Operational Strategies of Two-Spool Micro Gas Turbine With Alternative Fuels: A Performance Assessment

Renewable energy systems are being discussed globally, which has led to an increase in the use of low-emission energy resources. If low-emission technologies and renewable fuels are not introduced promptly and the electricity production continues to grow rapidly, a significant increase in greenhouse gas emissions will be observed in the coming decades. Decentralized and hybrid power systems present many benefits (i.e., improving overall efficiency and minimizing emissions) [1] in future power production. In that context, micro gas turbines (mGT) could play an important role in such systems, but their low efficiency and high specific price per kilowatt prevent them from being implemented effectively. Micro gas turbines power capacities are not well defined but are generally specified between 10 kW and 500 kW [2]. The typical performance of such an engine is lower than that of their competitor in the market (i.e., internal combustion engines—ICE).

Historically, microturbine technology has been researched since 1970, when the automotive industry began to consider replacing reciprocating piston engines [2,3]. Microturbines have not had great success in the automotive sector as they show lower transient response and higher cost compared to ICE [2]. However, they could play an imperative role in developing decentralized and hybrid energy systems in the current era. By increasing their performance coupled with state-of-the-art technologies and using renewable fuels, they could meet market demand in terms of emissions and efficiency.

Micro gas turbines are considered to be promising technologies due to their fuel flexibility, low emissions, high power density, and low maintenance costs [4,5]. Despite their relatively low exhaust temperatures (around 150–300 °C), their waste heat can still be utilized to generate low-pressure steam and/or hot water [2]. The heat provided by mGT exhaust allows it to be implemented in combined cycle concepts [6]. The thermal recovery performed by the organic Rankine cycle (ORC) bottomed to an mGT can increase the combined cycle thermal efficiency by 15% [7]. Also, concepts like the micro humid air turbine (mHAT) [8] proved to increase the performance of the engine significantly. Moreover, bladeless or Tesla technology can play a vital role in increasing the compressor's performance due to its reversible operation, low manufacturing costs, and low noise [9]. In that context, an alternative mGT concept, which is the two-spool intercooled mGT, was introduced. This engine seems capable of successfully conquering the small-scale combined heat and power (CHP) market. This two-spool intercooled mGT presents an electrical efficiency of around 40% and is capable of incorporating a range of fuels [10]. In the literature, two studies can be found that describe the design of this engine [11,12], where especially Malkamäki et al. have indicated that its efficiency and cost can be compared with a reciprocating engine from MWM GmbH of the same load [11].

Besides increasing the efficiency of decentralized production, the world also focuses on renewable fuels, such as hydrogen, syngas, and biogas [1,13], in response to the growing demand for low-emission technologies. The shift towards more renewable energy production also pushes the mGT to become more flexible in terms of fuel utilization. Therefore, apart from natural gas, alternative fuels with lower energy content, like syngas and biogas should be used. Moreover, the energy transition requires the utilization of hydrogen as the main fuel. Given the occasionally limited availability of these fuels, the mGT should thus be capable of running on both classical fossil-based gaseous fuels, hydrogen which has a high lower heating value (LHV), and alternative fuels like syngas which have a rather low LHV. The addition of alternative fuels could restrict the mGT operation as compressor surge can occur due to the decreasing air mass flow rate. Also, the combustion process when natural gas is mixed with syngas or hydrogen can cause combustion instabilities, leading to flameout (syngas) or flashback (hydrogen).

Having that in mind, researchers [14–17] studied the impact of using hydrogen and/or a blend of natural gas/hydrogen on the mGT both numerically and experimentally. Hydrogen addition in methane increased the flame temperatures, with cooling and NO_x emissions being the primary concerns [14,15]. Also, an experimental and numerical study on a pure hydrogen-fueled gas turbine highlighted the higher outlet velocity and pressure drop in the combustion chamber (CC) compared to pure natural gas operation [16]. Pappa et al. showed that the reactivity of hydrogen can be lower by applying steam as a diluent using large eddy simulation [17].

