This is a repository copy of CO2-enhanced and humidified operation of a micro-gas turbine for carbon capture. White Rose Research Online URL for this paper: http://eprints.whiterose.ac.uk/127756/ Version: Accepted Version Article: Best, T., Finney, K.N. orcid.org/0000-0002-2166-5277, Ingham, D.B. et al. (1 more author) (2018) CO2-enhanced and humidified operation of a micro-gas turbine for carbon capture. Journal of Cleaner Production, 176. pp. 370-381. ISSN 0959-6526 https://doi.org/10.1016/j.jclepro.2017.12.062 Reuse This article is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs (CC BY-NC-ND) licence. This licence only allows you to download this work and share it with others as long as you credit the authors, but you can’t change the article in any way or use it commercially. More information and the full terms of the licence here: https://creativecommons.org/licenses/ Takedown If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request. eprints@whiterose.ac.uk https://eprints.whiterose.ac.uk/ CO2-ENHANCED AND HUMIDIFIED OPERATION OF A MICRO-GAS TURBINE FOR CARBON CAPTURE1 Thom Best 1, Karen N. Finney 2*, Derek B. Ingham 2 and Mohamed Pourkashanian 2 1 Energy Technology & Innovation Initiative, Faculty of Engineering, University of Leeds, LS2 9JT, UK 2 Energy 2050, Mechanical Engineering, University of Sheffield, S10 1TN, UK * Corresponding author T: +44 (0)114 215 7226; E: k.n.finney@sheffield.ac.uk ABSTRACT As greenhouse gas emissions are a key driver of climate change, sources of CO2 must be mitigated, particularly from carbon-intensive sectors, like power production. Natural gas provides an increasingly large percentage of electricity; however its lower carbon intensity is insufficient to make proportional reduction contributions to circumvent 2°C global warming. The low partial pressure of CO2 in its flue gas makes postcombustion capture more challenging increasing the CO2 in the exhaust assists in enhancing capture efficiency. This paper experimentally investigates the impact of the combination of humidified air turbines and exhaust gas recirculation to increase CO2 partial pressures, with the aim of evaluating their effects on emissions and turbine parameters at various turndown ratios. It was found that CO2 levels could be increased from 1.5 to 5.3 vol%, meaning more efficient post-combustion capture would be possible. CO2 and steam additions increased incomplete combustion when used together at high levels for low turndown ratios (below 60%), with CO increasing from 49 to 211 ppm and CH4 from 2.5 to 52 ppm; this effect was negated at higher power outputs. Turbine cycle humidification resulted in net improvements to the turbine efficiency, by up to 5.5% on a specific fuel consumption basis. Keywords: gas turbine; CO2 emissions; exhaust gas recirculation; humidified air turbine; post-combustion carbon capture. Highlights:  investigation of humidification and exhaust gas recycle on gas turbine performance  incomplete combustion products (CO and CH4) were only seen at low power outputs  system temperatures were reduced by using steam and CO2 injections  combustor efficiency was reduced, whilst turbine efficiency improved  augmented CO2 levels and reduced gas volumes can improve capture performance ABBREVIATIONS AND NOMENCLATURE C specific heat capacity CCS carbon capture and storage EGR exhaust gas recirculation FR flowrate FTIR Fourier transform infrared HAT humidified air turbine LCP Large Combustion Plant mGT micro-gas turbine NOx oxides of nitrogen 1 PACT ppm PT rpm SFC TC UHC UKCCSRC 。 1 Pilot-scale Advanced Capture Technology parts per million pressure transducer revolutions per minute specific fuel consumption thermocouple unburnt hydrocarbons UK Carbon Capture Storage Research Council turbine efficiency 1. INTRODUCTION Due to the overwhelming evidence of climate change, it is necessary to mitigate greenhouse gas emissions. Average global temperatures were 0.87°C warmer in 2015 than 1951-1980 (GISTEMP.Team, 2016), which could have dramatic consequences, impacting ~660m people by 2030 (Global Humanitarian Forum, 2009). One of the largest static emissions sources is electrical power generation, accounting for 25% of greenhouse gases (IPCC, 2014). Whilst renewables are being deployed, it is necessary to find ways to decarbonise the conventional supply and address the remaining carbonaceous fuels that are likely to be utilised (Huisingh, et al., 2015), which can be achieved through carbon capture and storage (CCS). 1.1 Background Whilst carbon capture has been demonstrated at scale for coal (SaskPower, 2016; NRG Energy Inc., 2017), research into CCS for gas-based power is ongoing, due to the differences in flue gas composition between these two fuels. Whilst coal generates relatively high CO2 concentrations, natural gas, combusted under fuel-lean conditions, produces an exhaust with considerably lower levels, which is challenging for solventbased capture that relies heavily on the concentration gradient between the flue gas and capture media. Furthermore, lean firing results in significant remaining O2 in the flue gas, which can cause oxidative solvent degradation. This also means there are large volumetric flowrates of gas, necessitating capture plants with a substantial footprint. Consequently, gas turbine modifications have been sought to increase the partial pressure of CO2 in the exhaust stream going to the capture plant to improve overall plant efficiency. 1.1.1 Exhaust Gas Recirculation Exhaust gas recirculation (EGR) recycles a proportion of the exhaust to the front end of the process to increase the CO2 content of the oxidiser and thus the flue gas. This can increase the capture process efficiency, however, it can also reduce combustor performance, due to the reduction in peak temperatures, especially at the turbine inlet (Mansouri Majoumerd, et al., 2014). Optimising the EGR ratio is therefore of great importance to ensure the maximum CO2 concentration is reached with minimal impacts on system performance. The energy requirements for cooling and recycling the flue gas also need to be considered, especially when applied to large-scale implementation. Several modelling and simulation studies quantify the impacts of EGR and find the ideal recycle ratio for different system setups; showing net efficiency gains for the whole integrated process (Li, et al., 2011a; Li, et al., 2011b; Mansouri Majoumerd, et al., 2014). EGR displaces air in the combustor, lowering oxygen availability; this depletion is the limiting factor, where incomplete combustion products, such as CO and unburned hydrocarbons (UHC) increase at inlet O2 levels below 16% (ElKady, et al., 2009; Evulet, et al., 2009). Blowout has been seen to occur experimentally below 