1. Introduction
Micro Gas Turbines (MGT) represent a promising solution for distributed cogeneration due to important energy saving and reduction in pollutant emissions which characterize this type of power plant. Furthermore, we are witnessing a great variety of gas turbine-based hybrid power plants that combine two or more power generating devices and make use of the synergism to generate maximum power and heat recovery, thus offering high efficiencies. In this regard, for small-scale, distributed energy application, hybrid systems based on the integration between solar field and gas turbine (GT) technologies are recognized as a very interesting solution [
1,
2,
3,
4,
5,
6].
In addition, the flexibility of gas turbines offers the possibility to use alternative fuels that can further reduce the carbon footprint and energy demand. Particularly, two type of fuels are receiving growing attention—i.e., syngas and hydrogen—since they can be produced from wind or solar power and from any hydrocarbon-based feedstock gasification or pyrolysis of biomasses [
7,
8,
9,
10,
11,
12]. In particular, the syngas, consisting of a mixture of methane, carbon monoxide, hydrogen and a significant amount of inert gases, such as carbon dioxide and nitrogen, with percentages of each species being dependent on the gasification process used for its production, shows a lower LHV with respect to the natural gas. Consequently, the fuel change imposes a substantial increase in the fuel mass flow rate in order to meet the same turbine inlet temperature (TIT), leading to an off-design operation of the whole system and to a change in the reacting mixing quality in the combustion chamber, which in turn requires a trade-off between combustion efficiency, pollutant formation and overall system efficiency [
9]. Additionally, another characteristic issue related with use of syngas fuel in MGT combustors is a poor combustion completion.
On the other hand, the use of hydrogen fuel offers several appealing features, such as a high LHV and a carbon-free composition, although some other critical aspects are associated with its use in MGT combustors, namely the increase in NO
X production due to the higher flame temperature and the risk of flashback [
10].
The use of these new fuels has been investigated by several researchers, who tried to address the abovementioned issues. Concerning syngas fuel, Abagnale et al. [
7] investigated the performance of a 30 kW MGT annular combustor when fueled with two type of syngas and found that syngas from biomass gasification, despite a good combustion, can lead to an increase in NO
X formation. Cameretti et al. [
8] investigated the suitability of liquid and gaseous fuels for use in a lean premixed MGT combustor and also assessed the potential of external EGR for NO
X reduction. Published literature also reports other promising solutions for combustion completion improvements and NO
X mitigation, such as the increase in pilot injection rate and modification to the pilot location, respectively, [
9]. Mărculescu et al. [
13] analyzed the change in performance when the MGT is adapted to burn alternative low-quality gas fuel produced by biomass gasification with heating values 3 to 5 times lower than methane. At constant heat flow rate in the combustion chamber, the gas flow rate increases to keep the temperature of the flue gases constant. Then, the gross electric power increases but the net electric output decreases compared to methane use. Nicolosi et al. [
14] assessed the effect of the variation of the operating conditions on the performance of the recuperator and, therefore, of the whole MGT. The use of alternative fuels with low LHV shifts the operative points of the turbomachines; in general, compression ratio is reduced, as is the flow rate of the compressor. Therefore, attention must be paid to the compressor stall limit. The recuperator shows a slight variation in the temperature of the fluids, but a higher efficiency is recorded as the flow rate is typically reduced and a better heat recovery performance can be obtained.
Hasini et al. [
15] performed a CFD investigation, using ANSYS Fluent, of flow, combustion process and pollutant emission using natural gas, liquefied natural gas and syngas of different composition on a combustor can-type. The prediction of pollutant species concentration at combustor exit shows significant reduction in CO
2 and NO
X for syngas combustion compared to conventional natural gas and LNG combustion. Ammar et al. [
16] examined four syngas fuel compositions which differed by H
2/CO ratios. The volume of CO
2 at the exhaust decreases with the increase in hydrogen content, while the NO emissions increase as the hydrogen content increases. However, syngas fuel has lower emissions of NO and CO
2 and higher emissions of CO than those from natural gas fuel at the same operating conditions. In [
17] it is demonstrated that the reactivity of the syngas mixture is governed by hydrogen chemistry for CO concentrations lower than 50% in the fuel mixture, while for higher CO concentrations, an inhibiting effect of CO is observed.
