Energy Conversion and Management 171 (2018) 1213–1225
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Simulating combined cycle gas turbine power plants in Aspen HYSYS Zuming Liu, Iftekhar A. Karimi
⁎
T
Department of Chemical & Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, 117585, Singapore
A R T I C LE I N FO
A B S T R A C T
Keywords: Simulation Gas turbine Combined cycle Power plant Aspen HYSYS
Combined cycle gas turbine (CCGT) power plants are becoming increasingly important for electricity generation. Enhancing their thermal performance is essential for mitigating carbon emissions. This paper aims to present a methodology for simulating the off-design operation of a triple-pressure reheat CCGT plant in Aspen HYSYS. The modeling equations that rigorously capture the full off-design characteristics of various plant components (i.e. compressor, combustor, turbine, heat recovery steam generator, and steam turbines) are implemented in Aspen HYSYS, and a specially tailored procedure is proposed for solving them. The modeling strategy and solution procedure can be extended to simulate the off-design operation of any CCGT plants and are generically applicable to other process simulators (e.g. Aspen Plus, Unisim, and Pro II). To evaluate the model’s performance, its predictions are compared with those of an equivalent model from GateCycle. The results show the predictions of the two models (Aspen HYSYS and GateCycle) agree well. The average differences for the power outputs and thermal efficiencies of the gas turbine, steam cycle, and CCGT plant are less than 2.0%, 1.5%, and 0.6%, respectively. Besides, the differences arise primarily from the different gas enthalpy calculations. Since the model enables easy integration with various energy systems and can be made dynamic for predicting real-time behavior in Aspen HYSYS, it is very useful with wide applications.
1. Introduction Energy and environment are the two major global concerns of this century. The global warming caused by the greenhouse gas emissions is an existential threat. CO2 is considered as the main cause, and more than 40% of the CO2 emissions stem from the power industry [1]. As a result, much effort is underway on producing clean, green, and efficient electric power. Combined Cycle Gas Turbine (CCGT) power plants are one promising solution due to their high thermal efficiencies and low CO2 emissions [2]. Nowadays, CCGT plants are undergoing widespread installations. Some countries like Singapore produce more than 96% of their electric power from CCGT plants [3]. Power plants operate under off-design (especially part-load) conditions during most of their lifetimes. For example, a power plant in Nigeria produced only 64.3% of its design capacity from 2001 to 2010 [4]. The part-load operation arises from several factors. First, the power demand is hardly steady and rarely equals the plant design capacity. Second, a power plant is required to maintain spinning reserves (surplus capacity) to guard against unforeseen peaks in demands. Third, a power plant may often be overdesigned to buffer against demand uncertainties. The part-load operation decreases the plant’s thermal efficiency, incurring higher fuel consumption and CO2 emissions. Therefore, strong incentives exist for studying and optimizing the part-load
⁎
Corresponding author. E-mail address:
[email protected] (I.A. Karimi).
https://doi.org/10.1016/j.enconman.2018.06.049 Received 14 February 2018; Received in revised form 12 June 2018; Accepted 13 June 2018 0196-8904/ © 2018 Elsevier Ltd. All rights reserved.
