Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts

Applied Energy xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts Ward De Paepe a,c,⇑, Marina Montero Carrero b,a,c, Svend Bram b,c, Francesco Contino b,c, Alessandro Parente a,c a b c

Université Libre de Bruxelles (ULB), Aero-Thermo-Mechanical Department (ATM), Avenue Franklin Roosevelt 50, 1050 Brussels, Belgium Vrije Universiteit Brussel (VUB), Thermo and Fluid dynamics (FLOW), Pleinlaan 2, 1050 Brussels, Belgium BURN Joint Research Group, Université Libre de Bruxelles & Vrije Universiteit Brussel, Belgium

h i g h l i g h t s
Water injection in micro gas turbine (mGT) cycles has a large potential for waste heat recovery.
Current humidified mGT cycles do not fully exploit the potential for waste heat recovery.
Different cycle concepts of large scale gas turbines are applied on the small scale mGT to find the optimal cycle.
The cycle concept allowing the highest water injection rate and lowest stack temperature recovers most waste heat.
The REVAP cycle concept with preheat was identified as the optimal cycle.

a r t i c l e

i n f o

Article history: Received 20 January 2017 Received in revised form 26 May 2017 Accepted 1 June 2017 Available online xxxx Keywords: micro Gas Turbine (mGT) Cycle humidification Waste heat recovery micro Humid Air Turbine (mHAT) Advanced Humid Air Turbine (AHAT) REgenerative EVAPoration (REVAP)

a b s t r a c t Introduction of water in a micro Gas Turbine (mGT) has proven to be a very effective method to recover waste heat into the cycle, since it increases the mGT electrical efficiency significantly. Different routes exist for water introduction in the mGT cycle. Classical routes, like injection of steam/preheated water or the micro Humid Air Turbine (mHAT) concept, where water is introduced in the cycle by means of a saturation tower, have shown to have high potential. However none of the previously mentioned cycles exploits the full thermodynamic potential for waste heat recovery through water introduction. More advanced humidified Gas Turbine (GT) cycles have been proposed and studied for large scale GTs. So far, none of these concepts have been applied on mGT scale, despite their high potential. In this paper, we study the impact of these different, more advanced, humidified GT cycle concepts on the mGT performance. The different selected cycles – next to the classical steam injection or injection of (preheated) liquid water in the recuperated cycle and the mHAT – were: micro Humid Air Turbine Plus (mHAT+), Advanced Humid Air Turbine (AHAT) and the REgenerative EVAPoration (REVAPÒ) cycle concept. The impact of these concepts on the mGT cycle performance has been studied on the Turbec T100 mGT. Simulations indicated that humidifying the air of the mGT has a significant beneficial effect on cycle performance due to the increased waste heat recovery, resulting in a higher electrical power output (at constant rotational speed) or reduced fuel consumption (at constant power output), both leading to an increased electrical efficiency. Depending on the different cycle layout used, more or less waste heat could be recovered from the exhaust gas. The REVAPÒ concept with feedwater preheat was identified as the optimal cycle layout within the selected options. By applying this concept to the Turbec T100, most waste heat could be recovered, achieving the highest electrical efficiency increase. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction ⇑ Corresponding author at: Université Libre de Bruxelles (ULB), Aero-ThermoMechanical Department (ATM), Avenue Franklin Roosevelt 50, 1050 Brussels, Belgium. E-mail address: [email protected] (W. De Paepe). URL: http://www.burn.ulb.ac.be (W. De Paepe).

Despite the potential of micro Gas Turbines (mGTs) for smallscale (up to 500 kWe) Combined Heat and Power (CHP) production [1,2], they never fully penetrated the mGT market [3]. The heat-

http://dx.doi.org/10.1016/j.apenergy.2017.06.001 0306-2619/Ó 2017 Elsevier Ltd. All rights reserved.

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

2

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

Nomenclature Acronyms AHAT Advanced Humid Air Turbine CAF Corrected Air Flow CHAT Cascaded Humidified Advanced Turbine CHP Combined Heat and Power GT Gas Turbine HAT Humid Air Turbine HRSG Heat Recovery Steam Generator ICE Internal Combustion Engine mGT micro Gas Turbine mHAT micro Humid Air Turbine mHAT+ micro Humid Air Turbine Plus REVAPÒ REgenerative EVAPoration STIG STeam Injected Gas Turbine PIT Turbine Inlet Pressure TIT Turbine Inlet Temperature TOPHATÒ TOP Humid Air Turbine TOT Turbine Outlet Temperature WAC Water Atomizing inlet air Cooling

driven operation of CHP units in general, in combination with the lower electrical efficiency of the mGT compared to its main competitor, the Internal Combustion Engine (ICE) [4], was identified in the past as their major drawback. Since the mGT has still a relative high specific investment cost [5], any forced shutdown of the installation during periods with limited or no heat demand, has a negative effect on the economic performance [6]. By introducing water in the mGT cycle during periods with low heat demand to increase the electrical efficiency, the economic performance can be improved significantly [7–11]. In past research, several routes for mGT cycle humidification for waste heat recovery have been proposed and studied. Many researchers focussed on classical steam injection [6,10–13] or water injection [12,14] by studying these different options on different mGTs. All researchers reported a significant electrical efficiency increase (and possible electrical power output increase, when operated at constant rotational speed). This increase depends on the mGT size, the injection method (liquid water or steam) and on the injection point in the cycle. Finally, the concept of steam injection in the mGT cycle has already been validated experimentally for the Capstone C60 [13] and for the Turbec T100 [15–18], showing the potential for mGT humidification. Next to the more classical humidification methods of steam and liquid water injection, Parente et al. proposed the micro Humid Air Turbine (mHAT) concept [19]. This mHAT concept is based on the Humid Air Turbine (HAT) cycle, proposed by Rao [20], but modified for the mGT cycle by excluding the inter- and aftercooling. This mHAT cycle was identified, by our research group, as a perfect candidate for waste heat recovery through humidification, based upon its high efficiency in combination with the rather limited necessary cycle modifications [21]. These findings were confirmed by Nikpey et al. [22] and Majoumerd et al. [23], who also simulated the conversion of the Turbec T100 mGT into a mHAT. The mGT can easily be converted into a mHAT by introducing a saturation tower in between compressor outlet and recuperator inlet [24]. The high potential of the mHAT cycle was confirmed experimentally by our research group by transforming a Turbec T100 into a mHAT by introducing a spray saturation tower in the cycle [25–27]. Exergy analysis however indicated that the thermodynamic limit for water addition in the mGT is much higher than what can