Similarly, different combustion challenges are observed when syngas is used instead of hydrogen. Bompas et al. [18] numerically compared carbon monoxide levels when using natural gas or syngas, indicating that syngas does not significantly impact combustion efficiency. Diaz et al. [19] examined the energy and exergy efficiencies of a syngas mGT cogeneration system while also analyzing the total costs associated with the production of energy from syngas obtained by gasification. They showed that increasing the compression ratio improves the thermodynamic performance, but the cost of the system per unit of time is increased as well. Furthermore, it was found that the CC presented the highest rate of exergy destruction [19]. Thermodynamic and computational fluid dynamics (CFD) analyses have been performed on a small-scale integrated energy system composed of mGT-ORC-gasifier focusing on the performances of five different syngas-biogas fuels. This study indicated the necessity of operating the mGT in particular conditions, as determined by the compressor and turbine performance maps of the AE-T100 due to the low LHV of the fuels [20]. Renzi et al. assessed numerically the behavior of the same mGT cycle when it is fed with biomass-derived syngas showing that the use of syngas decreased the overall efficiency of the system by 5% due to the higher thermal input that is required to keep the produced power at the nominal value [21]. They also included a steam injection gas turbine (STIG) cycle to counteract the negative effects of syngas utilization on the cycle performance. De Paepe et al. extended the work of Renzi et al. by comparing two steady-state models and validating them in their capability to simulate steam injection in the T100 mGT fed with natural gas and syngas [22]. When the cycle runs with syngas some variation in performance parameters between both models is observed, which confirms the importance of the accurate and extensive modeling of the combustion chamber [22].

The above-mentioned studies analyze the performance characteristics of running mGT engines with alternative fuels. Numerical and experimental works studied the exhaust gas emissions and the required design modifications of CC to incorporate syngas or hydrogen. Although extensive performance analysis on one of the most promising mGT (i.e., two-spool mGT) in the market by utilizing several alternative fuel mixtures is still missing in the literature. The purpose of this article is thus to study the performance of a two-spool mGT from a thermodynamic point of view when applying various fuel mixtures, including natural gas (NG), syngas (SG), and hydrogen (H₂). For this analysis, a steady-state model of the two-spool engine was developed in Python programming language and the combustion mechanisms were embedded according to gri-mech 3.0 library. As validation, the 0D model is compared with another model of the two-spool mGT from the literature. Additionally, two part-load strategies are employed for the steady-state modeling of this cycle to simulate the part-load behavior. These strategies are compared regarding the electrical efficiency (⁠$ηel$⁠) and the behavior of both compressors. The structure of this paper is as follows: In the Methodology section, the numerical model is described thoroughly, and the two part-load strategies regarding the operation of the cycle are presented. In the Results section, the behavior of several fuel mixtures is discussed at nominal and part-load levels. Finally, the Conclusions section includes the outcomes of this study and the future perspectives of our work.

In this section, the mGT (two-spool engine) used for the numerical model is presented. Then, the description of the modeling techniques that are adopted for this analysis, along with the two part-load strategies, are presented.