operation. To this end, rigorous simulation models that accurately capture the full details of a power plant’s part-load operations are needed. Such simulation models provide the basis for a variety of routine operational tasks, such as benchmarking, process control, process optimization, condition monitoring, fault diagnosis, performance analysis, and performance improvement. Zhang and Cai [5] proposed some analytical formulas for compressor and turbine and combined them to predict the gas turbine performance. Aklilu and Gilani [6] adopted the normalized parameters from [7] to describe the characteristics of compressor and turbine and developed a simulation model in Matlab [8] to identify the plant operation mode from field data. Zhang et al. [9,10] presented a simulation program in Excel to study the off-design characteristics of combined cycles under different design parameters. While models in Matlab or Excel offer much freedom in model formulation and are attractive from a cost perspective, they are not user-friendly and require much programming and approximations. In addition to the modeling process being tedious, complex, and error-prone, the models may suffer from numerical and convergence issues due to the complex nonlinear iterative calculations. On the contrary, commercial software offer a nice graphical user interface, superior reliability, and enhanced accuracy with little or no programming. Hence, commercial software such as GateCycle [11],
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
Nomenclature
af c ca cc cor d g in map max min out st s t
Symbols A C c1,c2,c3 Fcu m L LHV N P PL PR ΔP Qloss R T U W
area, m2 swallowing capacity IGV angle correction factors copper loss fraction mass flow rate, kg/s generator load lower heating value shaft speed, rpm pressure, bar percent part-load pressure ratio pressure drop, kPa heat loss, kW gas constant temperature, K overall heat transfer coefficient, kJ/(s m2 K) power, kW
Acronyms BFD CCGT ECON EVAP GT HP HPP HRSG IP IPP LP LPP RHT RP SPHT SC TET IGVC
Greek letters Δα γ π η λ κ ν φ Ω
IGV angle specific heat ratio expansion ratio efficiency constant constant specific volume (γ − 1)/γ combustor loading
Subscripts a
air filter compressor cooling air combustion chamber corrected value design condition flue gas inlet performance map maximum minimum outlet steam turbine steam turbine
block flow diagram combined cycle gas turbine economizer evaporator gas turbine high pressure high pressure pump heat recovery steam generator intermediate pressure intermediate pressure pump low pressure low pressure pump reheater recirculation pump superheater steam cycle Turbine exit temperature inlet guide vane control
air
programs (e.g. Thermoflow/GateCycle and Aspen HYSYS/Unisim) with their different architectures and properties, and complex interactions between them are difficult to manage. Therefore, it is desirable to simulate both power plants and various energy systems in one seamless environment or platform such as a more versatile process simulator. This is crucial to facilitate easy integration between power plants and these energy systems. Aspen HYSYS [39] is a powerful process simulator with a large library of ready-made component models and in-built accurate property packages. By connecting the various components via material and energy streams, Aspen HYSYS can simulate both the steady and dynamic performance of complex chemical/hydrocarbon fluid-based processes [40–44]. This enables the simulation of both power plants and associated energy systems or options. Hence, Aspen HYSYS does not have the aforementioned shortcomings and offers an attractive platform for simulating power plants. However, modeling the CCGT plants under off-design conditions in Aspen HYSYS is challenging due to its sequential modular nature. In Aspen HYSYS, all plant components must be solved in a sequential rather than simultaneous manner. The highly complex steam circuits that involve mass/energy recycle in the CCGT plants require simultaneous solution and thus pose significant challenges to Aspen HYSYS. Furthermore, detailed compressor map and turbine characteristics have to be used for simulating CCGT off-design performance. This requires clever constructs and implementation in Aspen HYSYS. Therefore, a tailored non-obvious procedure is needed for simulating the CCGT plants under off-design conditions in Aspen
EBSILON Professional [12], and Thermoflow [13] have been preferred for studying power plants. Silva et al. [14] developed a thermodynamic information system in GateCycle for detecting plant operation anomalies and evaluating the performance gain from eliminating them. Lee et al. [15] proposed an analysis tool in GateCycle for predicting the plant generation capacity using the correction curves of gas and steam turbines. Liu and Karimi [16] presented the necessary correlations for simulating a CCGT plant in GateCycle and proposed a simulation-based method for maximizing its part-load performance. Aminov et al. [17] evaluated the fuel saving and reduction in CO2 emissions from replacing a thermal power plant by a CCGT plant using EBSILON Professional. Since GateCycle, EBSILON Professional and Thermoflow are principally designed for power plants, they offer a nice simulation experience. However, their versatility is limited in modeling other energy systems or options (e.g. CO2 capture [18–20], Organic Rankine Cycles (ORCs) [21–23], fuel cells [24–26], LNG terminals [27–29], air separation [30–32], and absorption chillers [33–35]). For instance, GateCycle is unable to model these energy systems. Although EBSILON Professional and Thermoflow offer special blocks for some energy options (e.g. CO2 capture, fuel cells, air separation and absorption chillers), they simulate them as black boxes. Hence, they cannot offer the full simulation details and freedom for process modification. To avoid these shortcomings, Nord et al. [36] and Karimi et al. [37] modeled CO2 capture process in Aspen HYSYS and Unisim respectively, while Lee et al. [38] simulated air separation for a gasification process in Aspen HYSYS. However, doing so requires interfacing two separate simulation 1214
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
temperature corrosion. The HP steam expands in an HP steam turbine (HPST) and then mixes with the IP steam. The mixed steam enters the reheaters and then expands in an IP steam turbine (IPST). The exhaust steam from the IPST mixes with the LP steam and enters an LP steam turbine (LPST). The three STs share a common shaft that rotates at the same speed as the GT. After expansion, the exhaust steam from the LPST goes to a condenser, and the condensate is pumped back via the LPP to the LP economizer.