Roman symbols A cross section area k heat capacity ratio _ m mass flow rate R universal gas constant Greek symbols g efficiency p pressure ratio Subscripts is isentropic turb turbine Superscripts ⁄ properties of standard air

be achieved with the steam/water injection in mGTs and the mHAT cycle [28]. As a consequence, the exergy loss at the stack is significant, even for the mHAT cycle (19 kW exergy loss at nominal power output of 100 kWe), as it can be clearly observed from the analysis of the exergy flows in a Grassmann diagram [29]. The main problem with the existing options for humidification, is the limited heat recovery from the stack, especially the recovery of the water latent heat in the flue gases. The latter is only released at low temperatures (below 67 °C), thus requiring a cooling stream ensuring a sufficient temperature difference (at least 10 °C lower with respect to the minimum temperature difference). In addition, such a stream should allow for the absorption of the heat, which requires that the stream remains at this temperature during the heat transfer process. This is only possible if the cold stream has a mass flow rate that is several orders of magnitude larger than the flue gas mass flow rate or if it is combined with a phase-change process. None of the above-mentioned cycles displays these options. In this paper, the potential of several more advanced humidified Gas Turbine (GT) cycle concepts on mGT scale will be presented. The aim of the study presented in this paper is to identify which of these more advanced humidified GT concepts has the largest potential for application on mGT scale for waste heat recovery and approaches the thermodynamic limit, which is defined based upon first and second law analysis. Based on the overview work of Jonsson and Yan, which discusses the different humidified GT cycles [30], a selection of possible large scale humidified GT concepts that can be applied on mGT scale, is made. The application of these cycles on mGT level is simulated on a typical mGT, the Turbec T100, by using Aspen PlusÒ [31]. This analysis allows for a full comparison of the performance of the different cycle concepts on mGT scale and identification of the optimal solution (from a thermodynamic point of view). The final feasibility study, taking into account investment and operating costs and possible technological challenges (e.g. material constraints) [32], is not within the scope of this paper. In the following sections, first the different selected cycles are presented. Second, the simulation approach for the humidified cycles is discussed, followed by the presentation of the results of the different simulations. Finally, the main findings of this study are summarized in the conclusion.

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

2. Advanced humidified mGT cycle selection The Turbec T100 mGT is selected as reference mGT to apply the advanced humidified cycle concepts. The nominal specifications of this unit, as indicated by the manufacturer, are summarized in Table 1. This Turbec T100 is a typical mGT, using the recuperated Brayton cycle (Fig. 1). The air is first compressed in a radial, variable speed compressor (1). By preheating the compressed air in a recuperator using the hot exhaust gases (2), high efficiency can be achieved. The compressed air is heated further in the combustion chamber by burning fuel (3). The hot compressed air is then expanded over the turbine (4) to deliver the necessary power to drive the compressor. The remaining power on the shaft in converted into electrical power by a variable speed generator (5). The Turbec T100 is, like most mGTs, a small-scale CHP unit, meaning both electricity and heat are produced simultaneously. By doing so, a very high total efficiency can be reached (80% for the T100, with 30% electrical and 50% thermal efficiency, as indicated by the manufacturer (Table 1)). For the Turbec T100 mGT, the thermal power is produced in the economizer ((6) in Fig. 1). In this economizer, the remaining heat in the exhaust gases coming from the recuperator, is converted into thermal power by heating water or producing steam for e.g. heating purposes. However, as mentioned before, when there is no heat demand, this heat in the exhaust gases cannot be used and is rejected as waste heat through the stack. This results in a total exergy loss of 69 kW through the stack, compared to a 7 kW loss in CHP-mode (when heat is recovered in the economizer) [29]. The main issue with the remaining waste heat in the exhaust gases is the low temperature at which it is available. The temper-

Table 1 Nominal specifications of the Turbec T100 mGT [33]. Electric power Thermal power Electric efficiency Thermal efficiency Total efficiency Maximal shaft speed

100 kWe 167 kWth 30% 50% 80% 70,000 rpm

Fig. 1. The Turbec T100 mGT, which is a typical engine representative for the stateof-the-art of mGTs, was used as reference machine for the development of the advanced humidified mGT cycles. The T100 mGT is a recuperated Brayton cycle, meaning that the compressed air coming from the compressor (1) is preheated in a recuperator using the heat available in the exhaust gasses (2), before sending the air to the combustor (3) and turbine (4) connected to a generator (5) for electrical power production. Thermal power for heating purpose is produced in the economizer (6).