The basis of this numerical model is a two-spool mGT, which has similar performance characteristics as the Aurelia A400 mGT [10]. This engine is a recuperated Brayton cycle and consists of two shafts, as shown in Fig. 1. The low-pressure compressor (LPC), low-pressure generator, and low-pressure turbine (LPT) are mounted on the low-pressure shaft (LPS). Additionally, the high-pressure compressor (HPC), high-pressure generator, and high-pressure turbine (HPT) are part of the high-pressure shaft (HPS). The recuperator uses the exhaust gases to increase the enthalpy of the working fluid before it enters the combustion chamber (CC) (5). The other heat exchanger (i.e., the intercooler) cools down the air at the HPC inlet (3) to increase the component's isentropic efficiency. The electrical power is generated from the two high-speed power generators and the power electronics needed to convert the two produced powers to grid frequency. This cycle can achieve an electrical efficiency higher than 40% for a nominal power of 400 kW [10]. The typical specifications of this engine are shown in Table 1. The two-spool machine operates at constant power and behaves similarly to a single-spool variable-speed mGT. Such machines allow the operator to keep the turbine inlet temperature (TIT) constant at high levels while reducing the rotational speed to operate at part load while still maintaining high electrical efficiency. In this specific engine, however, the two shafts can be controlled independently of each other. This allows the turbomachinery components to run at high efficiency for a larger operating range than with a single-shaft engine. Therefore, as the demanded power is decreased, the rotational speeds are decreased and TIT is kept constant by the controller. When the maximum allowable turbine exhaust temperature (TET) is reached at 40% load, the TIT is starting to reduce as well. Then, the electrical efficiency of the engine experiences a sharp decrease, as presented by Jaatinen-Värri et al. [23] with the cooperation of Aurelia Turbines Oy [10].

In this section, the in-house numerical model developed to study the impact of fuel alteration on the performance of a two-spool mGT is extensively described. This software can predict the nominal and part-load behavior of a two-spool mGT by using two different part-load schemes.

The modeling methods that are applied to each component of the mGT cycle are similar to a previous study from part of the authors [24]. In this study, the results of the transient T100 mGT were validated in steady-state and transient conditions. For the current study due to the lack of experimental results and data from the manufacturer (Aurelia Turbines Oy, Lappeenranta, Finland), our model adopted the design, boundary and initial conditions from the available studies on the Aurelia A400. After the description of the modeling methods, the complete model is compared with another study from the literature [11] and with the design specifications of Aurelia A400 [10] to confirm the accurately developed methodology (see later).

The fluid properties are modeled using different equations of state from the Coolprop library [25]. For example, the fluid behavior of water is calculated with IAPWS (International Association for the Properties of Water and Steam) 1995 [26]. The great advantage of this library is the representation of real gas behavior compared to most equations of state that emanate from the assumption of ideal gas behavior using reference states.

where $m˙c,sl$ is the mass flowrate at the surge line considering constant rotational speed.

where $Tw,in$ is the inlet temperature of the water at the cold side of the intercooler and it is kept at 15 °C during the simulations. The cold side of the intercooler utilizes water with a sufficient mass flowrate of 2.1 kg/s. The numbering of Eqs. (6) and (7) corresponds to the different states shown in Fig. 1.

η_cc corresponds to the combustion efficiency of CC and is 0.99. Also, the pressure losses at CC are taken into account as equal to 3%.

The power ratio (P_eR) between the generated power of HPS and LPS is extracted from the literature and has a value of 1.1473 [23]. This power ratio is considered when the LPS and HPS have the same rotational speed at nominal load. The two generators have 96% electromechanical efficiency and the efficiency of the inverter is 98%.

The complete steady-state model of the cycle operating with dry air is presented in Fig. 3. The model uses an iterative method that starts at the inlet of the cold side of the recuperator to accurately calculate the inlet hot side temperature T₈. We assume a value of T₈ to calculate the T₅ in the recuperator block and then determine the thermodynamic values in the inlet and outlet of CC and Turbine blocks until the error of T₈ between iterations is below 10^–5. The external iteration essentially controls the mGT cycle. The simulation converges by changing three user-defined constant values using the Secant method. Therefore, $m˙air$ controls K_c, $N=NLP=NHP$ controls the electrical power (P_e) and PR_HPT controls $PeR=PHP/PLP$⁠.