HYSYS, and to our knowledge, no study in the open literature has presented such a procedure. In this paper, a detailed model and a systematic procedure are presented for simulating the off-design operation of a triple-pressure reheat CCGT plant in Aspen HYSYS. The implementation of the rigorous modeling equations for various plant components in Aspen HYSYS is explained in detail. This produces an Aspen HYSYS model that captures the full details of the CCGT plant. A tailored procedure is then proposed for solving the Aspen HYSYS model. Finally, to evaluate the model’s performance, its simulation results are compared with those of an equivalent model from GateCycle.
2.2. CCGT simulation in Aspen HYSYS The following assumptions are made for simulating the triplepressure reheat CCGT plant in Aspen HYSYS.
2. Methodology
• The CCGT plant is at steady state. • The fuel combustion is complete in the combustor. • There are no leaks of water or flue gas from the HRSG. • The cooling water flow of the condenser is constant.
2.1. Combined cycle gas turbine (CCGT) power plant Fig. 1 shows the schematic of a triple-pressure reheat CCGT power plant. The plant comprises a Gas Turbine (GT), a Heat Recovery Steam Generator (HRSG), and three Steam Turbines (STs). The GT has an Air Compressor (AC) and a turbine running on a common shaft with a combustor in between. The common shaft rotates at a constant speed to deliver a fixed frequency (50 or 60 Hz) of power. The ambient air passes through an air filter to the AC, and the combustor uses the pressurized air from the AC to burn a gaseous fuel (e.g. natural gas) and feeds the hot gas into the turbine, where it expands to produce power. As the turbine blades are exposed to the hot gas from the combustor, some air from the AC exit is supplied to keep them cool. The exhaust gas from the turbine then goes through the HRSG, before being vented to the ambient as a flue gas. The HRSG recovers the remaining heat from the exhaust gas to produce steam. The HRSG comprises three steam generation subsystems: HighPressure (HP), Intermediate-Pressure (IP), and Low-Pressure (LP). Each subsystem has one feedwater pump (LPP, IPP, or HPP in Fig. 1), one or more economizers, one evaporator, and one or more superheaters. The feedwater from each pump gets preheated in the economizers, boiled in the evaporator, and superheated in the superheaters. Two reheaters (RHT1 and RHT2 in Fig. 1) are located between the HP superheaters. Moreover, two desuperheaters (DESHT1 and DESHT2 in Fig. 1) between the HP superheaters and reheaters moderate the temperatures of HP steam and reheat steam for safe operation by injecting water. Furthermore, a recirculation pump (RP in Fig. 1) recycles some hot water from the LP economizer exit back into its feed to prevent low-
