**1. Introduction**

The escalation of the world population and the unprecedented trend of energy consumption represent significant challenges of the current century. The massive utilization of internal combustion engines (ICE) has led to worldwide modernization while supporting the current living standards; however, such high-scale utilization has also resulted in uncontrolled fossil fuel consumption and alarming environmental pollution [1–3]. The latter represents a complex problem since ICEs play a central role in different sectors such as transportation, agriculture, power generation, and industry, thus setting intensified pressure on pollution deceleration [4–6]. Governmental and international organizations have made a tremendous effort to potentialize global energy transition to renewables while simultaneously setting restrictions in various sectors to mitigate the rate of greenhouse emissions [7,8].

**Citation:** Guillin-Estrada, W.; Maestre-Cambronel, D.; Bula-Silvera, A.; Gonzalez-Quiroga, A.; Duarte-Forero, J. Combustion and Performance Evaluation of a Spark Ignition Engine Operating with Acetone–Butanol–Ethanol and Hydroxy. *Appl. Sci.* **2021**, *11*, 5282. https://doi.org/10.3390/app11115282

Academic Editors: Cinzia Tornatore and Luca Marchitto

Received: 31 March 2021 Accepted: 7 May 2021 Published: 7 June 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 transition to alternative fuels in both compression ignited (CI) and spark-ignited (SI) engines is a pressing need, which emerges as a feasible solution to promote a more reliable operation, minimize global emissions and reduce fossil-fuel dependence. The investigation around CI engines is extensive and diverse as more than 144 biodiesel blends have been reported in the literature with promising results for a reliable and cleaner operation [9].

In the same way, the implementation of alternative fuels in SI engines has been broadly discussed. In this scenario, there is an inclined trend towards implementing oxygenated compounds such as bioethanol and biobutanol. Biobutanol offers different advantages as it can be blended in gasoline at relatively high mixing ratios without modifying the engine functionality. However, the main drawback of biobutanol is the high energy consumption associated with its production, which is based on the acetone–butanol–ethanol (ABE) fermentation process. Therefore, the direct implementation of ABE is a more reliable option from a techno-economic viewpoint. The utilization of ABE and other alcohol additives in SI engines has already been documented in the literature.

For instance, Masum et al. [10] investigated the influence on overall emissions and combustion performance of partial fuel substitution in SI engines with oxygenated fuels such as methanol, butanol, and pentanol in a volumetric replacement of 20%. The engine torque was maximized by using all the fuel blends, whereas emissions levels were minimized moderately. Similarly, Yacoub et al. [11] experimentally evaluated the overall performance of mixing gasoline with different straight-chain alcohol chains from methanol to pentanol (C1–C5) in a SI engine. This study unravels the importance of achieving optimal operating conditions within the engine to guarantee CO and HC emissions minimization. In contrast, the NOx emissions presented both upward or downward trends depending on the engine operating conditions. Nithyanandan et al. [12] examined the overall performance of ABE solution in different volumetric ratios, namely 3:6:1, 6:3:1, and 5:14:1. The study outlines the predominant role of increasing acetone content (6:3:1) as the brake thermal efficiency is improved since the combustion phasing resembles that of pure gasoline.

In the same vein, different perspectives in ethanol implementation have been derived from promoting sustainable operation in ICEs. Di Blasio et al. [13] implemented advanced optical methodologies to analyze the main structural and chemical characteristics of soot particles emitted by ethanol-fueled engines. The results demonstrated the low incidence of ethanol incorporation on the quality and nanostructure of soot emissions. Likewise, Gargiulo et al. [14] outlined that ethanol fumigation positively impacts the greenhouse emissions levels while reducing the concentration of emitted particles. Beatrice et al. [15] revealed the central role of engine calibration, pilot injection, and rail pressure to optimize the benefits of ethanol towards emissions minimization and higher thermal efficiency in CI engines. Similarly, Vassallo et al. [16] allocated the pressing need for research on advanced injection systems as a concrete driver of CO2 emissions targets in the future state of ICEs while maintaining high power density. Belgiorno et al. [17] elaborated on recent advances that integrate gasoline partially premixed combustion in Euro 5 diesel engines. The study focused on describing the effects of appropriate calibration parameters, pilot quantity, and exhaust gas recirculation to maximize thermal efficiency and reduce global pollutants.