where the shaft speeds ratio (NR) is the ratio between the high pressure (HP) and low pressure (LP) speeds and is a function of the electric output. The correlation between the generated power and the HPS rotational speed is derived from a study on the control strategies of Aurelia A400 [23]. Also, the power ratio (P_eR) should be a function of the produced electric power (⁠$PeR=f(Pe)/Pe,nom$⁠) as the two shafts are not loaded equally. Both correlations are made with a linear interpolation between the input values. Thus, by applying these two correlations to case 2, it is possible to utilize the part-load control scheme that is suggested by the Aurelia A400 studies in the literature [12,23]. Figure 2 shows the operating lines of case 1 (N_LP = N_HP, dashed line) and case 2 (⁠$NLP≠NHP$⁠, solid line) in both high-pressure and low-pressure compressor maps. The LPC operating line of case 2 is observed in the high-efficiency region as shown in Fig. 2. Whereas in case 1, the reduction in pressure ratio at part-load is less compared to the nominal point. Also, case 1 presents lower efficiencies for the same mass flow rate. The HPC operating line of case 2 shows higher pressure ratios than case 1 to compensate for the behavior of the LPC operating line. As a result, case 2 HPC operates in a small pressure ratio and shaft speed region compared to case 1.

Figure 4 depicts the electrical efficiency behavior at part load until 200 kW_e in both cases. As it is shown, case 2 presents slightly higher efficiency over the whole operating range. However, the efficiency difference is rather small and less than 1% even in 50% of the nominal load. Moreover, the efficiency increases as the generated electric power drops. On the other hand, in case 1, the efficiency starts to drop after 280 kW. Also, no significant difference in efficiencies is observed from 400 to 350 kW.

where the values of our model are shown as x and the values of the study [12] as x_p.

The values of temperature and pressure present an error below 0.7% which confirms the accurate use of the component methods that were applied for this study (Fig. 5). Also, the relative error of the electrical efficiency is 5.96%, as the electrical efficiency of the study [12] and the one we calculated are 45.8% and 43.07% respectively. The increased error in electrical efficiency could be associated with the fact that our simulation model does not use the turbine performance maps to calculate the operating parameters (pressure ratio and shaft speed) of HPT and LPT. Moreover, the study [12] has no information regarding the LHV of natural gas in this publication.

The part-load performance of the model is compared with the only available data regarding the Aurelia A400 in the literature. Therefore, the partial load efficiency of this machine was extracted with digitization from the Aurelia Turbines Oy datasheet [10]. This data is depicted with a blue line in Fig. 6. In this part-load comparison, the model runs with a TIT of 977 °C and T_w,in = 5 °C as suggested by Ref. [30] for each produced power that is set. As a result, the operating conditions of this comparison differ from the comparison that is conducted in Fig. 5. The results of Fig. 6 show that the efficiency relative error between the Aurelia Turbines Oy datasheet and the model is below 0.75%. The highest error is 0.742% and is observed at 200 kW_e. Also, the results of the model follow adequately the efficiency data and have a similar trend. Thus, both comparisons confirm the accuracy of our modeling methods and allow us to utilize this code to carry out a performance assessment in this study.

In this section, the fuels that are used for this analysis are described. First, the compositions of syngas should be chosen with care as the volume percentage of hydrogen (H₂) and carbon monoxide (CO) influence dramatically the LHV of the fuel and the combustion characteristics in general.

Syngas can be produced in several ways, while hydrogen percentages vary depending on the quality of the fuel input. Wood chips should have a moisture content of 20% to produce the maximum amount of hydrogen and lower tar. The presence of hydrogen in syngas generated by wood ranges between 5% and 30% based on literature data [32–35]. For this study, we considered an average amount of moisture in the wood chips. As a result, the amount of H₂ is 15.46% as shown in Table 3. Moreover, NG is composed mainly of methane with a percentage of 91.2% with the ethane and propane amounts being lower than 10%. Table 3 shows that the LHV of syngas is 47% of the LHV of natural gas. This difference in the stored energy of these two fuels could significantly alter the engine's operating conditions. The third fuel that we use for the testing of different mixtures in the two-spool mGT cycle is pure hydrogen with a 119.90 MJ/kg lower heating value.