The modeling equations describing the full off-design characteristics of various plant components rigorously are mainly from [16]. This paper focuses on (1) the details and challenges of implementing them in Aspen HYSYS, and (2) a tailored procedure for efficiently and reliably solving the resulting Aspen HYSYS model for the CCGT plant. The PengRobinson fluid package is used for air, fuel, and exhaust gas, while ASME steam table is employed for water and steam. Fig. 2 shows our complete Block Flow Diagram (BFD) for the CCGT plant in Aspen HYSYS. 2.2.1. Air filter The air filter is simulated by the Control Valve module (AFT in Fig. 2(a)) in Aspen HYSYS. The pressure drop through the air filter (ΔPaf ) is given by the following equation [11,16]. 1.84
m⎞ ΔP = ΔPd ⎛ ⎝ md ⎠ ⎜
⎟
−1
⎛ T ⎞⎛ P ⎞ ⎝ Td ⎠ ⎝ Pd ⎠
⎜
⎟⎜
⎟
(1)
where ΔP is the pressure drop, m is the mass flow rate, T is the temperature, P is the pressure, and subscript d denotes the design condition. While this is a precise approach to model the air filter, the pressure drop is usually quite small (< 100 Pa) even at the design condition. Hence, the pressure drop across the air filter is set as a fixed percentage (0.5%) of the ambient pressure. It is computed in a Spreadsheet module
Fig. 1. Schematic of a triple-pressure reheat CCGT power plant. 1215
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
CCGT plant – GT (a)
CCGT plant – SC (b)
CCGT plant – SC (c)
Fig. 2. Block flow diagram (BFD) for the CCGT plant in Aspen HYSYS: (a) Gas turbine (GT), (b–c) Steam cycle (SC).
1216
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
2.2.3. Combustor The combustor is simulated by the Conversion Reactor module (COMB in Fig. 2(a)) in Aspen HYSYS. The fuel combustion is defined as a set of conversion reactions with 100% conversions. Then, for a given fuel flow (mf ), the pressure drop in and heat loss from the combustor are computed using the following equations [16,47] in SS-1, and exported to COMB in Fig. 2(a).
(SS1 in Fig. 2(a)) and then exported to AFT in Fig. 2(a). 2.2.2. Air compressor (AC) The AC operating characteristics can be described by its performance map, which for a typical GT compressor is expressed in terms of pressure ratio (or isentropic efficiency) versus corrected mass flow and corrected speed (see Fig. 3). However, the compressor vendors do not share actual performance maps except with their customers. Hence, real compressor maps are hard to find in the open literature, and an example performance map is shown in Fig. 3. The map relates the following dimensionless operational variables [16,45].
Relative corrected mass flow:
mcor , r
2
⎡ min Tin ⎞ ⎛ min, d Tin, d ⎞ ⎤ ΔPcc = ΔPcc, d ⎢ ⎜⎛ ⎟/ ⎟⎥ Pin ⎠ ⎜ Pin, d ⎝ ⎠⎦ ⎣⎝
min Tin ⎞ ⎛ min, d Tin, d ⎞ = ⎜⎛ ⎟/ ⎜ ⎟ Pin, d ⎝ Pin ⎠ ⎝ ⎠
Qloss, d = (1−ηcc, d ) mf , d LHVf
(8a)
min Ω = 1.8 Pin exp(Tin/300)
(8b)
(2a)
Relative pressure ratio:
PRr = (PR−1)/(PRd−1)
Relative isentropic efficiency: Relative corrected speed:
ηr = η / ηd
Ncor , r = (N / Tin )/(Nd/ Tin, d )
x2−x x −x1 f (x1, y1) + f (x2 , y1) x2−x1 x2−x1
x −x x −x1 f (x , y2 ) = 2 f (x1, y2 ) + f (x2 , y2 ) x2−x1 x2−x1
mf ⎛ Ω ⎞1.6 mf , d ⎝ Ωd ⎠
(2b)
Qloss = Qloss, d
(2c)
where ΔPcc is the pressure drop in the combustor, LHV is the lower heating value, Qloss is the heat loss, Ω is the combustor loading, and subscript f denotes the fuel.