3

ature of the gas leaving the recuperator is too low for direct waste heat recovery in a heat exchanger. Therefore, to be able to use this waste heat, an additional cold source, in this case water, should be added to the cycle, enabling the recovery. Different possible routes exist to recover waste heat in an mGT cycle using water. Depending on the cycle, the existing heat exchange network of the mGT – consisting in a recuperator and an economizer (components (2) and (6) in Fig. 1) – is modified by introducing extra components, like a water injection system (a simple injector or a saturation tower) and possibly additional heat exchangers (aftercoolers or extra economizers). The final cycle layout depends on the chosen water introduction method. The aim of this study is to maximize the heat exchange in this network, using water to re-introduce the waste heat from the stack. This leads to a high heat recovery in the network, resulting in a higher overall electrical efficiency. Rather than using simple trial and error to find the optimal advanced humidified mGT cycle for waste heat recovery, a more systematic approach is used. The optimal cycle layout can be found using a 2-step method [34], where in a first step, the thermodynamic limit is found using a black box method in combination with second law analysis. In this first step, the heat exchanger network is replaced by a black box system on which first and second law analysis is applied. The thermodynamic optimal performance is then found by imposing a maximal exergetic efficiency (93%) and a minimal exergy destruction (5%) to the network. In a second step, the layout of the cycle corresponding to this optimal is searched by specifying the heat exchanger network. The feasibility of each of the networks is analysed using composite curve theory and pinch analysis. The cycle layout achieving similar performance as the thermodynamic optimal, is then the desired layout [34]. In a previous study conducted by our research group, the potential for mGT humidification was identified and, depending on the boundary conditions, the injection of preheated water injection or the mHAT were found to be the optimal solutions [28]. When assuming the heat exchangers are redesigned/optimized for the humidified cycle, preheated water injection is the option with the highest performance, while when using the original components of the dry mGT, the mHAT improves the performance the most [28]. However, neither of these cycles did reach the maximal possible efficiency identified from the black box analysis, applying first and second law analysis. The full potential could not be reached, since in this previous work, the number of heat exchangers was limited to 2. This limitation was set to keep the investment cost of the installation low [28], which puts some constraints and limits on the recovery of the waste heat from the exhaust gases. However, in the study presented in this paper, the number of possible heat exchangers in the network is not limited, which should allow for more heat recovery from the stack, especially the evaporation heat that is only available at low temperature. Both the options of aftercooling and a second economizer to preheat the feedwater are possible in this analysis, however intercooling is still not possible due to the single pressure level in the mGT. This should lead to higher heat recovery, a lower exergy destruction and higher exergetic efficiency of the network and a final cycle layout approaching the exergetic limit found in previous studies [21,28]. To avoid a random completion of the heat exchanger network for the mGT, different advanced humidified cycles for large-scale GTs, summarized by Jonsson and Yan [30], are converted to the smaller mGT scale. The following candidates are selected: GTs with injection of water that evaporates completely: Next to the more traditional (preheated) water injection in the recuperated cycle (injection in the compressor outlet, Fig. 2(b)), 2

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

4

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

Fig. 2. The advanced humidified mGT cycles, studied in this paper, covering: (a) inlet air cooling using WAC or inlet air fogging; (b) direct injection of water, possibly preheated (Economizer); (c) the REVAPÒ cycle which combines liquid water injection with aftercooling (Aftercooler), with possible feedwater preheat (Economizer 2); (d) injection of steam (STIG-case), both in the compressor outlet or in the combustion chamber and finally cycles using a saturation tower for humidification: (e) the mHAT (no aftercooling) or (f) mHAT+ (with aftercooling), both possibly with feedwater preheating (Economizer 2). The newly introduced part necessary for the humidification are indicated in pink. The AHAT cycle (not shown) combines the layout of the mHAT (e) with inlet air cooling, using WAC (a). (For interpretation of the references to colours in this figure legend, the reader is referred to the web version of this paper.)

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

other cycle concepts are selected as candidates for cycle analysis: compressor inlet air cooling, using fogging of Water Atomizing inlet air Cooling (WAC) (Fig. 2(a)), and the more advanced REgenerative EVAPoration (REVAPÒ) cycle [35], where part of the air/liquid water mixture after the injection is used for compressor aftercooling (Fig. 2(c)). GTs with injection of steam: In this category, only the traditional STeam Injected Gas Turbine (STIG) is selected, where the steam is either injected in between compressor outlet and recuperator inlet or in the combustion chamber (Fig. 2(d)). GTs cycles with injection of water in a humidification tower: For the final category, next to the classical mHAT cycle (where a saturation tower is introduced for humidification in between compressor outlet and recuperator inlet, Fig. 2(e)), the more advanced micro Humid Air Turbine Plus (mHAT+) (using aftercooling to preheat the water before injection in the saturation tower, Fig. 2(f)) [19] and the Advanced Humid Air Turbine (AHAT) (combining the mHAT cycle concept with compressor inlet air cooling, using WAC) [36] are selected. Important criteria for selection are: (1) the water is introduced to recover waste heat from the flue gases, e.g. the option of injecting liquid water in the combustion chamber to reduce the NOx exhaust is not considered in this analysis; (2) water can be introduced before or behind the compressor, however when injecting before the compressor, like when performing WAC or inlet air fogging, all water must be evaporated to avoid any damage to the component (wet compression, where droplets of liquid water enter the compressor, is thus not taken into account) and (3) the humidified GT concept must be applicable to the mGT cycle, meaning applicable at a single pressure level, since the mGT has only one compressor and one turbine stage. Due to this final constraints, cycle layouts, using intercooling and more advanced HAT cycles, like the Cascaded Humidified Advanced Turbine (CHAT) – cycle developed to compensate the mass unbalance between compressor and turbine [37] – or the TOP Humid Air Turbine (TOPHATÒ) [38] cannot be applied. Finally, partial humidification is not considered as an option, since the main goal is to optimize waste heat recovery. 3. Advanced humidified cycle modelling The modelling of the different advanced humidified mGT cycles, presented in the previous section (Fig. 2), is performed in Aspen PlusÒ [31], using modified dry and wet simulation models, created and validated in the past [15,16]. For the modelling of the turbo machinery parts of the mGT, the original compressor and turbine of the Turbec T100 are considered and are assumed to remain unmodified. For the compressor, the actual compressor map of the T100 (similar to the map presented in [39]) is introduced in Aspen PlusÒ (including the surge limit) indicating pressure ratio and isentropic efficiency. The turbine is assumed to be choked and having a dry isentropic efficiency of 85%. Since the introduction of water has a significant effect on the properties of the working fluid going through the turbine, the choking condition is corrected for the water injection by adapting the choking constant of the turbine according to the changing turbine inlet air composition:

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u pffiffiffiffiffiffiffi  kturb þ1 u kturb 1 _ turb TIT 2 m tkturb ¼ cte ½40: ¼A PIT kturb þ 1 R

ð1Þ

Identically, the isentropic efficiency is also corrected for the humidified cycles, using the correction method proposed by Parente et al.:

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi gis k  1 k þ 1 1  1=pc ½19: ¼ gis k  1 k þ 1 1  1=pc

5

ð2Þ

In Eq. (2), the superscript ( ) refers to the gas properties at standard dry air composition. The combustion chamber is also assumed to be unmodified and is modelled using a classical Gibbs reactor, with a combustion efficiency of 100% (complete combustion) and 10 kWth heat loss. Possible combustion instabilities are not considered in this paper, since the water fraction in the combustion air, even at the highest injection rate (10% in the preheated REVAPÒ case), is still below the limit for premixed combustion at which CO levels become too high (above 30% water content [41] for stable and complete combustion). Pure methane is considered as fuel, to make comparison with the exergy analysis from previous work possible [21,28], however similar results could be expected when using other gaseous fuels, like natural gas. In this combustion chamber, a 5% pressure loss is introduced (similar to the loss in the actual combustion chamber). Finally, the heat exchange and water injection network was simulated using generic counter flow heat exchangers. As design specification for these heat exchangers, a minimal pinch temperature of 50 °C for gas/gas heat exchangers (recuperator), 10 °C for gas/liquid or gas/gas-liquid heat exchangers (economizers and aftercoolers) and finally 10 °C for the entire heat exchange network are assumed. Additionally, a typical design pressure loss of 3% on the cold side of the network and 40 mbar at the hot side is considered [34]. Water is assumed to be introduced in the network at 15 °C and 1.013 bar. Finally, for the water introduction in the mGT working fluid, an adiabatic mixer is used for liquid water or steam injection, introducing a 0.5% pressure loss, which is a typical pressure loss related to the injectors. For the cycles where the working fluid is humidified using a saturation tower, the tower is simulated using the RadFrac module, as proposed by Queiroz et al. [42], also introducing a 0.5% pressure loss (typical pressure loss over the packing material [24]). The details about the used parameters in Aspen PlusÒ are summarized in Table 2. In the simulation models, 3 control loops are implemented:
Turbine Inlet Temperature (TIT) control;
speed control;
water injection control. The first control loop adjusts the fuel flow rate going into the combustion chamber. To optimize the electrical efficiency of the mGT, the Turbine Outlet Temperature (TOT) is typically kept constant (for the Turbec T100, TOT is kept constant at 645 °C) by changing the fuel flow rate. This TOT corresponds to a TIT close to 950 °C, which is the limit for the material of the turbine. Rather than keeping TOT constant, TIT was kept constant in this study. When water is introduced in the cycle, the corresponding TIT to a TOT of 645 °C reduces as a result of the changing properties of the working fluid. This lower TIT has a negative effect on the performance of the cycle. Therefore, TIT was kept constant, which leads to a TOT increase, which will however require special attention for the recuperator material selection [32]. The second control loop will regulate the rotational speed. The simulations are performed using two different operation modes of the mGT: constant rotational speed and constant electrical power output. The first simulation results on the advanced humidified mGT cycles presented in this paper are performed at constant rotational speed of the compressor and turbine. The speed is kept constant at 67230 rpm (see Table 2), which is the speed at which the mGT delivers the nominal electrical power output of 100 kWe in dry operation, at the same inlet conditions of 15 °C and 1.013 bar. By introducing water while keeping the rotational speed

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

6

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

Table 2 Boundary conditions used for the advanced humidified mGT cycle development. Compressor Pressure ratio Isentropic efficiency Inlet air temperature Rotational Speed Mechanical efficiency

Variablea Variablea 15 °C Variable/ 67320 rpmb 99%

Turbine Turbine back pressure Isentropic efficiency Turbine Inlet Temperature (TIT) Mechanical efficiency

40 mbar Variablec 950 °C 99%

Combustion chamber Combustor pressure loss Combustor heat loss

5% 10 kWth

Heat exchanger network Cold side pressure loss Hot side pressure loss Water injection pressure loss Heat exchanger pinch (gas/gas) Heat exchanger pinch (gas/liquid or gas/gas + liquid) Overall heat exchanger network pinch Stack temperature Feedwater inlet temperature

3% 40 mbar 0.5% 50 °C 10 °C 10 °C Variabled 15 °C

Fuel (methane) Lower Heating Value

50 MJ/kg

General Produced electrical power Combined generator and power electronics efficiency

100 kWe/Variablee 94%

a

The compressor map of the Turbec T100 was used in the simulations [39]. Depending on the operation mode, constant power output or constant rotational speed, the rotational speed is variable to control power or kept constant by the control system. c The isentropic efficiency depends on the water content of the working fluid (Eq. (2)). d The final stack temperature depends on the amount of injected water and recovered quantity of waste heat. e Depending on the operation mode, constant power output or constant rotational speed, the power output is kept constant at the nominal output or is variable. b

constant, the power output will increase due to the additional mass flow rate through the turbine. The second set of simulations is performed using constant power output mode, which corresponds to the standard operation mode of the mGT. In this constant power output mode, the Turbec T100 mGT operates at constant power output by adjusting the rotational speed of the shaft. This means that, when water/steam is introduced in the cycle behind the compressor, more power will be available on the shaft due to the mass imbalance between compressor and turbine [30], which leads to a power output increase. Therefore, the rotational speed will be lowered by the control system to keep the power output constant [15]. Because the turbine is choked, the operating point of the mGT shifts towards the surge limit (Fig. 3). Since an additional water mass flow rate is added to the cycle behind the compressor, the air flow rate through the compressor is limited (by the choking condition), reducing the Corrected Air Flow (CAF) and thus also the surge margin (Fig. 3). The final control loop sets the amount of introduced water into the cycle. The control system increases the feedwater flow rate till a maximum is reach. This maximal amount of water corresponds to the point were one component of the heat exchanger network has reached its minimal pinch temperature, but all the other boundary conditions from Table 2 are still respected. For both the mHAT+ and the REVAPÒ cycle concept, an additional control loop for setting the split fraction is added to the system. For the mHAT+, the split fraction is set in such a way, that the