In this section, the numerical results of the model using different fuel mixtures composed of NG, SG, and hydrogen (H₂) are presented. The steady-state model applied two different control schemes for the part-load operation and the results are depicted as well. In the control scheme of case 1, the rotational speeds of LPS and HPS have the same value at part-load. Whereas, in case 2 the shaft speeds are not equal except for the nominal conditions. Thus, the solid-line curves, in the figures below, refer to simulations performed using the control scheme of case 2 (⁠$NLP≠NHP$⁠) and the dashed lines belong to case 1 (N_LP = N_HP).

$m˙f$ is the fuel flow rate that is injected in the cycle.

As the volume percentage of hydrogen is increased in natural gas from 0% to 100%, the LHV of the fuel mixture increases significantly. This will lead to a decrease in the fuel flow rate. With pure natural gas as baseload, the fuel flow rate is 20.2 g/s, and with pure hydrogen 8.34 g/s. This results in a 58.2% fuel flow decrease. Due to the increased water content of the flue gases, the density and specific heat capacity at point 6 increase. This change in heat capacity of 1.32% leads to a slight drop in the air mass flowrate of 1.41%. All these minor changes, in the conditions of the engine at pure hydrogen, move the operating point of the LPC and HPC in the performance map approximately 0.1% toward the surge limit. This divergence in the operating conditions is very small and the change in electrical efficiency is below 0.2% not only at nominal but also at part-load for both control schemes (cases 1 and 2).

Although we do not see any significant performance alteration when we increase the amount of hydrogen in the fuel, syngas presents a major impact. The amount of syngas in natural gas is increased from 0 to 100 vol. % at 200, 300, and 400 kW of electric load with a 10 vol. % step. The results of electrical efficiency are shown in Fig. 7. We observe that at nominal load ("P_e = 400 kW_e" line) the performance of the cycle remains somewhat constant around 39.88%. Only from 90 to 100 vol. %, the efficiency is slightly decreased. When applying the operational strategy of case 1 (N_LP = N_HP, dashed lines) at part-load, the efficiency at pure natural gas on 300 and 200 kW is almost identical at 41.2% (dashed lines at 0 vol. % SG content) which is also confirmed by Fig. 4. In Fig. 4 we can see that from 310 to 200 kW the efficiency line of case 1 flattens. The performance of case 1 starts to decrease dramatically when the syngas volume percentage exceeds 50% in natural gas (Fig. 7. The efficiency at 300 kW shows 1.3% and at 200 kW 2.31% absolute decrease at 100% syngas. Case 2 (⁠$NLP≠NHP$⁠, solid lines) presents a higher efficiency at the two part-loads with a variable syngas fraction than case 1. At 300 kW the efficiency of pure syngas even shows an absolute increase of 0.28% and at 200 kW of 0.94%. Furthermore, the efficiency of case 2 does not increase dramatically as we decrease the load in the range of 70–100% of syngas in the fuel mixture. Thus, at high amounts of syngas, the efficiency remains rather constant as the power drops. At the range of 0–70% of syngas, the efficiency increases monotonously as the electric power decreases.

The behavior of the part-load efficiency in cases 1 and 2 is associated with the operating conditions of the LPC. Figure 8 depicts the isentropic efficiency of the LPC (⁠$ηLPC,is$⁠) of the two cases while we increase the volume percentage of syngas into the fuel. At full load (400 kW) the isentropic efficiency of the component increases as we inject more syngas. This helps the electrical efficiency of the cycle to remain constant (see Fig. 7). The efficiency at 300 kW does not change significantly in both cases and remains almost constant even at high amounts of syngas which is in line with the behavior of electrical efficiency at 300 kW in Fig. 7. At 200 kW the superiority of case 2 is obvious as the isentropic efficiency remains around 83.4% and enhances the total electrical efficiency of the cycle.