(2d)
where η is the efficiency, N is the shaft speed, and PR = Pout / Pin . Subscript cor denotes the corrected value, r denotes the relative value, in denotes the inlet, and out denotes the outlet. The AC is simulated by the Compressor module (AirCOMP in Fig. 2(a)) in Aspen HYSYS. However, AirCOMP can only accept operating curves expressed in terms of pressure head (or isentropic efficiency) versus volumetric flow, and not the one in Fig. 3. Hence, a special procedure is needed to overcome this limitation. For supplying the performance map in Fig. 3 to AirCOMP, equispaced parabolic lines indexed by an auxiliary coordinate called β [46] (0.4 ≤ β ≤ 1.0 ) are introduced on the map, as shown in Fig. 4. The β lines intersect the speed lines (Ncor , r ) and each (β , Ncor , r ) defines a unique point on the compressor map. Every point (β , Ncor , r ) on the map represents a unique triplet of PRr , mcor , r , and ηr . Each of these three is stored as a two-dimensional look-up table in SS-1 with β and Ncor , r as arguments. Given any (x , y ) = (β , Ncor , r ) , PRr , mcor , r and ηr are obtained from these tables via bilinear interpolations (see Fig. 5) as follows.
f (x , y1) =
(7)
⎜
⎟
(8c)
2.2.4. Turbine In a heavy-duty turbine, blade cooling is necessary to prevent turbine blades from overheating. In this paper, the turbine blade cooling is simulated by bleeding two air streams from the AC exit and injecting them into the turbine inlet and exit, respectively. As shown in Fig. 2(a), the stator cooling air mixes with the main hot gas at the turbine inlet. The mixed gas then expands in the turbine. Finally, the rotor cooling air mixes with the expanded gas at the turbine outlet. The stator and rotor cooling flows can be computed by Eq. (9) [48] in SS-1. Their mixing with the turbine inlet and outlet gases is simulated by the Mixer module in Aspen HYSYS.
Tca, d ⎞ P mca = mca, d ⎜⎛ ca ⎟⎞ ⎛ P ⎝ ca, d ⎠ ⎝ Tca ⎠ ⎜
0.5
⎟
(9)
where mca is the mass flow rate of the cooling air, and Pca and Tca are the pressure and temperature of the cooling air. The turbine flow characteristics can be described by the following constant swallowing capacity [49–51].
(3a)
C= (3b)
min, d Tin, d min Tin = = Cd κPin κd Pin, d
(10a)
γ+1
f (x , y ) =
y−y1 y2 −y f (x , y2 ) f (x , y1) + y2 −y1 y2 −y1
κ=
(3c)
γ ⎛ 2 ⎞ γ−1 ⎜ ⎟ Rg ⎝ γ + 1 ⎠
where f denotes PRr , mcor , r or ηr , x denotes β , y denotes Ncor , r , and ( x1, y1), ( x1, y2 ), ( x2 , y2 ), and ( x2 , y1) denote the closest four points that surround ( x , y ) in a table of f (x , y ) . Modern ACs have variable inlet guide vanes (IGVs) whose openings are varied to regulate the air flow. This opening is measured by an IGV angle Δα (normally 0 ≤ Δα ≤ 40º) , where Δα = 0 corresponds to the fully open IGVs. For a given Δα , Eqs. (4)–(6) [47] can be used to correct the PRr , mcor , r , and ηr read from the map.
PRr , IGV = PRr (1−c1 Δα )
(4)
mcor , r , IGV = mcor , r (1−c2 Δα )
(5)
ηr , IGV = ηr (1−c3 Δα 2)
(6)
where c1, c2 , and c3 are vane angle correction factors. Then, given β and Ncor , r , the AC can be simulated in Aspen HYSYS as follows. Compute PRr , mcor , r and ηr from Eq. (3), PRr , IGV , mcor , r , IGV and ηr , IGV from Eqs. (4)–(6), and PR , min and η from Eq. (2) all in SS-1. Export PR and η to AirCOMP in Fig. 2(a) and min to its inlet stream (S1).