The syngas supply and combustion efficiency of the MGT Capstone C30 were already experimentally investigated in [
18,
19], by testing different compositions that have various hydrogen and carbon monoxide ratios. It was found that the high amount of hydrogen content in syngas leads to an increase in the combustion efficiency and emits more NOx emissions. Indeed, higher hydrogen content elongates the flame because of the presence of more H radicals. This H radical promotes the chain branching and chain propagation, enhancing the reaction zone length and flame speed. Thus, the flammability limits are extended so that the flame is able to achieve stability at leaner conditions. On the other hand, CO, emissions are released when the combustion efficiency is low and increase with the CO content in the syngas.
Hydrogen fuel is considered another appealing alternative to fossil fuels, potentially capable of relieving global CO
2 emissions. Pambudi et al. [
20] analyzed the impact of hydrogen in the power generation sector by comparing a base scenario without hydrogen with one in which hydrogen substitutes part of the fossil fuel supply from 2020 onwards. The use of hydrogen would result in substantially less CO
2 being released into the atmosphere, leading to emission reductions of nearly 60% by 2050. However, the percentage of hydrogen to the energy supply must be around 10% due to the increased NOx emissions.
The suitability of the use of CH
4/H
2 blends in MGT was investigated in various ratios both in lean premixed and in RQL combustors by Tuccillo et al. [
10], who found that the lean premixed combustor type appears more prone to flashback at lower H
2 concentrations and that RQL combustors offer a better chance for NO
X control via the separate injection of methane and hydrogen. However, the increased combustor outlet temperature might be a problem from a turbine blade thermal resistance point of view, as well as combustor design modifications in the dilution zone, which might be necessary [
11].
In this regard, the design and construction of a 100 kW pure hydrogen fueled MGT is illustrated in [
21]. The progressive optimization of the compressor–combustor system is achieved starting from a full thermodynamic cycle analysis and though CFD simulations in steady and transient conditions.
In [
22], steam injection (STIG) is proposed as a solution to limit the drawbacks caused by the use of alternative fuels in MGT. A proper injection strategy is evaluated to enhance the electric power by 24% and the electric efficiency up to 29%. Moreover, steam injection in the combustion chamber, allows for dramatically reduced CO and NOx emissions. In particular, the higher CO emissions can be reduced by injecting the majority of the steam in the dilution zone, allowing higher temperatures, which enhance the CO oxidation, while the NOx emissions can be reduced by injecting part of the steam in the pilot zone, where the majority of the NOx forms.
However, despite the scientific effort witnessed by a large number of publications in the field, none of the abovementioned approaches allowed us to solve all the issues related with the use of alternative fuels within MGT designed for natural gas, which still presents several criticalities.
In the present study, a hybrid micro gas turbine/solar field plant was considered to evaluate the contribution of the solar radiation to increase the combustor inlet temperature for improving the efficiency of a recuperated thermal cycle. As a result of the introduction of the solar field, the gas turbine operating conditions change during the day, varying the fuel input requirement with considerable effect on the exhaust emissions.
Additionally, two fuels able to reduce CO
2 and pollutant emissions—hydrogen and a syngas from agriculture biomass—were tested and compared with natural gas as a reference fuel. The authors present a thermodynamic model of a CHP plant integrated with a solar field by Thermoflex
® solver. The micro gas turbine is a 30 kW Capstone C30, able to produce electric and thermal power (hot water). The syngas considered is obtained from gasification of olive pits biomass [
1], which is present copiously in Southern Italy. As a matter of fact, several agriculture products are considered in the literature as potential fuels [
23,
24,
25,
26], depending on geographic sites rich in a particular biomass.
The thermal cycle analysis was carried out over daily and yearly periods and it aimed, in particular, at estimating the effectiveness of thermal storage in the solar field as a feasible way to extend both energy saving and pollutant reduction to nightly hours operation. The analysis compared the solar assisted CHP plant performance for several fuels and a final second law analysis aimed at the estimation of the increase in residual mechanical energy potential obtained by means of the solar contribution.
In the final part of this paper, to evaluate the pollutant emissions, in particular NOx, a CFD calculation of MGT combustor was performed for each fuel, considered by using ANSYS Fluent solver. Many studies were performed by the authors themselves to evaluate the behavior of this combustor by varying fuels and operating conditions [
3,
5,
27,
28].