At this part of the analysis, it is necessary to also present the LPC map (not only efficiency, as presented before) to get a full picture regarding the behavior of the two-spool mGT when syngas as fuel is used. Figure 9 presents the LPC performance map in normalized values. We can observe that by increasing the volume percentage of syngas, the normalized reduced mass flowrate moves toward the surge line in both control strategies. This can be explained by the difference in fuel LHV. As the LHV of the fuel mixture drops in the syngas case, the demanded fuel flowrate in the CC, to reach the required TIT, increases. However, the flue gas mass flowrate is constant as the LPT is choked. This results in a decrease of the air flowrate of both compressors. In case 2 (solid operating lines) the points of pressure ratio and mass flowrate remain in the region of $ηis/ηis,des=1$ while the dashed lines of case 1 operate at higher pressure ratios. This higher-pressure ratio of the line at 200 kW of case 1 decreases the isentropic efficiency of the component as is also depicted in Fig. 8.

This significant decrease in air flowrate at pure syngas, which is presented in Fig. 9, is linked to a large amount of nitrogen in the fuel (see Table 3). Almost half of the fuel volume is taken by nitrogen which does not take part in the combustion process that gives energy to the flow. As a result, a higher fuel flow rate is needed for the combustion to reach the required TIT. This increased fuel flow rate decreases the air flow rate due to the choking condition on LPT. Thus, the additional normalized airflow rate decrease is 7.3%. Also, the two part-load strategies do not present any significant conclusions on the HPC component. The HPC map presents the operating lines from cases 1 and 2 to be located in the high isentropic efficiency region. Thus, as the HPC isentropic efficiency and also the surge margin remain rather constant, the performance results of HPC in its performance map are not depicted.

As the control scheme of case 1 lets the LPC operate at higher pressure ratios compared to case 2, it is important to calculate from Eq. (2) and evaluate the surge margin behavior in a natural gas—syngas mixture. Figure 10 shows the surge margin (SM), which is calculated from Eq. (2), in the three different generated electric powers. We can see that the SM decreases by 4% from 60 to 100 vol. % of syngas at full-load. However, SM remains over 25% even at 200 kW of case 2 which keeps the LPC at a safe operating region even with high-nitrogen syngas fuel. Although, as the load decreases, case 1 moves the operating points more closely to the surge line. Pure syngas fuel has a surge margin of 22.2% at 300 kW which is considered sufficient, however, at 200 kW the SM goes below 20% with 70 vol. % of syngas. Moreover, at pure syngas, the SM value at 200 kW load is 17%. If the nitrogen amount in syngas increases even more, the safe operation of the mGT is not ensured. We should also consider that the utilized performance maps are obtained using digitization and curve fitting methods. As a result, the current modeling uncertainties lead us to consider that case 1 is not a safe operating strategy at loads below 50% of nominal with pure syngas.

In this subsection, different fuel mixtures of natural gas, syngas, and hydrogen are tested regarding their performance behavior at different loads. Figure 11 presents the electrical efficiency of five different fuel mixtures at an operating range of 200 kW, adopting part-load control strategy case 2. As we discussed in the first paragraph of this chapter, the efficiency curves of pure natural gas (100% NG) and pure hydrogen (100% H₂) are very similar and show the highest electrical efficiency. Only below 280 kW, the two lines start to diverge with slightly higher efficiency for NG. However, the difference remains below 0.2%.

Using syngas lowers the efficiency compared to pure natural gas fuel. One way to compensate, for the lower performance observed with syngas, is the addition of hydrogen to the mixture. As the amount of hydrogen, in a syngas–hydrogen mixture increases (lines red, orange, and brown), the efficiency improves and the maximum value moves toward 200 kW. Furthermore, a 15 vol. % hydrogen addition in the fuel (orange line), increases the efficiency at 200 kW of 0.46%. Also, at 30% of hydrogen (brown line) the efficiency presents almost a constant behavior at the power range of 240 to 320 kW. Therefore, even a 30 vol. % hydrogen addition counteracts the performance disadvantages of syngas consisting of high nitrogen vol. %. This is logical considering the significantly higher LHV of hydrogen compared to syngas.