Fig. 3. Relativized compressor map. 1217
(10b)
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
P (S7) is unknown, and depends on the HRSG pressure drop (ΔPHRSG ). ΔPHRSG is computed from Eq. (1), and simulated by a Control Valve module (DUCT in Fig. 2(a)) before the HRSG. P (S7) needs to be iterated such that P (S7)−ΔPHRSG = Pamb . This is achieved by using an Adjust module (ADJ-TP in Fig. 2(a)) as follows. Given a P (S7) , compute ηt from Eq. (10) in SS-1, and export its value to TURB in Fig. 2(a). Aspen HYSYS simulates TURB and gives T (S6) , where T denotes temperature. Com' = P (S7)−ΔPHRSG in SS-1. In ADJ-TP, select P (S7) as the adpute Pamb ' justed variable, Pamb as the target variable, and specify Pamb as the specified target value. The complete GT simulation procedure in Aspen HYSYS is shown in Fig. 6. 2.2.5. Heat recovery steam generator (HRSG) The HRSG comprises a series of economizers, evaporators, and superheaters. The economizers and superheaters are heat exchangers that extract the waste heat from the exhaust gas for heating water and steam, respectively. Hence, they are simulated by the LNG Exchanger module in Aspen HYSYS. However, the evaporators involve change in state along with phase equilibrium, and hence their simulation is different and more challenging than the economizers and superheaters. Normally, an evaporator consists of a boiler and a steam drum. The boiler is a heat exchanger that extracts the waste heat from the exhaust gas to produce water/steam mixture, while the steam drum is a phaseseparator that separates water/steam mixture into saturated water and steam. Thus, the boiler and steam drum are simulated by the LNG Exchanger and Separator modules in Aspen HYSYS. The steam generation process in the evaporator is simulated as follows. The boiler extracts the waste heat from the exhaust gas to generate water/steam mixture. The water/steam mixture mixes with the subcooled water from the economizer in the steam drum to produce saturated steam and water. The saturated steam goes to the superheater while the saturated water returns to the boiler. Note that the LNG Exchanger module is essentially the same as the Heat Exchanger module here, since there are only two streams involved in heat exchange and mass and energy balances are of our only interest. Our motivation is to just make the BFD look cleaner (less convoluted) as shown in Fig. 2(b–c). The LNG Exchanger module in Aspen HYSYS needs a UA value (overall heat transfer coefficient × heat transfer area) for heat exchange calculation. Since U is mainly affected by the exhaust gas flow under off-design conditions, UA is computed in SS-1 as follows [53], and then exported to each LNG Exchanger module.
Fig. 4. β lines on a relativized compressor map.
Fig. 5. Bilinear interpolation for reading the compressor map expressed in terms of (β, Ncor , r ) as coordinates.
mg ⎞ UA = (UA)d ⎛⎜ ⎟ m ⎝ g, d ⎠
where C is the swallowing capacity, κ is a constant, γ is the specific heat ratio, and Rg is the gas constant. The ambient temperature and GT shaft speed fix Ncor , r . Then, for a given mf , β that satisfies C = Cd fixes the AC operating point on the map. Determining the correct β requires iterations, and is done by using an Adjust module (ADJ-BETA in Fig. 2(a)) in Aspen HYSYS. In ADJ-BETA, β is selected as the adjusted variable, C computed in SS-1 is chosen as the target variable, and Cd is supplied as the specified target value. Once a β is given, Aspen HYSYS can simulate AirCOMP and COMB, and adjust β to achieve C = Cd . Then, the gas into the turbine is fully known, and Aspen HYSYS can simulate the turbine fully. The turbine is simulated by the Expander module (TURB in Fig. 2(a)) in Aspen HYSYS. The turbine isentropic efficiency (ηt ) is estimated by the following semi-empirical formula [10,52].