14% (Ditaranto, et al., 2009). There have been few experimental studies in this area, but these show stable operation is possible with high levels of EGR (Røkke and Hustad, 2005; ElKady, et al., 2009; Evulet, et al., 2009; Jansohn, et al., 2011). Recycling 25% of the flue gases has been shown to increase flue gas CO2 levels to 8%, with 35% recycle giving exhaust emissions containing over 10% (ElKady, et al., 2009; Evulet, et al., 2009). This is dependent upon the fuel-air mix, with Li, et al. (2011b) showing 10% CO2 is only achieved with 55% EGR. Best, et al. (2016) experimentally investigated the whole system impacts of EGR, considering the effects on turbine and capture performance. This showed a reduction in turbine efficiency but an increase in CO2 levels, and thus a net capture process efficiency gain. 1.1.2 Humidified Turbine Cycles Humidified air turbines (HAT) use water/steam injections before combustion, with air displaced by water; when condensed out, the CO2 is then a relatively larger proportion of the total exhaust, enabling more efficient capture (Rao and Day, 1996; Jonsson and Yan, 2005). De Paepe, et al. (2012) found humidification improved efficiencies by up to 2.4%, however their humidification percentage was not quantified and emissions were not published. Mansouri Majoumerd, et al. (2014) looked at both HAT and EGR separately 2 through validated modelling, but did not investigated their combined impacts; they found improvements in efficiency with steam, but a reduction with EGR. In conventional turbines, the pressure ratio and turbine inlet temperature are the biggest factors driving turbine performance. The rate of moisture injection is a strong influence and the addition of steam increases the specific heat capacity (C) of the working fluid, implying greater heat transfer through the recuperator is achievable, increasing efficiency (Horlock, 1998; Li, et al., 2011a). Steam addition also increases the mass flow through the turbine, augmenting its specific power output (Jonsson and Yan, 2005). If steam is added post compression, the additional mass flow through the turbine (for which the compressor has not had to do work for) can improve system performance through greater power generation. HAT has been used historically to improve turbine efficiency and reduce emissions. Older low-NOx combustors used water for temperature control (Jonsson and Yan, 2005), since it acts as a diluent, reducing system temperatures. Although the addition of steam may appear to improve the power output, this does not take into account the energy of rendering the steam (Wan, et al., 2010). If water is added, efficiency gains would be reduced due to the energy required to convert the water into steam via evaporation. The addition of high levels of steam to create humidified air can cause combustion instabilities, shown by higher levels of UHCs and CO (Day, et al., 1999). Using appropriate levels of humidification, it is possible to reduce NOx without significantly increasing incomplete combustion; suggestions range from 5 to 14 vol% (Takahashi, et al., 2002; Li, et al., 2011a; De Paepe, et al., 2012; De Paepe, et al., 2013; Wei and Zang, 2013). 1.1.3 Combining CO2-Enahnced and Humidified Cycles HAT can improve turbine efficiency through increased mass flow post-compression, reducing compressor work. It may therefore be possible to negate some of the reduced turbine efficiency experienced with EGR by combining it with HAT. However HAT has its own detrimental impact on combustion, and hence all the benefits must be balanced. It may be the case that for the turbine utilised here, there may be a limited impact on combustion due to the lean fuel-air mix (Nikpey Somehsaraei, et al., 2014), hence HAT may provide significant performance/efficiency benefits. Simultaneously, as previously seen, CO2 enhancement can have limited impacts on combustion whilst providing significant capture efficiency gains (Best, et al., 2016). However in a combined cycle gas turbine with significantly leaner air mixes, both EGR and HAT may have significantly larger implications on the combustor, and thus cause flame instabilities and lean blowout issues. These methods will however also address the other challenges posed by CCS on natural gas. 1.2 Aims and Objectives This paper explores the possibility of combining EGR (CO2 injection to the compressor inlet) and HAT (steam injection at the compressor outlet) on a mGT, with the purpose of evaluating the impacts on turbine operation and assessing the improvements to post-combustion capture through the resultant increase in flue gas CO2 levels. This article explores the relative impacts of each process separately and then in combination, compared to a baseline. The effects on emissions as an indicator of incomplete combustion, characterised by CO and UHC, and on a range of turbine parameters are quantified, including engine speed, fuel consumption and efficiency. Whilst others have simulated the impact of these processes, e.g. Taimoor, et al. (2016), this is the first paper to experimentally investigate the combination of both techniques. 2. EXPERIMENTAL METHODOLOGY The UK Carbon Capture Storage Research Centre (UKCCSRC) Pilot-scale Advanced CO2 Capture Technology (PACT) Facilities are the national specialist research and development facilities for combustion and carbon capture. The core site houses combustion rigs and a solvent-based, post-combustion carbon capture plant. Comprehensive analytical facilities are available to monitor the impacts of changing the operating regimes, including numerous turbine parameters and thorough exhaust gas characterisation. 3 2.1 Turbec T100 PH Gas Turbine The PACT facility currently houses two natural gas Turbec T100 PH micro-gas turbines (mGT). Each has an electrical power generation capacity of up to 100 kW, with an electrical efficiency of 30%, and both have combined heat units that can recover a further 165 kW of thermal power, improving overall system efficiencies to ~80% (Turbec, 2009). This is a simple Brayton cycle, operating with the addition of a recuperator before the combustion chamber. This paper utilised the Series 1 mGT, which has been extensively modified to include both CO2 pre-compression and steam post-compression, as depicted in Figure 1, to simulate EGR and HAT. The single-stage centrifugal compressor compresses ambient air, and CO2 when injected, to pressure ratios of up to 4.5:1. This can then have steam added at this point, before passing through the recuperator and then to the combustor. The recuperator increases the oxidiser temperature to improve the electrical efficiency, exchanging heat with the exhaust gases from the turbine outlet; the efficiency of which is enhanced by both the steam and CO2 addition, which increase C. The lean pre-mixed natural gas flame in the combustor is swirl-stabilised to give low emissions of CO, UHC and NOx. The products expand through the turbine to near atmospheric pressure and drive the single shaft for the compressor and generator. After passing through the recuperator, the exhaust gases go through a counterflow water-gas heat exchanger. However in this experimental set up, to maintain high exhaust temperatures and reduce the possibility of condensate from the HAT process, the heat exchanger was drained and the exhaust gas remained hot. Figure 1: Schematic of the components of the Turbec T100 mGT system at PACT, including the modifications for CO2 and steam injection, and thermocouples (TC), pressure transducers (PT) and flowrate monitors (FR). 