2. Thermodynamic Analysis
With the aim of assessing the suitability of the use of alternative fuels in a hybrid solar field-mGT plant, from a retrofit point of view, the model of a Capstone C30 micro Gas Turbine (mGT), whose characteristics are described in
Table 1 [
1,
3,
4,
7,
9,
10], was implemented in a zero-dimensional solver (Thermoflex). Additionally, a solar field with storage was integrated in the plant schematic, as depicted in
Figure 1 and the mGT was simulated in cogenerating mode. Finally, the plant was provided with a Heat Recovery Unit placed downstream of the recuperator #5 (
Figure 1), on the exhaust gas path. Concerning the solar field, although the adoption of parabolic trough collectors with a heat transfer fluid—namely, a diathermic oil—is known to be less efficient than solar fields with direct radiation from heliostats to the operating fluid in the micro gas turbine [
1,
4,
5], the former solution was chosen as it is less expensive and it allows an easier thermal storage. Additionally, the solar field was sized defining a solar network exit temperature of 420 °C, referring to the irradiance curve of July to obtain a reduced collector area for cost containment, although a thorough economic analysis was not performed. For additional details about the solar field, the reader is referred to
Table 2.
The heat collected by solar field was provided to the air flowing through the mGT via the recuperator # 12 (
Figure 1), placed downstream of the mGT’s compressor. Consequently, the solar contribution was expected to provide a two-fold effect:
The increased air temperature at combustor inlet was expected to moderately reduce the fuel consumption for a given firing temperature;
An enhanced performance of the CHP power plant was expected, thanks to the higher temperatures reached by the exhaust gases at the inlet of heat recovery unit.
Mass and energy balances for the abovementioned plant, of which a schematic representation is provided in
Figure 1, were solved by a zero-dimensional solver (Thermoflex).
Firstly, a thermodynamic analysis was performed with and without energy storage to highlight its potential benefit on the daily fuel demand considering 5th July, for which hourly ambient temperature is depicted in
Figure 2. Secondly, the thermodynamic analysis was extended, for the thermal storage case only, including different fuels. More specifically, three fuels were considered—namely conventional natural gas, which was treated as the baseline case for comparisons; syngas from agriculture waste biomass, (i.e., olive pits) and pure hydrogen. Syngas and natural gas compositions are reported in
Table 3.
However, it is to be considered that different fuels require slight modifications to the plant layout to account for either fuel compression to the combustion chamber pressure or pressure regulation. Nonetheless, it was deemed that such modifications might have introduced a bias in the results, and so in the fuels comparison. Hence, in order to leave the plant unmodified from case to case, still holding the likelihood of the results, the following approach was followed by the authors:
For the natural gas case, the hypothesized scenario is a delivery via net. Hence, a compression power was accounted in the plant net power computation, based on the fuel mass flow rate, compression ratio (delivery-combustion chamber) and compressor polytropic efficiency.
For the syngas case, the envisaged scenario is an on-site production. Consequently, plant net power was diminished by the compression power from ambient pressure to reaction chamber pressure, and further to combustion chamber pressure, of the reacting air flow rate for syngas production.
For the hydrogen case, the purchase of pressurized gas from a producing plant is considered the most likely scenario. Accordingly, being the average storage pressure available on the Italian market about 200 bar, no compression work was accounted in the plant net power computation.
Based on the above-mentioned assumptions, the thermodynamic analysis was conducted accounting for ambient temperature variation during the day and the year by imposing as a boundary condition the ambient temperature retrieved from [
29], and the results were compared both on an hour-to-hour basis for a given day, as well as averaged monthly, taking power, efficiency, thermal power supplied by the fuel, heat available to HRU and CO
2 emissions as figures of merit. Furthermore, an analysis on the cogeneration process and its characteristic indexes was also performed. Additionally, due to the increased amount of recoverable heat from exhausts at higher temperatures, a proper exergy analysis was carried out to identify the potential of still available mechanical energy.
In the next section, the results of the thermodynamic analysis together with those of cogeneration and exergy analysis are presented.