ηt = ηt , d
N Nd
Tin, d Tin
πdφ−1 ⎛ N λ−(λ−1) π φ−1 ⎜ Nd ⎝
Tin, d Tin
πdφ−1 ⎞ π φ−1 ⎟ ⎠
0.8
(12)
where mg is the gas mass flow rate. The water/steam pressure losses in HRSG heat exchangers vary as follows during off-design operation [11]. 1.98
m⎞ ΔP = ΔPd ⎛ m ⎝ d⎠ ⎜
1.98
m⎞ ΔP = ΔPd ⎛ m ⎝ d⎠ ⎜
for water
⎟
⎟
⎛ν⎞ ⎝ νd ⎠
⎜
⎟
for steam
(13)
(14)
where ν is the specific volume of steam. The water/steam pressure losses are computed in SS1, and exported to the corresponding LNG Exchanger modules. The HP steam and reheat steam from the HRSG may exceed their maximum allowable temperatures (THPSmax and TRHSmax ) during offdesign operation. For safe operation, two desuperheaters are installed to moderate their temperatures by injecting water. The two desuperheaters are simulated by two Mixer modules (DeSH1 and DeSH2 in Fig. 2(b)) in Aspen HYSYS. Moreover, two Adjust modules (ADJ-SH1 and ADJ-SH2 in Fig. 2(b)) are employed to control the temperatures of HP steam and reheat steam under off-design operation. In ADJ-SH1 and ADJ-SH2, m (S26a) and m (S41a) are selected as the adjusted variables, T (S36) and T (S51) are chosen as the target variables, and THPSmax and
(11)
where π = Pin/ Pout , φ = (γ −1)/ γ , λ is a known constant, and subscript t denotes turbine. For solving TURB, its outlet pressure [P (S7) ] is needed. Here, P denotes pressure, and S7 denotes the stream in Fig. 2(a). However, 1218
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
Fig. 6. GT simulation procedure in Aspen HYSYS.
steam flow during off-design operation, and thus can be corrected as follows [12]:
TRHSmax are supplied as the specified target values. Here, m refers to mass flow rate, T refers to temperature, and S26a, S41a, S36, and S51 refer to the streams in Fig. 2. The LP economizer feedwater temperature cannot fall below the minimum allowable temperature (TFWmin ) to avoid low-temperature corrosion. Hence, some warm water from the LP economizer exit is recycled back to its feed. An Adjust module (ADJ-RCF in Fig. 2(c)) is employed to control the LP economizer feedwater temperature. In ADJRCF, m (S65b) is selected as the adjusted variable, T (S63) is chosen as the target variable, and TFWmin is supplied as the specified target value.
5
=
2.2.8. Condenser The condenser is simulated by the LNG Exchanger module (CONDR in Fig. 2(c)) in Aspen HYSYS. It condenses the water/steam mixture from the LPST to saturated water. Thus, the vapor fraction of S60 in Fig. 2(c) is set to 0. Because the heat transfer in the condenser is very efficient and the cooling water flow is kept unchanged, the UA value for the condenser is assumed constant under off-design conditions. Then, the condenser operating pressure varies to fully condense the water/ steam mixture. An Adjust module (ADJ-CDP in Fig. 2(c)) is employed to find the right condenser pressure. In ADJ-CDP, P (S59) is selected as the adjusted variable, the relative UA error of the CONDR is chosen as the target variable, and 0 is supplied as the specified target value. Finally, the generator efficiency [57] and the power outputs for the GT, Steam Cycle (SC), and CCGT plant are computed in SS-1 as follows.
ms, d Tin, d 2 Pin2 , d−Pout ,d
(16)
where subscript s denotes steam, and st denotes steam turbine. The HPST, IPST, and LPST are simulated by three Expander modules (HPT, IPT, and LPT in Fig. 2(b–c)) in Aspen HYSYS, respectively. The isentropic efficiencies for HPST, IPST, and LPST are computed in SS-1 using Eq. (16), and exported to HPT, IPT, and LPT.