2.2 Experimental Conditions Three key parameters were varied for the tests conducted; the steam and CO2 levels across various turn down ratios. Steam injection varied between 0 and 40 kg/hr, CO2 enhancement from 0 to 125 kg/hr and power outputs between 50 and 70 kW. The test matrix was designed accordingly and the 44 permutations of the conditions tested are outlined in Table 1. All tests were carried out over a minimum of 15 minutes of continuous stable operation as recommended by ISO 2314 (BSI, 2010) and each condition was repeated at least twice, with data averaged for each. 2.3 Data Acquisition All testing was carried out at the PACT Core Facilities, using the modified Series 1 mGT. The experimental set-up involved extensive monitoring systems for various gas turbine parameters and flue gas species. Firstly, the gas turbine parameters were assessed with the instrumentation added in-house throughout the turbine to ensure full systems monitoring and more comprehensive characterisation of the cycle (Figure 1). Data were collected using a Compact RIO-9022 Real-Time controller, displayed and recorded with a custom LabView program. K-type thermocouples were installed for additional temperature measurements in °C. Pressures (in bar g) were measured using Rosemount pressure transmitters, with errors up to 0.065% of the 4 calibrated range, equivalent to 0.8 mbar for PT1, PT3 and PT5 and 7.5 mbar for PT2, PT4 and PT6. The instrument errors (Table 2) and standard deviations (Table 3) are too small for precise depiction on graphs and hence error bars are not plotted. Secondly, internallyT instrumentation was recorded (via WinNAP software), which monitors the following parameters:  air inlet temperature (°C)  turbine inlet (calculated) and outlet temperatures (°C)  power output and set point (kW)  turbine speed (rpm and % maximum of 70,000) Table 1: Test matrix outlining the conditions used. STEAM ADDITION (kg/hr) 20 0 CO2 ADDITION (kg/hr) 40 0 50 75 100 125 0 50 75 100 125 0 50 75 100 125                                             POWER (kW) 50 55 60 65 70 Most species in the flue gas were measured using a Gasmet DX4000 FTIR, Fourier transform infrared spectroscopy. This included the main combustion gases, such as CO2, H2O, CO and speciation of UHCs, as well as NOx. In this study, oxygen was monitored using a paramagnetic oxygen transducer in a ServoFlex Mini Multi-Purpose 5200 analyser. The fuel flow (in m3/hr) was measured using a Quantometer turbine flow meter, with a maximum relative error of 0.63% as obtained from factory calibration (Table 2). Table 2: Instrument standard errors for the devices used with the mGT. INSTRUMENTAL ERROR INSTRUMENT Quantometer K-type thermocouples Rosemount pressure transmitters 2051CDC2A Rosemount pressure transmitters 2051TG2A % 0.63 0.40 0.07 0.07 unit 1.5°C 0.8 mbar 7.5 mbar Table 3: Maximum standard deviation errors for the instruments used with the mGT. INSTRUMENT MAXIMUM STANDARD DEVIATION baseline EGR & HAT GASMET DX4000 FTIR CO (ppm) CH4 (ppm) CO2 (vol%) NOx (ppm) 17.3 3.9 0.02 1.3 17.3 6.0 0.02 1.2 TURBEC T100 Engine speed (rpm) Power (kW) Temperatures (°C) 59 0.5 0.6 139 0.5 0.6 ADDITIONAL INSTRUMENTATION Pressures (mbar) 0.01 Temperatures (°C) 0.5 0.01 11.6 5 2.4 Emission Corrections and Data Reporting All emissions are reported on a dry basis. NOx results were corrected to the industrial standard of emissions reporting and EC Large Combustion Plant (LCP) Regulation standards at 15% O 2, 101.3 kPa and 273.15 K (European Commission, 2006). Correcting NOx to 15% O2 however may not be appropriate, since the O2 in the exhaust will be lower than in turbines that do not utilise EGR or steam addition; the lower O 2 results mean NOx appear to be higher (ElKady, et al., 2009). A better comparison is on the basis of the fuel combusted and net power output from the turbine, giving a NOx emissions index (g/kg of fuel and g/kWh, corrected to 15% O2). The mGT however does not fall under the EC LCP regulations, as the efficiency is below 35%, where NOx is limited to 50。/35 (。 is turbine efficiency, giving 34 mg/m3 here). Local air quality standards, such as those of the Greater London Authority (AMEC, 2013), do apply to small mGTs though, and therefore imposes a 5 g/kWh limit, which the turbine achieves under normal operating conditions. Emissions of CO and UHCs, measured with the FTIR, were used as an indicator of combustor performance. Turbine efficiency was calculated using the net calorific value of the fuel, its flowrate and the power output. 3. RESULTS AND DISCUSSION The modified mGT was utilised for these tests, with its additional capabilities for steam and CO 2 to simulate the processes of HAT and EGR. During the commissioning phase, the turbine power electronics were damaged by water ingress from the condensation of the injected moisture. After its replacement, the heat exchanger was drained to prevent condensate back flow into the turbine housing, due to the higher exhaust gas temperatures. Condensate was not seen at the new bleed spots, with no moisture observed within the turbine housing. The baseline and CO2-enhanced results have been discussed previously in much detail by Best, et al. (2016); the results are briefly reported here for comparative purposes. 3.1 Baseline Performance The results for the baseline tests are summarised in Table 4. Whilst fuel consumption increases with power output, the specific fuel consumption decreases, resulting in improved efficiencies at higher load factors. There is a notable increase in performance with turndown ratio, with the lowest efficiency and highest emissions at 50 kW. As the power output increases, the efficiency improves, with better combustion at 60 kW, particularly CO, which highlights incomplete combustion. Since the system operates under very fuellean conditions, there is no issue with oxygen availability; thus the fuel-oxidiser mixing is likely to be inadequate here, resulting in lower temperatures within the combustion zone. Furthermore, with high levels of both O2 and NOx present at low turndown ratios, this indicates insufficient diffusion, combined with a lower pressure ratio that suggests longer residence times. Higher NOx emissions at 70 kW are expected, due to the hotter temperatures throughout the system than for low load factors. Additional assessments of baseline performance were considered by Best, et al. (2016). 