Fig. 3. Due to the choking of the turbine, the operating point of the mGT shifts towards the surge limit with increasing water injection.

temperature differences between both streams on both sides of Economizer 1 (Fig. 2(f)) are equal to 10 °C. For the REVAPÒ cycle, the working fluid is split between Economizer 1 and the After Cooler (Fig. 2(c)) in such a way that the temperatures of the cold streams at the outlet of both heat exchangers are equal. The final simulation results of the advanced humidified mGT cycles are compared to the dry cycle performance, i.e. electrical power output and efficiency. This dry model has been validated before [6]. Reference values for the dry mGT cycle are: the nominal electrical power output of 100 kWe is produced with a rotational speed of 67320 rpm and an electrical efficiency of 32.8%. This simulated electrical efficiency of the dry mGT is higher than the nominal efficiency of 30% reported by Turbec (Table 1) due to the use of the same design and control conditions as used for the advanced humidified mGT cycles of Table 2 (which allows for a correct comparison). This means that the recuperator has a slightly higher effectiveness due to the use of a fixed hot side temperature difference rather than using a predefined heat exchanger model (50 °C hot side temperature difference compared to 70 °C when using the actual heat exchanger [15]), while the TIT is also slightly increased (950 °C compared to 920 °C at nominal load [15]), due to the TIT control instead of the TOT control. Next to the electrical performance of the different humidified mGT cycle concepts, the waste heat recovery is assessed by focussing on two parameters: the amount of injected water and the temperature of the flue gases. Previous process-based simulations from our group proved that the amount of injected water needed to be maximized to boost the efficiency of the recovery system and thus the waste heat recovery, leading to a stack temperature minimization [28]. The total exergy flow of the exhaust gas is a possible alternative parameter that can be used to determine the waste heat recovery. However, in the case of the mGT, there is a oneto-one relationship between the temperature of the flue gases and the exergy flow. The higher the water injection rate, the lower the stack temperature will become. At the same time, this higher water injection rate also leads to lower total mass flow rates (since the turbine is choked). In combination with the lower exhaust gas temperature, this results in a lower exergy flow rate. 4. Results Analysis of the simulation results of the different proposed cycles clearly indicates that the cycle layout allowing the lowest stack temperature, thus allowing for the highest waste heat recovery from the stack, achieves the highest cycle efficiency (Figs. 4 and 5). For both the simulations at constant rotational speed (Fig. 4) and constant power production (Fig. 5), the REVAPÒ cycle concept

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

7

Fig. 4. The cycle allowing the lowest stack temperature (here the REVAPÒ cycle with feedwater preheat), allows for the highest amount of injected water and waste heat recovery, resulting in the highest cycle efficiency and electrical power output at constant rotational speed.

with feedwater preheat has the highest efficiency increase (4.6 and 4.2%points respectively), followed by the REVAPÒ without feedwater preheat (4.6 and 4.2%points) and direct water injection with feedwater preheat (3.8 and 3.6%points), although the difference in performance between these concepts is limited. So despite the much more complex cycle layout of the REVAPÒ, the efficiency is not much higher than simple direct preheated water injection. Both mHAT (3.8 and 3.6%points) and mHAT+ (3.8 and 3.6%points) achieve high efficiency (slightly higher for the mHAT+ due to the positive effect of the aftercooling on the evaporation process in the saturation tower), but they do not reach the high injection levels and performance of the direct injection humidification. Finally, the STIG concept with injection in the combustion chamber achieves the lowest cycle efficiency increase (even results in an efficiency decrease of 0.3 and 0.7%points), followed by the STIG with injection in the compressor outlet (2.5 and 2.4%points efficiency increase). This lower performance is a result of the high exergy destruction in the Heat Recovery Steam Generator (HRSG) (see later). This shows that direct injection of water for mGT humidification (method I for GT cycle humidification as defined by Jonsson and Yan [30]) is superior over – meaning: achieving higher waste heat recovery and thus higher electrical efficiency – the humidification using a saturation tower (method III), which in turn is superior over the humidification through steam injection (method II). Using the REVAPÒ cycle concept with feedwater preheat, the final stack temperature could be lowered the most (as low as 55.3 °C at constant rotation speed and 53.9 °C at constant power

output), allowing the highest amount of injected water (69.5 and 58.0 g/s respectively) and the highest electrical efficiency increase of 4.6 and 4.2%points of all cycles. This maximal efficiency increase and waste heat recovery of the REVAPÒ cycle is however still below the exergetic limit found during the black box analysis [28]. This limit was a water injection of 155 g/s corresponding to an electrical power production of 147.4 kWe and an electrical efficiency of 42.3% at constant rotational speed of 67320 rpm and 123 g/s water injection with an electrical efficiency of 41.7% at constant power production of 100 kWe respectively. Both limits are set by the compressor, which reaches its surge limit due to the choking condition (Fig. 3) and not because of a violation of the second law. For both cases, the exergetic efficiency of the heat recovery network is below the limit of 93% and the exergy destruction is above the minimal 5% [34]. This indicates that from a thermodynamic point of view, even more water can be injected, but this would require a redesign of the compressor. Despite this potential, we see that for large scale GTs the REVAPÒ and HAT cycles fulfil the full potential for water introduction [43], while for the mGT, the REVAPÒ cycle still does not reach the thermodynamic limit. From this observation, 2 possible conclusions can be drawn. A first possible conclusion is that from a thermodynamic point of view it should be possible to design a heat exchanger network that allows for this optimal waste heat recovery. The main challenge lies within the condensation heat of the water, which must also be recovered to reach an exergetic heat recovery efficiency of 93%. Already in the mGT with REVAPÒ concept, the final stack temperature is below the dewpoint, meaning part of the condensation heat of the water

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

8

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

Fig. 5. Similar to the simulations at constant rotational speed, for constant power output, it is also the cycle allowing the lowest stack temperature (here also the REVAPÒ cycle with feedwater preheat), that allows for the maximal amount of water injection and highest waste heat recovery, resulting in the highest cycle efficiency.