On the other hand, hydrogen technology gradually becomes a prominent candidate as an energy carrier that can promote an enhanced operation in both CI and SI engines based on environmental and operational perspectives. Hydrogen production is primarily led by gas reforming technologies representing nearly 60% of the global production [18]. Nonetheless, the carbon footprint of reforming-based production schemes features several challenges to contribute to greenhouse emissions minimization. Therefore, the role of hydrogen production via water-splitting and biomass technologies will be of increased interest in the mid-long term of the hydrogen market [19]. The continuous research on electrolyzers has facilitated the construction of sophisticated and feasible components that maintain proper operation, high-purity reactant agents, and reasonable production rates [18–20].

Therefore, since water electrolysis produces hydrogen and oxygen, the gaseous fuel enrichment in ICEs can be performed either with standalone hydrogen operation and hydroxy gas (HHO).

Shivaprasad et al. [21] experimentally evaluate the influence of hydrogen doping from 5% to 25% in a single-cylinder. Increasing hydrogen replacement increases the incylinder pressure while minimizing both HC and CO emissions; however, the adverse effect of such implementation was the intensification of NOx formation. Ismail et al. [22] envisioned HHO enrichment as a secondary fuel in SI engines encountering promising results towards enhancing thermal efficiency and power output and decreasing emissions. Yilmaz et al. [23] revealed that a constant volumetric HHO enrichment in the engine triggers adverse effects in power output, fuel consumption, and emissions levels. Therefore, the authors implemented a hydroxy control unit to control the volumetric rates of gaseous fuel replacement via voltage and current variations to guarantee the optimal rate based on the engine operation. In this way, they managed to reduce fuel metrics, overall emissions and enhance thermal performance.

The main contribution of this investigation is to evaluate the performance of the dual-fuel operation in spark ignition (SI) engines while simultaneously implementing hydroxy (HHO) gas enrichment and acetone–butanol–ethanol (ABE) as additive. The study incorporates evaluation metrics based on combustion performance, thermal efficiency, fuel consumption, and emissions levels. The novelty of this paper relies on the incorporation of a complete methodology to predict combustion performance and energy/exergy distributions. This study examines a combined fuel operation mode in SI, which has not drawn sufficient attention in published studies. In the development of the experimental assessment, ABE is used in different volumetric ratios, namely 5% (ABE 5) and 10% (ABE 10), whereas hydroxy gas is implemented as gaseous fuel in volumetric flow additions of 0.4 LPM. Moreover, the study includes a complete characterization of the experimental test bench, hydroxy generation system, instrumentation, and measuring uncertainty. Therefore, this work represents a further effort on closing the knowledge gap in the implementation of alternative fuels in SI engines while pinpointing experimental and numerical guidelines to evaluate the performance of dual-fuel technologies.

This paper is structured as follows: Section 2 outlines the main features of the experimental test bench, tested fuels, instrumentation characteristics, and describes the constitutive formulation of the combustion and thermodynamic modeling. Section 3 provides the core findings while critically discussing the outcomes. Finally, Section 4 states the concluding remarks while describing the limitations and future avenues in the field.

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

#### *2.1. Experimental Test Bench*

The experiments were performed in a 4T, naturally aspirated spark-ignition engine (model MZ175, YAMAHA®). The engine has a volumetric capacity of 171 cm<sup>3</sup> and a compression ratio of 8.5:1. It is worth discussing the relevance of the volumetric capacity in ICEs since it provides a clear perspective of the context of the present application. In essence, this matter is essential, considering that international and governmental regulations concerning greenhouse emissions in the transportation sector are based on the volumetric capacity. The typical taxation margin is classified as low (<1000 cc), middle (1200–1500 cc), and high (>1500 cc) capacity [24]. Note that the engine used in this study is intended for power generation applications. However, its operational characteristics resemble those in commercial vehicles, extending the impact spectrum of the proposed dual-fuel technology. Table 1 lists the main features of the SI engine.