3.2 mGT Performance using CO2 Injections for EGR Operation Various turndown ratios were employed with CO2 enhancements of up to 125 kg/hr; the results are summarised in Table 4. The use of CO2-enhanced operation slowed turbine speeds compared to the baseline, due to the denser CO2 enabling the same mass throughput at lower volumetric flowrates. System temperatures were notably reduced compared to the baseline, due to the higher C of the added CO2. The turbine inlet temperature, for example, was ~10°C lower for the EGR tests. In terms of emissions, NOx decreased for the comparative CO2 tests, as thermal NOx formation is favoured by higher temperatures; the temperatures throughout the system were lower when operating with EGR. CO was notably increased by CO2 injections, but only for baseload operation (50 kW), as demonstrated in Table 4. Low emissions levels were reported for power outputs 60 kW. The higher levels of CO2 injection at high turndown ratios 6 resulted in considerably more CO2 in the exhaust gas, which would benefit post-combustion capture performance. Further details of the analysis of CO2-enhanced operation can be found in Best, et al. (2016); the results all corroborate the assessments of previous work (e.g. ElKady, et al., 2009; Evulet, et al., 2009). Table 4: Summary of baseline and CO2-enhanced performance of the mGT (after Best, et al. (2016)). 0 kg/hr CO2 INJECTION 50 kW 60 kW 70 kW 75 kg/hr CO2 INJECTION 50 kW 60 kW 70 kW 125 kg/hr CO2 INJECTION 50 kW 60 kW 70 kW 23.8 0.38 59,784 21.36 27.3 0.36 62,476 22.33 30.6 0.35 64,886 23.07 23.4 0.37 58,372 21.98 27.0 0.36 61,384 22.89 30.2 0.34 63,857 23.86 23.6 0.37 58,481 21.89 26.9 0.35 61,247 22.94 30.3 0.34 63,707 23.76 TEMPERATURES AND PRESSURES TC2 compressor outlet temp (°C) PT2 compressor outlet pressure (bar) Turbine inlet temperature (°C) Turbine outlet temperature (°C) 161 2.2 883 643 174 2.5 898 644 186 2.8 913 643 153 2.2 872 641 167 2.5 888 641 180 2.8 905 642 154 2.2 873 641 167 2.5 888 641 180 2.8 904 642 FLUE GAS COMPOSITION O2 (vol%, dry) CO (ppm, dry) NOx (O2 corrected, g/kWh) CO2 (vol%, dry) 19.0 49 1.63 1.5 19.0 1 1.19 1.5 19.0 2 1.56 1.6 17.9 154 1.72 3.6 18.0 3 0.33 3.6 17.9 4 0.95 3.5 17.8 132 1.65 5.0 17.7 5 0.59 4.9 17.6 4 0.75 4.7 VARIABLE TURBINE PARAMETERS Fuel flowrate (m3/h) Specific fuel consumption (kg/kWh) Engine speed (rpm) Turbine efficiency (%) 3.3 mGT Performance using Steam Injections for HAT Operation With the addition of steam post-compression, it is expected that the CO2 concentration would increase on a dry basis in the exhaust, since steam is displacing the air; this is observed in Figure 2. There is a good correlation of higher CO2 in the exhaust with steam addition with the volume percentage fractionally larger due to the HAT process after the water has been condensed out (emissions converted to a dry basis), supporting the conclusions of Li, et al. (2011a) and De Paepe, et al. (2012), among others. This trend is observable at all power outputs. Figure 3 shows the impacts of steam injection on engine speed. As with the use of CO2-enhanced operation above, steam also results in a denser oxidiser and thus the same mass throughput for the system can again be achieved at a lower volumetric flow, slowing turbine speeds by ~2000 rpm, as also found by De Paepe, et al. (2012). Whilst steam injection overall had a significant impact on the rotational speed of the engine, there was much less difference between the two steam injection cases tested compared to the baseline. 1.70 CO2 Concentration (vol%, dry) 1.65 1.60 1.55 1.50 1.45 1.40 45 50 55 60 65 Power Output (kW) 70 75 Figure 2: CO2 level measured in the exhaust at various levels of humidification for different power set points. 7 66,000 65,000 Engine Speed (rpm) 64,000 63,000 62,000 61,000 60,000 59,000 58,000 57,000 45 50 55 60 Power Output (kW) 65 70 75 Figure 3: Engine speeds at various levels of humidification for different power set points. 3.4 mGT Performance using Steam and CO2 Injections for HAT and EGR Operation The use of steam and CO2 injections combined were then assessed. The impacts of changing these primary variables on system temperatures, emissions and overall efficiency are considered. 3.4.1 Impacts on System Temperatures A number of system temperatures were monitored for the mGT, as outlined in Section 2.3. Both steam and CO2 have the ability to alter the C and their impacts are considered here. Cold-side Recuperator Inlet: Figure 4 outlines the temperatures recorded at TC2 the compressed oxidiser at the cold-side inlet to the recuperator, which included the steam and/or CO2, where used. The baseline tests, with no steam or CO2, had the highest temperatures. This is consistent for all power outputs, but temperatures here increase as the load factor increases; the highest temperature for the highest power output. The impact of simulated EGR reduced the temperatures here consistently for all power outputs by ~7°C. The use of steam however was shown to have a greater impact. Temperatures for all power outputs when using steam injection were lower than for both the baseline and CO2-enhanced cases; the maximum difference seen was 44°C. The temperatures recorded when using 40 kg/hr steam are slightly higher than the 20 kg/hr cases as the steam temperatures required to obtain the higher flowrates were greater. Turbine Inlet and Outlet: Figure 5, which outlines the turbine inlet temperatures for power outputs of 50 and 60 kW for a range of cases, reveals that this is greatly impacted by the power output. Those for 50 kW (Figure 5a) are much lower than those for 60 kW (Figure 5b). Furthermore, the use of both HAT and EGR affect these temperatures. As the level of CO2 injection increases, temperatures decrease due to the alterations in the C of the flue gas. The addition of steam also increases the C and thus these temperatures were also found to be lower across all power outputs. Overall, the temperatures at the turbine inlet were up to 15°C lower. As can be seen for Figure 5, the temperature appears to be less affected and more consistent with the addition of CO2 to an already steam-enhanced system there are smaller differences between the baseline and steam cases with high levels of CO2. This is probably because the C is already higher here, and the combined addition of steam is a less radical change. The turbine controls a number of parameters to ensure a turbine outlet temperature of ~645°C is maintained. Whilst steam and CO2 altered this temperatures, it was by relatively small amounts (<5°C). The CO2 additions resulted in lower temperatures at higher levels of injections, however there were no trends seen for steam injections. 