in the stack is recovered. To reach the black box limit, an even larger fraction of this condensation heat must be captured [28]. The challenge herein lies in the recovery of a large fraction of heat that is only available at low temperature (below 55 °C), which is not straightforward. As mentioned in the introduction, this is only possible if the cold stream has a mass flow rate that is several orders of magnitude larger than the flue gas mass flow rate or if it is combined with a phase-change process. The REVAPÒ cycle only partially fulfils this. A second possible conclusion is that the limits set for the black box analysis (93% exergetic efficiency and 5% exergy destruction) are too high for mGTs. Due to the lower maximum allowed TIT of 950 °C, mGTs operate with a much larger airto-fuel ratio compared to large scale GTs. This higher air-to-fuel ratio makes that the energy flow circulating in the heat exchanger network of the humidified cycles is larger than the energy input through the fuel in the combustion chamber (even in the dry cycle, both flows are of the same order of magnitude [29]), this in comparison to large GTs operating at higher TIT and lower air-to-fuel ratio. Therefore, it is possible that the limits must be adjusted for the mGT scale. The final result of the REVAPÒ cycle with preheat corresponds to an exergy efficiency of 84.0% and 85.1% and a destruction of 13.9% and 14.4% for respectively constant rotational speed and constant power output. For all cycles with water injection (with or without saturation tower), it is clear that the effect of feedwater preheat is rather limited on the final cycle performance. Feedwater preheat results in a 0.01, 0.2 and 0.3%points extra efficiency increase for respectively the REVAPÒ concept, direct water injection and the mHAT cycle

(Fig. 4). This limited effect can be explained by the small mass flow rates of feedwater, that only allow for a limited amount of waste heat recovery. Nonetheless, there is still a minor positive effect, but when constructing the final cycle layout, one should consider the additional cost of an extra heat exchanger for the preheat over the rather limited gain in efficiency of the cycle (this is a consideration that should be made for all cycles, but is outside the scope of this paper). The results of the simulations indicate that direct injection of water which fully evaporates (direct water injection and REVAPÒ concept) is superior over cycle humidification using a saturation tower (mHAT and mHAT+ concept) (Figs. 4 and 5). By introducing liquid water directly in the cycle, the hot compressed air is cooled further due to the immediate evaporation of the water in the dry air, strongly reducing the temperature. In addition, after injecting the water mass flow rates indicated in Figs. 4 and 5, the compressed air is not only fully saturated, but part of the water is still present in its liquid form (2-phase flow mixture). This liquid water allows for a higher heat recovery due to the evaporation heat, causing a slower rise of the temperature of the 2-phase flow, leading to a higher cycle performance. To achieve this, one should accept 2phase flow in the cycle. This 2-phase flow will make the design of the heat exchangers more complex. The behaviour of the 2phase flow is also more complex to predict and model. The presence of 2-phase flow could lead to an imbalance in the heat exchangers. In this component, the flow is split, passing through different channels. If different liquid fractions enter different channels, every channel may have a different temperature, therefore

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

9

Fig. 6. The composite curves of the different cycles show the transfer of the heat between the cold and hot streams in the heat exchanger network and the corresponding temperature change.

leading to the imbalance. Depending on the gas/liquid fraction, several flow regimes exist and depending on these regimes, flow characteristics and heat transfer are different [44]. When the water injection rate is limited, e.g. in the water injection case, to avoid 2phase flow (injection of liquid water up to full saturation of the compressed air, but not further), the amount of injected water is limited to 34.6 and 30.6 g/s (constant speed and constant power with no feedwater preheat), leading to a limited efficiency increase of 2.7 and 2.6%points, which is lower than the increase achieved

with the mHAT and mHAT+. So in case 2-phase flow is allowed, direct injection of water that fully evaporates (direct water injection and REVAPÒ cases) are superior over cycle concepts using a saturation tower for humidification. When 2-phase flow is not allowed, cycle concepts using these saturation tower for humidification have the highest potential. Composing the different composite curves of the different heat exchanger networks provides insight to better understand the different cycle performances (Fig. 6). These composite curves show

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

10

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

Fig. 6 (continued)

the heat exchange between the hot and cold streams at each temperature. They indicate the heat exchanged between the hot and cold stream and the corresponding temperature change of both streams during the heat transfer process. The total energy transferred by the hot stream and absorbed by the cold stream is then the width of the x-axis. For optimal performance, the cold and hot curves must be as close as possible to each other, preferably a distance equal to the pinch temperature of 10 °C, over the full range. It is clear from the different composite curves depicted in Fig. 6 (curves are obtained at constant rotational speed, but similar results can be expected for the constant power simulations) that the REVAPÒ cycles achieve the highest heat recovery due to the very low cold curve starting temperature and the small distance between the hot and cold curves. These composite curves also explain why steam injection (STIG-case) is always inferior to water injection (with or without saturation tower). When the feedwater reaches the boiling temperature, the temperature will remain constant during the steam generation in the HRSG (which depends on the water pressure), until all water has turned into steam. During this boiling process, the cold stream keeps absorbing heat from the hot stream to turn the liquid water into vapour, cooling down this hot stream, while the temperature of the cold stream remains constant. To avoid composite curve crossing, there is no other possibility than to accept a higher temperature difference between the hot and cold stream, as a result of the isothermal phase change process. The larger temperature difference results in more exergy destruction in the heat exchange network, leading to a lower cycle

performance. The cycles with water injection are not bound by this limitation, even the cycles with a saturation tower, since evaporation happens at variable temperature, allowing smaller temperature differences, higher heat exchange and cycle performances. Finally, the effect of performing inlet air cooling, using water atomization, depends on the conditions of the inlet air. When assuming 0% relative humidity, the mGT efficiency increases from 32.9% to 34.2% (WAC-case) at constant rotational speed and from 32.9% to 34.4% at constant power output, due to a significant drop in inlet air temperature (by 11.4 °C). When assuming a more common 60% relative humidity, this effect is limited, resulting in a final electrical efficiency of 33.3% and 33.5% at, respectively, constant rotational speed and constant power output. When applying inlet air cooling using water evaporation on the mHAT (the AHAT), assuming a relative humidity of 60% of the inlet air, the efficiency of the cycle can be improved to 36.5% (constant rotational speed) and 36.4% (constant power output), compared to the 36.3% and 36.1% efficiency of the mHAT. The inlet air humidification has little effect on the humidification of the compressed air at the compressor outlet in the saturation tower, since the added water fraction in front of the compressor is much smaller than the fraction added in the saturation tower (less than 3%). As a final remark, one can conclude that inlet air cooling using fogging or WAC can improve the performance of all cycles, but the effectiveness depends mainly on the inlet air conditions. Furthermore, the inlet air cooling does not allow for a higher waste heat recovery.