**Table 1.** Specifications of the gasoline engine.

A picture of the test bench is shown in Figure 1. The experimental setup consists of the SI engine, hydroxy generation system, and DAQ system, as shown in Figure 2. Firstly, the SI engine is integrated into a dynamometer to control the load condition. A crankshaft angle sensor (Beck Arnley 180–042) allows measuring engine speed. The in-cylinder pressure is measured with a piezoelectric transducer (KISTLER type 7063-A) placed in the cylinder head. The engine fuel consumption rate is obtained via a scale (OHAUS PA313) and a chronometer. On the other hand, intake airflow is measured employing a hot-wire type mass sensor (BOSCH 22,680 7J600). Additionally, the temperatures of the exhaust gases were measured using K-type thermocouples. Lastly, the measurement of CO, NOx, and HC emissions was carried out using two different gas analyzers, namely BrainBee AGS-688 and PCA® 400. An additional gas analyzer (BrainBee OPA-100) measured the opacity levels of the exhaust gases. The measuring instruments were integrated into a data acquisition system that processes the output data. Table 2 lists the main features of the measuring instruments of the test bench.

**Figure 1.** Experimental test bench.

**Figure 2.** Schematic representation of the experimental test bench. (1) AC-DC converter; (2) electrolyzer cell; (3) electrolytic tank; (4) bubbler; (5) HHO storage tank; (6) HHO flowmeter; (7) flame arrester; (8) charge amplifier; (9) pressure sensor; (10) silica gel filter; (11) flowmeter; (12) gasoline tank; (13) ABE tank; (14) fuel pump; (15) BrainBee AGS-688 emission gas; (16) PCA 400 emission gas analyzer; (17) opacimeter BrainBee OPA-100; (18) alternator; (19) encoder; (20) data acquisition (DAQ) system.

**Table 2.** Specifications of measuring instruments.


The hydroxy gas addition was carried out by means of an HHO gas generator installed in the engine intake system (see Figure 2). Hydroxy gas is obtained through a dry cell made of stainless steel, with the ability to withstand high temperatures and currents. Additionally, this type of material does not cause a chemical reaction with the electrolytic substance. To improve cell performance, KOH was used as a catalyst at a concentration of 20% (gram of solute/volume of solution). This allows improving the conductivity of the dry cell.

An electrolytic tank constantly supplies a flow of water to the dry cell to maintain constant hydroxy production. Additionally, a bubbler tank was installed to retain the water content in the hydroxy gas. Two flame arresters and a silica gel filter were installed to prevent flashback.

The measurement uncertainty results from various factors such as measuring instrumentation, calibration, and external environmental conditions [25]. The type A evaluation method, which comprises the statistical evaluation of a series of measurements, was employed. The type A method calculates the best estimate (*bi*) of a set of measurements (*<sup>x</sup>*1, *x*2, *x*3, ... *xn*) using Equation (1):

$$b\_i = \overline{\mathfrak{X}} = \frac{1}{n} \bullet \sum\_{i=1}^{n} \mathfrak{x}\_i \tag{1}$$

In the uncertainty model, the standard deviation (S) assists in calculating the error dispersion of a set of measurements (*<sup>x</sup>*1, *x*2, *x*3, ... *xn*) as expressed in Equation (2).

$$S = \sqrt{\frac{1}{n-1} \bullet \sum\_{i=1}^{n} (x\_i - \overline{x})^2} \tag{2}$$

Lastly, the model uses the standard uncertainty to measure the mean experimental standard deviation *u*(*xi*) from the measurements.

$$f(x\_i) = \frac{1}{\sqrt{n}} \bullet \sqrt{\frac{1}{n-1} \bullet \sum\_{i=1}^{n} (x\_i - \overline{x})^2} \tag{3}$$

where *n* refers to the number of repetitions in the measurements.