8 200 Cold-side Recuperator Inlet Temperaute (°C) 175 150 125 100 75 50 25 0 45 50 55 60 Power Output (kW) 65 70 900 900 895 895 890 890 Turbine Inlet Temperature (°C) Turbine Inlet Temperature (°C) Figure 4: Impacts of steam and CO2 injections (separately) on the temperature at TC2. 885 880 875 870 865 885 880 875 870 865 860 860 0 50 100 CO2 Enhancement Level (kg/hr) (a) 150 0 50 100 CO2 Enhancement Level (kg/hr) (b) 150 Figure 5: Impacts of steam and CO2 injections on turbine inlet temperature, for (a) 50 kW; and (b) 60 kW. 3.4.2 Impacts on Emissions Although low oxygen levels in the combustor cause unstable flames and blowout, the relatively fuel-lean conditions of the mGT ensure this is not an issue. Even with high levels of steam and CO2 injections for low power outputs (40 kg/hr steam and 125 kg/hr CO2 at 50 kW), where the oxidiser flowrates are lowest, the remaining O2 in the flue gas was still 17.7 vol% on a dry basis. However flame speeds and temperatures are reduced due to the increased C under these conditions (De Santis, et al., 2016), and the combination of EGR and HAT further exacerbate this effect compared to when they are used individually. As previously considered in Best, et al. (2016), it has been shown that system modifications, such as EGR (and for HAT and/or EGR herein), have the greatest impact on emissions at the lowest turndown ratios. This is due to the slightly richer air-fuel ratio at the bottom of the turbines turndown ratio, with changes in the oxidiser composition from the steam and/or CO2 injections having more effect and the mixing being more diffuse. 9 CO2 Levels: With the additions of both CO2 and steam, a linear CO2 increase was seen, as observed in Best, et al. (2016) for the use of CO2 injections alone. At 40 kg/hr steam enhancement with increasing CO2 injection across the turndown ratios of 50 and 60 kW (shown in Figures 6a and 6b, respectively), the highest concentration of CO2 was found, as would be expected; peak CO2 concentrations of 5.3 vol%, dry were seen for the most extreme case at the lowest turndown ratio. This is a particularly strong trend and is consistent for both power outputs and across all levels of CO2 enhancement. This quantifies the extent to which the CO2 concentrations can be augmented by the combination of the HAT and EGR. 6.5 6.5 50kW 0 kg/hrBaseline (baseline) 6.0 50kW 20kg 20 kg/hr 50kW 40kg 40 kg/hr 5.0 60kW 20kg 20 kg/hr 5.5 CO2 Concentration (vol%, dry) CO2 Concentration (vol%, dry) 5.5 60kW Baseline 0 kg/hr (baseline) 6.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 1.0 0.5 0.5 0.0 60kW 40kg 40 kg/hr 5.0 0.0 0 50 100 CO2 Enhancement Level (kg/hr) (a) 150 0 50 100 CO2 Enhancement Level (kg/hr) (b) 150 Figure 6: CO2 concentrations measured in the exhaust from the mGT at various levels of humidification for different levels of CO2 enhancement for, (a) 50 kW and (b) 60 kW. CO Emissions: Carbon monoxide levels were monitored in the exhaust as an indicator of incomplete combustion; Figure 7 outlines the impact of steam and CO2 injections. Even with a small addition of steam, emissions of CO2 increase dramatically for low power outputs (50-55 kW). Greater additions of steam result in higher CO emissions with levels over 200 ppm. Furthermore, these are significantly affected beyond that of the simple EGR, as indicated by Table 4 and seen in Figure 7. For CO2 injections alone, the peak CO concentrations in the flue gas were recorded as ~150 ppm. When combining both steam and CO2 injections, the high levels seen with humidified operation alone are increased even further. Here, the highest CO levels were recorded for maximum steam (40 kg/hr) and CO2 (125 kg/hr) injection rates at the lowest turndown ratio 260 ppm for 50 kW. Elevated levels of CO were only seen for low power outputs (50-55 kW) and were much reduced by increasing the power output; by 65 kW, levels were below 5 ppm for all cases. This shows increased incomplete combustion occurs only at low load factors when mixing is insufficient and flame speeds are reduced, resulting in only partial fuel combustion. Therefore, the negative impacts of operating with high levels of EGR and HAT can be mitigating by increasing the power output of the turbine above baseload operation. Overall efficiencies are greater for higher power outputs, as shown in Table 4 and considered further below (Section 3.4.3). Unburned Hydrocarbon Content: In addition to CO, emissions of methane were also monitored in the exhaust as an indicator of incomplete combustion, along with a number of other hydrocarbon species. A similar trend in CO and CH4 emissions (Figures 7 and 8, respectively) can be observed. The use of EGR or HAT alone increased emissions compared to the baseline 125 kg/hr CO2 increased CH4 levels from 2.5 to 22 ppm at baseload and 40 kg/hr steam increased CH4 to 52 ppm. However, the combination of EGR and HAT further elevated the levels of CH4 present, increasing the concentration to over 70 ppm. This clear 10 evidence of increased incomplete combustion is most likely due to reduced flame speeds (De Santis, et al., 2016) and lower efficiencies. As with the results for CO above, increasing the power output rapidly improve efficiency (Table 4) and reduces the levels of CH4 with minimal emissions at 60+ kW for all cases. Emissions of other UHCs, such as ethane, showed a similar trend increasing with the level of HAT and EGR used, but decreasing to minimal levels by 60 kW. This is where the temperatures would be higher and mixing improved, resulting in faster flame speeds and superior combustion efficiencies, minimising the levels of incomplete combustion products. 350 CO Concentration (ppm, dry) 300 baseline - no steam/CO2 CO2 20 kg/hr steam - no CO2 CO2 40 kg/hr steam - no CO2 CO2 0 kg/hr steam - 125 kg/hr CO2 CO2 20 kg/hr steam - 125 kg/hr CO2 CO2 CO2 40 kg/hr steam - 125 kg/hr CO2 250 200 150 100 50 0 45 50 55 60 Power Output (kW) 65 70 Figure 7: CO concentrations as an indicator of incomplete combustion, comparing various levels of humidification at different power settings, for 0 and 125 kg/hr CO2 injection. 100 baseline - no steam/CO2 CO2 20 kg/hr steam - no CO2 CO2 40 kg/hr steam - no CO2 CO2 CO2 0 kg/hr steam - 125 kg/hr CO2 20 kg/hr steam - 125 kg/hr CO2 CO2 40 kg/hr steam - 125 kg/hr CO2 CO2 CH4 Concentration (ppm, dry) 80 60 40 20 0 45 50 55 60 Power Output (kW) 65 Figure 8: CH4 concentrations at various levels of humidification at different power settings, for 0 and 125 kg/hr CO2 injection. 