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

5. Conclusion In this paper, several advanced humidified GT concepts of large scale GTs have been applied to mGT scale to study their potential for waste heat recovery in the mGTs. The results of this analysis clearly showed that the cycle which is capable of reducing the stack temperature the most, has the highest waste heat recovery from the exhaust gases, resulting in the highest electrical efficiency. The key for optimal waste heat recovery lies in the possibility to recover the evaporation heat from the condensing water in the stack. However, to recover this additional waste heat, a more complex cycle layout is necessary. The optimal heat recovery can be accomplished using the REVAPÒ cycle concept. Within this concept, the stack temperature can be lowered the most (up to 55.3 °C and 55.9 °C) for both operation at constant rotational speed and constant power output. Applying the REVAPÒ concept with feedwater preheat to the Turbec T100 results in a final electrical power output of 121.6 kWe and an electrical efficiency of 37.4% (increase of 4.6%points compared to the dry cycle) for constant rotational speed and a reduced fuel consumption of 11.5% at constant nominal power output of 100 kWe, also leading to an electrical efficiency of 37.1% (increase of 4.3%points). The simulation results also indicated that cycle humidification by direct injection of water is superior over the humidification using a saturation tower, which in turn is superior over the humidification through steam injection. As a final remark, we would like to highlight that, although, the optimal found cycle layout, the REVAPÒ concept with feedwater preheat, has a much more complex cycle layout compared to e.g. direct preheated water injection, the difference in performance (efficiency increase) is rather limited. A more complex cycle layout does not necessarily leads to much higher waste heat recovery and significantly better thermodynamic performance. So next to this thermodynamic performance, when choosing the final cycle layout for implementation in an actual mGT, the cycle complexity and possible technological challenges linked to the cycle concept, should also be considered. Given the limited market share of mGTs on the small-scale CHP market, commercial usage of humidified mGTs is not likely in the short term. However, the results presented in this paper can be used by future researchers/developers focusing on mGT applications for electrical power production only. The results of this investigation can help assessing the need for humidification, to increase the power output and electrical efficiency of the cycle, and to select the method to use based on the design specification (e.g. maximal performance or limited cycle changes). Additionally, several technological challenges, like for instance better material, still need to be overcome. Acknowledgement The first author was supported by a fellowship from the Fonds National de la Recherche Scientifique, FRS-FNRS (Communauté Française de Belgique). References [1] Gamou S, Ito K, Yokoyama R. Optimal operational planning of cogeneration systems with microturbine and desiccant air conditioning units. J Eng Gas Turbines Power 2005;127(3):606–14. [2] Gamou S, Yokoyama R, Ito K. Parametric study on economic feasibility of microturbine cogeneration systems by an optimization approach. J Eng Gas Turbines Power 2005;127(2):389–96. [3] Frost & Sullivan. Combined heat and power: integrating biomass technologies in buildings for efficient energy consumption; 2011. [4] Pilavachi PA. Mini- and micro-gas turbines for combined heat and power. Appl Therm Eng 2002;22(18):2003–14. [5] Galanti L, Massardo AF. Micro gas turbine thermodynamic and economic analysis up to 500 kWe size. Appl Energy 2011;88(12):4795–802.