Our study set a total of five repetitions (*n* = 5) in the experiments for each operational variable following previous studies [26]. Table 3 shows the uncertainty associated with each operational variable.

**Table 3.** Measurement uncertainty of measured variables.


#### *2.2. Tested Conditions and Fuel Characteristics*

Table 4 lists the properties of the gasoline and the ABE blend oxygenates. Here we implemented ABE additive in a mixing ratio of 3:6:1, which has been implemented in SI engines [27].

The study also used two different alcohol blends, namely ABE 5 and ABE 10, representing the replacement rate in gasoline fuel. Table 5 shows the main properties of these blends.


**Table 4.** Properties of the fuels used in this study [28].

**Table 5.** Properties of the ABE blends.


On the other hand, hydroxy doping has been implemented directly in the intake air system, and the volumetric flow replacement follows the methodology proposed by Ismail et al. [29]. The latter states that a suitable gas substitution rate is 0.25 LPM for an engine capacity of 1000 cm3. Therefore, this study used a 0.04 LPM replacement rate. The main properties of the hydroxy gas that serves within the modeling are: density (0.49 kg • m<sup>−</sup>3) and the LHV (21.99 MJ • kg−1). The experimental assessment alternates the dual-fuel operation in the engine to relate the influence of hydroxy and ABE compounds. A series of 15 runs were established, as shown in Table 6.

**Table 6.** Nomenclature and composition of fuels.


#### *2.3. Fundamentals of the Combustion and Thermodynamic Models*

The fundamental formulation implemented in this paper represents a simplified model of the physical phenomena due to required modeling assumptions. First, all combustion gases in all stages follow ideal gas behavior [30]. Moreover, the flame propagation speed is considered to operate below the supersonic condition, assuming a uniform pressure in the combustion chamber [31]. The combustion reactants are assumed in stoichiometric amounts, considering that a significant part of the combustion results from diffusion interactions. Similarly, reactant's properties are calculated using an average temperature in the combustion chamber, which entails thermal stabilization as a result of the diffusion process. Finally, the model accounts for heat exchange interactions in the cylinder liner to reinforce the model prediction capabilities.

As a first insight, the model establishes the first law of thermodynamics for an open system, which constitutes the combustion chamber. In essence, this model enables the characterization of the heat release curves as a function of the engine operating conditions. Accordingly, Equation (4) gives an energy balance for the control volume neglecting macroscopic effects [32]:

$$\frac{d\mathcal{U}}{d\theta} = \frac{d\mathcal{Q}}{d\theta} - \frac{d\mathcal{W}}{d\theta} + \sum\_{i} \frac{dH\_i}{d\theta} \tag{4}$$

where *U* refers to internal energy, *Q* and *W* represent heat and mechanical work, respectively. *H* refers to the system enthalpy, and *θ* relates to the crank angle.

The heat release rate (*HRR*) can be expressed as defined in Equation (5):

$$H\_{RR} = \frac{m\_{comb} \bullet C\_{\upsilon} \bullet \frac{dT}{d\theta} + P \bullet \frac{dV}{d\theta} + R \bullet T \bullet \frac{dm\_{bb}}{d\theta} + \frac{dQ\_{r}}{d\theta} - \frac{dm\_{fuel}}{d\theta} \bullet (h - u)}{m\_{comb} \bullet LHV} \tag{5}$$

where *mcomb*,*mf uel*, and *mbb* refer to the mass of combustion gases, fuel, and blow-by gas, respectively. Additionally, *R*, *h*, *u*, *V*, and *T* represent the ideal gas constant, specific enthalpy, specific internal energy, volume, and temperature, respectively, which are parameters that assist in determining the thermodynamic state of the fuel mixture. Lastly, *Qr* , *Cv*, and *LHV* refer to rejected heat, specific heat at constant volume, and lower heating value, respectively.