11 70 NOx Emissions: The NOx Emissions Index was calculated based on the experimental data from the FTIR using the method outlined in Section 2.4. As previously investigated in Best, et al. (2016), NOx can be reduced by CO2 injections, due to the alterations in C of the working fluid, limiting thermal NOx formation (De Santis, et al., 2016). From Figure 9, it can be seen that there was much overlap in the results herein, particularly at higher turndown ratios that are likely to be used in power stations. Despite the fact that both process modifications increase the C of the oxidiser, and thus have the ability to reduce system temperatures, as reported in Section 3.4.1, the impact was not clearly seen here in terms of reductions in NOx levels for the combined steam and CO2 cases compared to the baseline. Emission levels were similar for all tests at 65 kW. There was little variation in the turbine outlet temperature monitored, and this may impact NOx formation despite the changes in oxidiser composition. Kinetic modelling of these variations may enable further insight here, specifically in terms of the variations in combustion/flame temperatures, flame stability and emissions between the different operating conditions. 4.0 baseline - no nosteam/CO2 steam/CO2 0 kg/hr of steam125 125kg/hr kg/hrCO of 2CO2 steam 20kg Steam 125kg125 CO2kg/hr CO2 20 kg/hr steam 40kg H2Osteam 125kg CO2 40 kg/hr 125 kg/hr CO2 NOx Emission Index (g/kWh) 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 45 50 55 60 65 70 Power Output kW Figure 9: NOx Emissions Index for various levels of humidification at different power settings, for baseline and 125 kg/hr CO2 injection. 3.4.3 Turbine Parameters and System Efficiency A number of turbine parameters were monitored and from these, a range of key performance indictors can be established, such as the specific fuel consumption and the overall system efficiency. Engine Speed: Figure 10 delineates the impacts of steam and/or CO2 additions on the mGT engine speed, which both have notable implications. As outlined for HAT and EGR individually, the use of either of modification reduces the engine speed significantly, since a lower volumetric flowrate through the system is achieved for the same mass flow, due to the increased density of the steam and CO 2. CO2 additions decreased the turbine frequency by up to 1600 rpm (in line with previous results published by Best, et al. (2016)), whereas humidification reduced the turbine frequency by up to 1850 rpm. This reduction is ~2.9%, similar to the savings in specific fuel consumption of 3.2% at 50 kW with 40 kg steam addition considered in the subsequent section. When these processes are combined, further reductions in engine speed were observed; Figure 10 clearly shows that the highest flowrates of steam impacted the most on this parameter when used in conjunction with simulated EGR. The largest decrease in turbine engine speed (of ~2500 rpm) was observed at 40 kg/hr steam addition with high levels of CO2 enhancement. 12 60,500 60,000 Engine Speed (rpm) 59,500 59,000 58,500 58,000 57,500 57,000 0 20 40 60 80 100 120 140 CO2 Enhancement Level (kg/hr) Figure 10: Turbine engine speed for various levels of humidification and CO2 injection at 50 kW. Specific Fuel Consumption (SFC) and Efficiency: With the addition of CO2, it would be expected that the combustor would experience reduced efficiency due to the increased C and reduced O2 availability, caused by the displacement of air with CO2. In Best, et al. (2016) and ElKady, et al. (2009), this is shown to occur, and was replicated here by the presence of increased levels of CO and CH4 (Figures 7 and 8). These emissions strongly evidence the reduction in complete combustion and lower flame temperatures. In terms of SFC and overall system efficiency however, the trends are quite different. Figure 11 shows that the baseline tests with no steam or CO2 had the highest SFC per unit energy generated across all power outputs. Furthermore, Figure 12 reveals the baseline cases had the lowest efficiency. The efficiency takes into consideration a number of factors, including the power generated and fuel flowrate. The efficiency improves with power output, as seen in Figure 12a and 12b for 50 and 60 kW, respectively. The use of CO 2 injection alone improves the SFC and efficiency slightly for all power settings. The same trend is seen for the addition of just steam, although the changes seen with this are more prominent. These observed improvements may be due to the reduced engine speeds, for the reasons considered in the preceding section, which appear to outweigh the apparent losses in combustion performance indicated in the emissions at low turndown ratios. With the addition of both steam and CO2 though, there appears little difference in the results for SFC between these and the steam only cases. It would appear that whilst HAT improves the SFC beyond that of the baseline turbine performance, the impact of this alone is so strong that in combination with CO2 enhancement little difference is seen. The improvements noted at 20 and 40 kg/hr steam injection alone is 3-5% better than the baseline over the operating envelope tested. Figure 12 also reveals that whilst CO2 and steam separately improve the efficiency of the system (as found by Li, et al. (2011a) and De Paepe, et al. (2012) for EGR and HAT, respectively), there is little change in the efficiencies when both are utilised; this is noted not only for power outputs of 50-60 kW, as shown in Figures 12a and 12b, but also for the other turndown ratios investigated. This is the maximum efficiency achievable under these conditions and whilst steam addition has improved this compared to the baseline and EGR-only cases, greater levels of steam injection appear to have little impact and the efficiency cannot be improved further; the maximum efficiency has already been attained for these conditions. 13 Specific Fuel Consumption (kg/kWh) 0.39 0.38 0.37 0.36 0.35 0.34 0.33 45 50 55 60 65 70 75 Power Output (kW) 24.0 24.0 23.5 23.5 23.0 23.0 System Efficiency (%) System Efficiency (%) Figure 11: Specific fuel consumption for various levels of humidification at different power settings, for 0 and 125 kg/hr CO2 injection. 