In this paper, a 0D steady-state model predicted the performance behavior of a two-spool mGT with alternative fuels by applying two different part-load strategies. The model was developed in Python programming language adopting cantera object-oriented library to calculate the combustion process effectively. Two different control schemes for the calculation of the part-load operating conditions of the engine are adopted. The first control strategy used identical rotational speeds (case 1). The second strategy was taken from the literature as it is considered to be one of the most efficient for the two-spool mGT (case 2). In that strategy, the rotational speed of the LPS controls the produced power and the shaft speed of the HPS is a function of this generated power using a correlation from the literature. Moreover, the model is compared with another study about a two-spool mGT; temperature and pressure results showed an error below 0.7%.

Three different fuels are utilized for this performance assessment. We observed the performance behavior of the mGT for natural gas, syngas, and hydrogen in the two part-load scenarios (cases 1 and 2). As we add hydrogen to natural gas, the divergence in the operating conditions is insignificant. On the other hand, including syngas in natural gas leads to lower efficiencies. A mixture of natural gas and syngas decreases the air flow rate of the cycle. Pure syngas presents a 7.1% decrease in normalized air flow rate. The nitrogen content mostly influences the air flow rate decrease in LPC toward the surge line. For this reason, hydrogen could be injected into syngas to counteract the efficiency and surge margin decrease. The superiority of control strategy case 2 is highlighted as it presented increased efficiencies at part-load for five different fuel mixtures. Also, control scheme case 1 showed a surge margin below 20% when the amount of syngas is over 70 vol. % in natural gas at 200 kW. This indicates that the LPC could present instabilities in its operation for lower loads and higher nitrogen contents in syngas.

The results show the potential of applying alternative fuels with an effective operational strategy. However, in future work, specific attention should be paid to the calculation of the flame speed of the different fuel mixtures using CFD analysis and chemical kinetics software to determine the occurrence of flashback. Furthermore, a fuel dilution method by injecting water or nitrogen to avoid this phenomenon could be applied.

This project has received funding from the European Union's Horizon 2020 research and innovation program under the Marie Sklodowska-Curie grant agreement No 861079 (“NextMGT—Next Generation of Micro Gas Turbines for High Efficiency, Low Emissions, and Fuel Flexibility”). This paper reflects only the authors' view, and the Research Executive Agency and the European Commission are not responsible for any use that may be made of the information it contains.

• Université Catholique de Louvain (Grant No. 86107; Funder ID: 10.13039/501100005041).

Roman Symbols

c_p =

constant pressure heat capacity (kJ/(kg°K))

h =

specific enthalpy (J/kg)

K_c =

choke constant

$m˙$ =

mass flow rate (kg/s)

N =

rotational speed (rpm)

p =

pressure (Pa)

P =

power (kW)

T =

temperature (°C)

Greek Symbols

ϵ =

heat exchanger effectiveness

η =

efficiency (%)

π =

pressure ratio

Subscripts

air =

properties of air

des =

design value

el =

electrical

fuel =

properties of fuel

int =

intercooler

rec =

recuperator

red =

reduced

Acronyms

CC =

combustion chamber

CFD =

computational fluid dynamics

CHP =

combined heat and power

FAR =

fuel-to-air ratio

HPC =

high-pressure compressor

HPT =

high-pressure turbine

ICE =

internal combustion engine

LHV =

lower heating value (MJ/kg)

LPC =

low-pressure compressor

LPT =

low-pressure turbine

mGT =

micro gas turbine

mHAT =

micro humid air turbine

NG =

natural gas

ORC =

organic Rankine cycle

PR =

pressure ratio

SG =

synthesis gas (syngas)

STIG =

steam-injected gas turbine

TET =

turbine exhaust temperature (°C)

TIT =

turbine inlet temperature (°C)