4. CFD Analysis
Beside carbon dioxide, it is important to take into account other pollutant species. Actually, relevant nitric oxides emissions represent an important issue for gas turbine operation. The high peak flame temperature and high oxygen concentration are the main causes of this harmful pollutant production.
A CFD based investigation can be extremely useful to check combustion effectiveness and pollutant formation. Simulations have been carried out by using FLUENT
® flow solver. The reversed flow annular combustor was reproduced by a 2D mesh to reduce computational effort (
Figure 18). In
Figure 18, the features of the grid are listed, and a three-dimensional view of the combustor is displayed to better show its configuration.
The chemical kinetics sub-model chosen for the combustion simulations is the well-established GRI-Mech 3.0, also tested by the authors in [
10,
27]. It consists of 53 species and 325 reactions and it includes detailed kinetics for methane, lighter hydrocarbons and hydrogen. A chain of reactions for the nitric oxide formation is also included.
The combustion development is modelled by the Eddy dissipation concept (EDC), where the detailed Arrhenius chemical kinetics are incorporated in turbulent flames, and this make this model suitable to deal with simultaneous reactions.
Due to the high computational efforts required by calculations in order to deal with this complex kinetic mechanism, the heat transfer from the core to external liner was simulated by assigning a fixed temperature of 1000 K to the intermediate wall. Such a value was chosen based on previous simulations carried out on the same combustor by the authors, since this procedure allows a satisfactory evaluation of the temperature gradients within the primary region.
The CFD calculations use the boundary conditions shown in
Table 8. The mass flow of each fuel is related to its LHV (
Table 6): in order to reach the same TIT, the lower the LHV, the higher the amount of fuel introduced in the combustion chamber, as reported in
Table 8.
It is important to highlight that the air mass flow rate reported in
Table 8 is actually split into two parts. The main one is assigned to the inlet from the recuperator. This air crosses the external liner and is progressively introduced in the internal liner for the completion of oxidation and gas dilution.
As a matter of fact, the injector also crosses the external liner and faces directly in the primary zone (
Figure 18); a small amount of the air coming from the recuperator is trapped in the injector through holes that are properly oriented to provide a swirl motion to the air that is mixed with the fuel inside the injector. In this way, a premixed mixture is formed and burns in the primary zone of the internal liner. However, the difficulties encountered in the simulation of this configuration in a 2D domain suggested to the authors to split the total mass flow rate and to assign a fixed direct injection of air from the fuel inlet for each fuel.
In
Table 9, the results, in terms of temperatures and emissions, are listed. As already stated, the characteristics reported in
Table 1 were used for the calculations in the zero-dimensional solver. In particular, the fuel mass flow was calculated in order to reach a TIT (or combustor outlet temperature) of 1173 K. The amount of each fuel is different since the three fuels have different LHV, and this could lead to a different combustion development. However, the target temperature at the outlet is reached for all cases. Moreover, as expected, the higher temperature peak in the combustor are attained in the case with hydrogen since it has the higher adiabatic flame temperature. The lower temperature obtained by the syngas combustion leads to an effective reduction in nitric oxides at the outlet with respect to the other two fuels. Additionally, in
Table 9 the concentrations of the main species for each fuel, respectively, H
2, CO and CH
4, as reported in
Table 2, are compared. As it is noted, hydrogen and natural gas present an efficient combustion while the syngas case shows that a significant amount of carbon monoxide is not able to be converted in CO
2. A possible cause can be found in the evaluation of the fuel and air mass flow rates in
Table 8. Indeed, to achieve the same TIT, a higher amount of syngas must be introduced, while a lower air flow is necessary to keep the same operating conditions of the micro gas turbine. Therefore, the air amount decrease prevents the complete oxidation of the fuel. Finally, it is important to evidence that the use of hydrogen does not produce emissions of CO
2 and CO at the outlet.
In
Figure 19, temperature distributions show higher peaks for the case with hydrogen, as already highlighted in
Table 9. However, this case presents the same emissions of NO
x as the natural gas case. Indeed, despite NG does not achieve the same maximum temperature as hydrogen, the wider extension of the zone promotes the nitric oxides formation. This is confirmed by the NO distribution in
Figure 20 and its reaction rate in
Figure 21, as the higher concentrations and rates of reaction are detected in correspondence with the higher temperatures.