22.5 22.0 21.5 22.5 22.0 21.5 21.0 21.0 20.5 20.5 20.0 20.0 0 50 100 CO2 Enhancement Level (kg/hr) (a) 0 150 50 100 CO2 Enhancement Level (kg/hr) (b) 150 Figure 12: Calculated efficiencies for various levels of humidification at different levels of CO2 injection up to 125 kg/hr for, (a) 50 kW and (b) 60 kW. 3.5 Impact of EGR and HAT on Gas-CCS Performance Although several previous studies have investigated the use of EGR or HAT, as outlined in Section 1.1, none have combined these process modifications to assess their collective impacts; thus it is difficult to drawn direct comparisons. Where relevant in the above sections (Sections 3.5 through 3.4), the new data presented here has been compared to that previously published, and they concur with the main findings where similar trends were observed (e.g. ElKady, et al., 2009; Evulet, et al., 2009; Li, et al., 2011a; De Paepe, 14 et al., 2012). Further investigation however is required into the combination of different process modifications, including humidification, EGR, supplementary firing, etc. to fully assess the potential of gas turbine improvements that could be applied positively in gas-CCS scenarios. These could augment overall efficiencies, as well as increase the CO2 concentrations for enhanced capture performance. In addition to the implications for the mGT here, upscaling the technology would also result in further factors to be considered for commercial-scale combined cycles. When operating with real EGR, for example, cooling the recycled stream sufficiently will be vital to ensure optimal performance, maintaining high efficiencies in the compressor. Additional intercooling here though, between compressor stages, may not be appropriate for combined HAT and EGR operation, since this may result in the condensing of moisture at inappropriate locations depending on when the moisture is introduced to the cycle. Other adaptations to this system could potentially further improve efficiencies, such as the use of selective EGR (Merkel, et al., 2013) or a supercritical bottoming cycle (Li, et al., 2011a). In terms of the micro-turbine here, to improve efficiencies would entail significant overhaul (renewal/reconditioning) of the compressor and turbine seals and blades. Solvent-based, post-combustion capture is an energy intensive process, due to the regeneration of the capture solvent large amounts of energy are required to break the bonds between the CO2 and the solvent, releasing it for compression. As gas turbines have a low partial pressure of CO 2 in the flue gas that is treated, this means more energy is required per tonne of CO2 captured; increasing the partial pressure of CO2 reduces this energy consumption (Akram, et al., 2016). Combining the options of EGR and HAT has been shown to increase the CO2 concentration in the flue gas more than either of these technologies on their own (Figure 6). These increases equate to reductions in specific reboiler duty for the 50 kW case here, an increase in CO2 from 5.1 vol% without steam to 5.3 vol% with steam under EGR conditions could mean a drop in specific reboiler duty of over 0.1 MJ/kg CO2 captured (Akram, et al., 2016). This would mean a saving in energy consumption of over 1%, in addition to the reduction in SFC of the turbine of 3%. At 60 kW, similar improvements are noted with CO2 increasing from 4.9 to 5.1 vol% with 40 kg/hr steam addition, again reducing the specific reboiler duty by over 0.1 MJ/kg CO2 captured (Akram, et al., 2016), whilst turbine SFC is reduced by 4%. These reductions in reboiler duty may appear to be modest but on the larger scale can add up to significant energy and thus cost reductions. These are even more pronounced when the compared to a baseline of no EGR or HAT. When operating with real EGR at scale, this would result in reduced volumes of flue gas to the capture plant, further minimising costs (Li, et al., 2011a). Whilst these tests have been conducted at a pilot-scale for economic reasons, scaling-up these technologies will be required for the widespread deployment of CCS from natural gas. Their deployment can have much larger benefits, in terms of reducing energy consumption and lowering the costs of capture at larger scales. There may however be some negative consequences of the use of EGR/HAT under certain operating conditions and thus these need to be highlighted to ensure such process modifications are deployed under the correct circumstances. As seen from the discussion on emissions (for both CO and CH4 in Section 3.4.2), elevated concentrations of incomplete combustion products were noted for low turndown ratios when high levels of simulated EGR and HAT were used in combination; these however rapidly decreased as the power output increased. Gas turbine plants are utilised for their high degree of flexibility, enabling load following and rapid response times to ramping requirements. Moreover, these are often used for baseload power and thus do not often run at full output, generating electricity at relatively low levels compared to their maximum capacity. If systems operate under such conditions with high levels of EGR and HAT though, this could cause potential problems with emissions compliance. Increasing the power output has been shown to mitigate these issues, however this may impact on grid supply and stability. Furthermore, the ramping behaviour of the turbines would also necessitate a flexible capture strategy to deal with the variations in the power output and thus in the exhaust gas flowrates received by the capture plant. 4. CONCLUSIONS Decarbonising conventional power is vital to limit climate change and gas-CCS could significantly address this issue. There are challenges associated with this, however modifications, such as EGR and HAT, 15 can optimise the process. These were experimentally researched here with a mGT, to assess their collective impacts, through CO2 and steam injections, evaluating the effects on emissions and turbine parameters at various turndown ratios. It was found that CO2 levels in the exhaust could be substantially increased when using both methods, more so than when used individually; enabling more efficient capture. Peak CO2 concentrations of 5.3 vol% were recorded for the most extreme levels of steam and CO2 augmentation. Other benefits included net improvements to the turbine efficiency, by up to 5.5% on a SFC basis, indicating an offset in the reduction in turbine efficiency caused by EGR. This is due to the higher density working fluid, which results in lower engine speeds (reduced by up to ~2500 rpm) for high levels of EGR and HAT. Since the engine is working less hard to achieve the same throughput, efficiencies are improved. Some negative consequences were seen under certain conditions and it would be advisable to avoid these when operating at large-scale. Incomplete combustion increased at low load factors, but was rapidly negated as power outputs increased. Emissions were negligible above 60 kW <3 ppm for UHCs and ~40 ppm for CO. This was due to poor fuel-oxidiser mixing, lower pressures and the higher C of the modulated working fluid, which reduces system temperatures, flame speeds and