#### 2.3.1. Calculation of Combustion Gases Properties

Firstly, the average temperature inside the combustion chamber is calculated via Equation (6), which includes the universal gas constant (*R*). The latter is calculated via Equation (7).

$$T = \frac{P \bullet V}{m\_{comb} \bullet R} \tag{6}$$

$$R = X\_{air} \bullet R\_{air} + X\_{sl} \bullet R\_{sl} + X\_{\mathcal{S}} \bullet R\_{\mathcal{S}} \tag{7}$$

where *Rair*, *Rst* and *Rg* refer to the air gas constants of air, stoichiometric combustion, and gaseous fuel, respectively. Similarly, *Xair*, *Xst* and *Xg* correspond to the mass fraction of the gases mentioned above.

On the other hand, the specific heat ratio of combustion gases uses the Zucrow and Hoffman correlation [33], as described in Equation (8).

$$\gamma(T) = 1.46 - 1.63 \bullet 10^{-4} \bullet T + 4.14 \bullet 10^{-8} \bullet T^2 \tag{8}$$

Subsequently, the specific heat at constant volume of the combustion gases is given by Equation (9):

$$\mathbb{C}\_{\upsilon}(T) = \frac{\mathbb{R}}{\gamma(T) - 1} \tag{9}$$

Lastly, both the specific enthalpy and internal energy are calculated as a function of the specific heat ratio, which is temperature-dependent, as shown in Equations (10) and (11), respectively.

$$h(T) = R \int \frac{\gamma(T)}{\gamma(T) - 1} \, dT \tag{10}$$

$$
\mu(T) = R \int \frac{1}{\gamma(T) - 1} \, dT \tag{11}
$$

#### 2.3.2. Blow-by Gas Losses

As previously discussed, the blow-by gas losses phenomena account for a significant share of energy losses. Therefore, incorporating such effects in the thermodynamic model becomes a determinant factor in predicting the operating conditions [34]. Hence, the study implements the formulation introduced by Irimescu [35] to predict the energy losses derived from exhaust gas leakage inside the combustion chamber as described in Equations (12) and (13).

$$\frac{dm\_{bb}}{d\theta} = \frac{P\_{ext} \bullet A\_v \bullet C\_D}{N \bullet \left(R \bullet T\_{ext}\right)^{1/2}} \bullet \left(\frac{P\_{int}}{P\_{ext}}\right)^{1/\gamma(T)} \bullet \left[2 \bullet \frac{\gamma(T)}{\gamma(T) - 1} \bullet \left(1 - \left(\frac{P\_{int}}{P\_{ext}}\right)^{\frac{\gamma(T) - 1}{\gamma(T)}}\right)\right]^{1/2} \tag{12}$$

$$\frac{dm\_{bb}}{d\theta} = \frac{P\_s \bullet A\_v \bullet C\_D}{\gamma(T)^{1/2} \bullet N \bullet \left(R \bullet T\_{ext}\right)^{1/2} \bullet \left[1 - \left(\frac{2}{\gamma(T) + 1}\right)^{\frac{\gamma(T) + 1}{\Delta \gamma(T) - 1}}\right]}$$

$$if\_f \frac{P\_{int}}{P\_{ext}} \le \left(\frac{2}{\gamma(T) - 1}\right)^{\frac{\gamma(T) - 1}{\gamma(T)}}\tag{13}$$

where *Pext* and *Pint* represent the exhaust and intake flow pressures, respectively. *N* refers to the engine speed. *CD* and *Av* account for the discharge coefficient and engine valve area calculated via Equations (14) and (15), respectively.

.

$$\mathcal{C}\_D = \frac{m\_{\text{fuel}}}{\dot{m}\_t} \tag{14}$$

$$A\_v = \frac{\pi \bullet D\_v^{\cdot 2}}{4} \tag{15}$$

where .*mf uel* is the experimental mass flow of the inlet valve and .*mt* is the theoretical mass flow rate, calculated considering a constant compressible flow through a valve orifice. Besides, *Dv* is the diameter of the valve and amounts to 28 mm.