11

[6] Delattin F, Bram S, Knoops S, De Ruyck J. Effects of steam injection on microturbine efficiency and performance. Energy 2008;33(2):241–7. [7] Loujendi D, Sani K, Tofigh A, Majidian A, Jouybari H. An economical evaluation of the water/steam injection in a CHP microturbine cycle. J Eng Sci Technol Rev 2012;5(2):20–5. [8] Montero Carrero M, De Paepe W, Parente A, Contino F. T100 mGT converted into mHAT for domestic applications: economic analysis based on hourly demand. Appl Energy 2015;164:1019–27. [9] Montero Carrero M, De Paepe W, Bram S, Musin F, Parente A, Contino F. Humidified micro gas turbines for domestic users: an economic and primary energy savings analysis. Energy 2016;117(2):429–38. [10] Stathopoulos P, Paschereit C. Retrofitting micro gas turbines for wet operation. A way to increase operational flexibility in distributed CHP plants. Appl Energy 2015;154:438–46. [11] Stathopoulos P, Paschereit CO. Operational strategies of wet cycle micro gas turbines and their economic evaluation. J Eng Gas Turbines Power 2015;138. V003T20A003. [12] Lee JJ, Jeon MS, Kim TS. The influence of water and steam injection on the performance of a recuperated cycle microturbine for combined heat and power application. Appl Energy 2010;87(4):1307–16. [13] Mochizuki K, Shibata S, Inoue U, Tsuchiya T, Sotouchi H, Okamoto M. New concept of a micro gas turbine based co-generation package for performance improvement in practical use. In: ASME conference proceedings. ASME Paper PWR2005-50364; 2005. p. 1305–10. [14] Zhang S, Xiao Y. Steady-state off-design thermodynamic performance analysis of a humid air turbine based on a micro turbine. In: ASME conference proceedings. ASME paper GT2006-90335; 2006. p. 287–96. [15] De Paepe W, Delattin F, Bram S, De Ruyck J. Steam injection experiments in a microturbine – a thermodynamic performance analysis. Appl Energy 2012;97:569–76. [16] De Paepe W, Delattin F, Bram S, Contino F, De Ruyck J. A study on the performance of steam injection in a typical micro Gas Turbine. In: ASME conference proceedings. ASME paper GT2013-94569; 2013a. p. V05AT23A011. [17] Ferrari ML, Pascenti M, Traverso AN, Massardo AF. Hybrid system test rig: chemical composition emulation with steam injection. Appl Energy 2012;97:809–15. [18] Renzi M, Riolfi C, Baratieri M. Influence of the syngas feed on the combustion process and performance of a micro gas turbine with steam injection. Energy Procedia 2017;105:1665–70. [19] Parente J, Traverso A, Massardo AF. Micro humid air cycle: Part A – thermodynamic and technical aspects. In: ASME conference proceedings. ASME paper GT2003-38326; 2003. p. 221–9. [20] Rao AD. Process for producing power. US patent no. 4829763; 1989. [21] De Paepe W, Delattin F, Bram S, De Ruyck J. Water injection in a micro gas turbine – assessment of the performance using a black box method. Appl Energy 2013;112:1291–302. [22] Nikpey H, Mansouri Majoumerd M, Assadi M, Breuhaus P. Thermodynamic analysis of innovative micro gas turbine cycles. In: ASME conference proceedings. ASME paper GT2014-26917; 2014. p. V03AT07A029. [23] Majoumerd MM, Somehsaraei HN, Assadi M, Breuhaus P. Micro gas turbine configurations with carbon capture – performance assessment using a validated thermodynamic model. Appl Therm Eng 2014;73(1):172–84. [24] De Paepe W, Contino F, Delattin F, Bram S, De Ruyck J. New concept of spray saturation tower for micro humid air turbine applications. Appl Energy 2014;130:723–37. [25] De Paepe W, Montero Carrero M, Bram S, Contino F. T100 micro gas turbine converted to full humid air operation: test rig evaluation. In: ASME conference proceedings. ASME paper GT2014-26123; 2014b. p. V03AT07A020. [26] De Paepe W, Montero Carrero M, Bram S, Contino F. T100 micro gas turbine converted to full humid air operation: a thermodynamic performance analysis. In: ASME conference proceedings. ASME paper GT2015-43267; 2015. p. V003T06A015. [27] Montero Carrero M, De Paepe W, Magnusson J, Parente A, Bram S, Contino F. Experimental characterisation of a micro humid air turbine: assessment of the thermodynamic performance. Appl Therm Eng 2017;118:796–806. [28] De Paepe W, Contino F, Delattin F, Bram S, De Ruyck J. Optimal waste heat recovery in micro gas turbine cycles through liquid water injection. Appl Therm Eng 2014;70(1):846–56. [29] Montero Carrero M, De Paepe W, Bram S, Parente A, Contino F. Does humidification improve the micro gas turbine cycle? Thermodynamic assessment based on Sankey and Grassmann diagrams. Appl Energy 2017b. http://dx.doi.org/10.1016/j.apenergy.2017.05.067 [in press]. [30] Jonsson M, Yan J. Humidified gas turbines – a review of proposed and implemented cycles. Energy 2005;30(7):1013–78. [31] Aspen Technology Inc.. Aspen plus v8.8. 2016. [accessed: 17-01-2017]. [32] De Paepe W, Montero Carrero M, Bram S, Parente A, Contino F. Towards higher micro gas turbine efficiency and flexibility – humidified mGTs. In: ASME conference proceedings. GT2017-64857; 2017. 13 pages [accepted for publication]. [33] Turbec AB. T100 microturbine CHP system: technical description ver 4.0; 2000–2001. [34] Bram S, De Ruyck J. Exergy analysis tools for ASPEN applied to evaporative cycle design. Energy Convers Manage 1997;38(15-17):1613–24. [35] De Ruyck J, Bram S, Allard G. REVAPÒ cycle: a new evaporative cycle without saturation tower. J Eng Gas Turbines Power 1997;119(4):893–7.

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001

12

W. De Paepe et al. / Applied Energy xxx (2017) xxx–xxx

[36] Takahashi T, Koda E, Mimaki T. A systematic analysis of the effect of air humidification to gas turbine systems. JSME Int J Ser B – Fluids Therm Eng 2002;45(3):530–5. [37] Nakhamkin M, Swensen EC, Wilson JM, Gaul G, Polsky M. The cascaded humidified advanced turbine (CHAT). J Eng Gas Turbines Power 1996;118 (3):565–71. [38] Van Liere J, Laagland G. The TOPHATÒ cycle. VDI-Berichte; 2000. p. 161–75. [39] Caresana F, Comodi G, Pelagalli L, Renzi M, Vagni S. Use of a test-bed to study the performance of micro gas turbines for cogeneration applications. Appl Therm Eng 2011;31(16):3552–8. [40] Çengel Y, Boles M. Thermodynamics: an engineering approach. McGraw-Hill series in mechanical engineering. Boston, US: McGraw-Hill Higher Education; 2006. ISBN 9780072884951..

[41] Belokon AA, Khritov KM, Klyachko LA, Tschepin SA, Zakharov VM, George Opdyke J. Prediction of combustion efficiency and NOx levels for diffusion flame combustors in HAT cycles. In: ASME conference proceedings. ASME paper GT2002-30609; 2002. p. 791–7. [42] Queiroz JA, Rodrigues VM, Matos HA, Martins F. Modeling of existing cooling towers in ASPEN PLUS using an equilibrium stage method. Energy Convers Manage 2012;64:473–81. [43] De Ruyck J, Bram S, Allard G. Humid air cycle development based on exergy analysis and composite curve theory. In: ASME conference proceedings. ASME paper 95-CTP-39; 1995. p. 10. [44] Rouhani S, Sohal M. Two-phase flow patterns: a review of research results. Prog Nucl Energy 1983;11(3):219–59.

Please cite this article in press as: De Paepe W et al. Waste heat recovery optimization in micro gas turbine applications using advanced humidified gas turbine cycle concepts. Appl Energy (2017), http://dx.doi.org/10.1016/j.apenergy.2017.06.001