reaction rates. Measured temperature differences of up to 44°C were seen, however greater differences in combustion temperatures (70°C or more), particularly in the flame region, have been predicted (De Santis, et al., 2016). Whilst turbine cycle humidification alone has large impacts, the combination with EGR results in greater levels of CO2 in the flue gas and thus reductions in reboiler duty, allowing more efficient capture. Furthermore, the efficiency gains will help offset the energy consumption for implementing these developments. The use of either/both of these process modifications could be a key enabling technology for the widespread deployment of efficient gas-CCS. ACKNOWLEDGEMENTS The authors are very grateful to both the UK Engineering and Physical Sciences Research Council (EPSRC Gas-FACTS EP/J020788/1 and SELECT EP/M001482/1) and the University of Leeds Low Carbon CDT, for their financial support of this project. The authors acknowledge that the UKCCSRC PACT Facilities, funded by the Department for Business, Energy and Industrial Strategy and the Engineering and Physical Sciences Research Council, have been used for experimental work reported in this publication. REFERENCES Akram, M., Ali, U., Best, T., Blakey, S., Finney, K.N., Pourkashanian, M., 2016. Performance evaluation of PACT pilot-plant for CO2 capture from gas turbines with exhaust gas recycle. International Journal of Greenhouse Gas Control 47, 137-150. AMEC, 2013. Greater London Authority Air Quality Support: Biomass and CHP Emission Standards. Best, T., Finney, K.N., Ingham, D.B., Pourkashanian, M., 2016. Impact of CO2-enriched combustion air on micro-gas turbine performance for carbon capture. Energy 115, 1138-1147. BSI, 2010. Gas Turbines - Acceptance Tests ISO 2314. BSI Group. Day, W.H., Kendrick, D., Knight, B., Bhargava, A., Sowa, W., Colket, M., Casleton, K., Maloney, D., 1999. HAT cycle technology development program. netl.doe.gov/publications/proceedings/99/99ats/2-8.pdf. De Paepe, W., Delattin, F., Bram, S., De Ruyck, J., 2012. Steam injection experiments in a microturbine thermodynamic performance analysis. Applied Energy 97, 569-576. A De Paepe, W., Delattin, F., Bram, S., De Ruyck, J., 2013. Water injection in a micro gas turbine - Assessment of the performance using a black box method. Applied Energy 112, 1291-1302. 16 De Santis, A., Ingham, D.B., Ma, L., Pourkashanian, M., 2016. CFD analysis of exhaust gas recirculation in a micro gas turbine combustor for CO2 capture. Fuel 173, 146-154. Ditaranto, M., Hals, J., Bjørge, T., 2009. Investigation on the in-flame NO reburning in turbine exhaust gas. Proceedings of the Combustion Institute 32, 2659-2666. ElKady, A.M., Evulet, A.T., Brand, A.R., Urson, T.P., Lynghjem, A., 2009. Application of exhaust gas recirculation in a DLN F-Class combustion system for postcombustion carbon capture. Journal of Engineering for Gas Turbines and Power 131 (034505), 1-6. European Commission, 2006. Integrated Pollution Prevention and Control: Reference Document on Best Available Techniques for Large Combustion Plants. eippcb.jrc.ec.europa.eu/reference/BREF/lcp_bref_0706. pdf. E AT EK AM B AR C D O GE NOx combustors utilizing exhaust gas recirculation for post-combustion carbon capture. Energy Procedia 1, 3809-3816. GISTEMP.Team, 2016. GISS Surface Temperature Analysis (GISTEMP). data.giss.nasa.gov/gistemp. Global Humanitarian Forum, 2009. The Anatomy of a Silent Crisis. Global Humanitarian Forum. Horlock, J.H., 1998. Heat exchanger performance with water injection (with relevance to evaporative gas turbine (EGT) cycles). Energy Conversion and Management 39, 1621-1630. Huisingh, D., Zhang, Z., Moore, J.C., Qiao, Q., Li, Q., 2015. Recent advances in carbon emissions reduction: Policies, technologies, monitoring, assessment and modeling. Journal of Cleaner Production 103, 1-12. IPCC, 2014. Climate Change 2014 Synthesis Report Summary for Policymakers. ipcc.ch/pdf/assessmentreport/ar5/syr/AR5_SYR_FINAL_SPM.pdf. Jansohn, P., Griffin, T., Mantzaras, I., Marechal, F., Clemens, F., 2011. Technologies for gas turbine power generation with CO2 mitigation. Energy Procedia 4, 1901-1908. Jonsson, M., Yan, J., 2005. Humidified gas turbines - A review of proposed and implemented cycles. Energy 30, 1013-1078. Li, H., Ditaranto, M., Berstad, D., 2011a. Technologies for increasing CO2 concentration in exhaust gas from natural gas-fired power production with post-combustion, amine-based CO2 capture. Energy 36, 11241133. Li, H., Haugen, G., Ditaranto, M., Berstad, D., Jordal, K., 2011b. Impacts of exhaust gas recirculation (EGR) on the natural gas combined cycle integrated with chemical absorption CO2 capture technology. Energy Procedia 4, 1411-1418. Mansouri Majoumerd, M., Nikpey Somehsaraei, H., Assadi, M., Breuhaus, P., 2014. Micro gas turbine configurations with carbon capture Performance assessment using a validated thermodynamic model. Applied Thermal Engineering 73, 172-184. Merkel, T.C., Wei, X., He, Z., White, L.S., Wijmans, J.G., Baker, R.W., 2013. Selective exhaust gas recycle with membranes for CO2 capture from natural gas combined cycle power plants. Industrial and Engineering Chemistry Research 52, 1150-1159. Nikpey Somehsaraei, H., Assadi, M., Breuhaus, P., Mørkved, P.T., 2014. Experimental evaluation and ANN modeling of a recuperative micro gas turbine burning mixtures of natural gas and biogas. Applied Energy 117, 30-41. NRG Energy Inc., 2017. Petra Nova. nrg.com/generation/projects/petra-nova/. Rao, A.D., Day, W.H., 1996. Mitigation of greenhouse gases from gas turbine power plants. Energy Conversion and Management 37, 909-914. Røkke, P.E., Hustad, J.E., 2005. Exhaust gas recirculation in gas turbines for reduction of CO 2 emissions: Combustion testing with focus on stability and emissions. International Journal of Thermodynamics 8, 167173. S P S saskpowerccs.com. P CCS T W F 17 P -Combustion Coal-Fired CCS Facility. Taimoor, A.A., Muhammad, A., Saleem, W., Zain-ul-Abdein, M., 2016. Humidified exhaust recirculation for efficient combined cycle gas turbines. Energy 106, 356-366. Takahashi, T., Koda, E., Mimaki, T., 2002. A systematic analysis of the effect of air humidification to gas turbine systems. Japan Society of Mechanical Engineers: International Journal Series B 45, 530-535. Turbec, 2009. T100 Microturbine System: Technical Description 09/12/29. T100 Natural Gas. D 14127-03 Version 3 Wan, K., Zhang, S., Wang, J., Xiao, Y., 2010. Performance of humid air turbine with exhaust gas expanded to below ambient pressure based on microturbine. Energy Conversion and Management 51, 2127-2133. Wei, C., Zang, S., 2013. Experimental investigation on the off-design performance of a small-sized humid air turbine cycle. Applied Thermal Engineering 51, 166-176 18