesel engine.

**Author Contributions:** Conceptualization, M.A.A. and M.N.N.; formal analysis, M.A.A. and M.T.I.; investigation, M.A.A. and M.N.N.; resources, M.A.A. M.W.A.; data curation, M.A.A., M.T.I. and M.W.C.; writing—original draft preparation, M.A.A. and M.N.N.; writing—review & Editing- M.A.A., M.N.N., M.T.I., M.W.C. and M.W.A.; visualization, M.A.A., M.N.N., M.T.I., M.W.C. and M.W.A.; project administration, M.A.A. and M.N.N.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding

**Conflicts of Interest:** The authors declare no conflict of interest

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Energies* Editorial Office E-mail: energies@mdpi.com www.mdpi.com/journal/energies