In
Figure 22, the distributions of OH are illustrated because this radical is an important intermediate species of the hydrocarbons chain of reactions, this means that the highest concentrations of this species identify the location of combustion activity. As it is possible to notice in all cases, the combustion is delimited in the primary and secondary zones, that are characterized by the most significant OH presence. A reduced OH concentration is also observed after the last row of dilution holes, especially in the syngas and Natural Gas cases, thus indicating that the completion of the oxidation process involves the dilution zone.
Finally, in
Figure 23 and
Figure 24, the reaction rate of the main species for each fuel and the related mass fractions are displayed. Indeed, as already demonstrated by the authors [
7,
10,
27], the combustor behavior approaches that typical of an RQL combustor. Hence, the fuel distribution is promoted by the swirl motion generated inside the combustor, which does not vary with the fuel used and has the aim to form in a primary zone a rich mixture with a poor availability of oxygen in order to prevent the nitric oxides formation. However, in the syngas case this behavior is accentuated by the increased amount of fuel introduced. The rich mixture is even farther away from the stoichiometric value and, different from H
2 and CH
4 that start to burn near the injector, the main oxidation occurs only in proximity of the first row of dilution holes, as evidenced in
Figure 23. This could explain the higher concentration of CO in the secondary zone, the slower combustion of syngas and the reduced amount of NO
x at the outlet.
As illustrated in
Figure 24, despite the different scales used for the distinct fuels, in all cases the primary zone is characterized by high concentration of fuel while the mixing with air and then the main oxidation process occurs between the primary and secondary zone.
5. Conclusions
The present work showed a thermodynamic analysis carried out on a mGT-based power plant with solar field and thermal storage, provided with a Heat Recovery Unit for cogeneration purpose. The analysis was first carried out for a given day with and without thermal storage to assess its beneficial impact on fuel demand during night hours. Subsequently, the analysis was extended to the assessment of the plant response to two alternative fuels—i.e., pure hydrogen and syngas from agriculture waste, namely olive pits, which were compared against the natural gas case. Calculations were first performed on an hour-to-hour basis for a single day, and were later, extended to 12 months.
To assess the relative performance of the three fuels, several figures of merit were considered—e.g., net power output, fuel-based thermal efficiency, thermal power to the HRU and specific CO2 emission. Additionally, to enhance the legibility of the impact of thermal storage on energy availability and fuel demand during night hours, two groups of results were presented—i.e., daylight and night hour average quantities.
Concerning fuel response analysis, from a specific CO2 emissions point of view, pure hydrogen was obviously the best alternative with basically negligible values, whereas syngas was shown to produce the largest amount of CO2 per kWh. Nevertheless, it is to be kept in mind that syngas, being a fuel from vegetable origin, is considered neutral by legislation. Furthermore, pure hydrogen fuel proved to be the best solution from power and efficiency point of view, both for single day and 12 months analyses. Conversely, syngas stood out from heat transferred to the HRU point of view, both for single day and 12 months analyses.
Subsequently, in view of the large amount of thermal energy available to the HRU, a detailed analysis of the cogeneration process as well as an exergy analysis were performed. Results showed that the adoption of the solar field led to a substantial increase in the thermal power available to the HRU, independently of the fuel considered, which, in turn, largely increased RISP index and FEUF, with the best results achieved by syngas and hydrogen. Likewise, exergy analysis proved that the introduction of the solar field almost doubled the exergy flow rate availability at HRU, with a large improvement in second law efficiency.
Finally, a CFD analysis carried out using the well-established GRI-Mech 3.0 on the FLUENT® flow solver demonstrated that, as expected, the higher temperature peak in the combustor are attained in the case with hydrogen since it has the higher adiabatic flame temperature, presenting the same emissions of NOx as the natural gas case, since this last case is characterized by wider extensions of the zones at high temperature that promote the nitric oxides formation. On the contrary, the lower temperature obtained by the syngas combustion leads to an effective reduction in nitric oxides at the outlet with respect to the other two fuels. However, the syngas case shows that a significant amount of carbon monoxide is not able to be converted in CO2, different from hydrogen and natural gas, that present an efficient combustion; in particular the use of hydrogen does not produce emissions of CO2 and CO at the outlet.