2.2.7. Steam turbines (STs) CCGT plants usually adopt sliding pressure operation for STs under off-design conditions. This implies that the throttling valves of STs are fully open and the steam pressures in the HRSG are regulated by water pumps to match ST characteristics. Since valve throttling is eliminated, sliding pressure operation produces a better plant performance than constant pressure operation [54]. The off-design characteristics of an ST can be described by the Stodola’s method [55,56].
2 Pin2 −Pout
3
2
m m ⎤ + 0.0585 ⎜⎛ s ⎟⎞ + 0.0163 ⎜⎛ s ⎟⎞ + 0.98⎥ ⎝ ms, d ⎠ ⎝ ms, d ⎠ ⎦
2.2.6. Water pumps The water pumps (HPP, IPP, LPP, and RP) are simulated by the Pump module along with the Control Valve module in Aspen HYSYS. The pump curves for the water pumps can be either supplied as user input or generated automatically inside the Pump module. The control valves (HPCV, IPCV, and LPCV) serve as regulating the steam pressures in the HRSG to match the ST operation.
ms Tin
4
m m m ⎡ ηst = ηst , d ⎢−0.1035 ⎜⎛ s ⎟⎞ + 0.2357 ⎜⎛ s ⎟⎞ −0.1872 ⎜⎛ s ⎟⎞ m m m s , d s , d ⎝ ⎠ ⎝ ⎠ ⎝ s, d ⎠ ⎣
(15)
ηgen =
The isentropic efficiency (ηst ) of an ST is mainly affected by the 1219
Lgen ηgen, d 2 Lgen ηgen, d + (1−ηgen, d )[(1−Fcu ) + Fcu Lgen ]
(17)
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
WGT = (Wt −Wc / ηmech ) ηgen
(18)
and STs must be solved jointly and special constructs are necessary for back-pressure calculations.
(19)
The procedure in Fig. 7 is designed ingeniously to address the above challenges. Simulating the evaporator is its key first step. Consider the HP evaporator (HP drum and HP boiler in Fig. 2(b)). For simulating it, Recycle modules (R3 and R9 in Fig. 2(b)) are needed for specifying stream conditions (e.g. flow, pressure, temperature, etc.). The Recycle module in Aspen HYSYS is a mathematical operation and has an inlet stream and an outlet stream. For example, R9 has S30a as the inlet stream and S30b as the outlet stream. In the Recycle module, the stream conditions can be transferred forwards from the inlet to the outlet. Aspen HYSYS first utilizes the outlet stream conditions as assumed values to solve the flowsheet sequentially around the Recycle module. Based on the differences between the inlet and outlet stream conditions, Aspen HYSYS updates the outlet stream conditions iteratively until the inlet stream conditions match the out stream conditions within the tolerances specified in the Recycle module. The simulation of the HP evaporator is performed with Recycle modules as follows. Since the HP boiler produces steam/water mixture, the vapor fraction of S32 can be set to any value between 0 and 1. Then, two Recycle modules (R3 and R9) are employed for specifying P (S30b) , T (S12b) , and T (S30b) . Aspen HYSYS uses them to solve the HP boiler and HP drum, as the flow, pressure, and composition of the exhaust gas streams within the HRSG are already known. The HP boiler computes m (S32) from the energy balance and heat transfer equations, which enables the HP drum to calculate m (S30b) and m (S33) from mass and energy balances and water/steam equilibrium. This means that the HP evaporator can automatically compute its own water flow. Next, m (S30a) is set to m (S30b) in SS-1, which fixes the water flow in the HP circuit. The pressure losses in the HP economizers are computed by Eq. (13) in SS-1. In the following, two Recycle modules (R2 and R10) are used for specifying T (S11b) , P (S26b) , T (S26b) and m (S26a) . Meanwhile, m (S26b) is set to m (S26a) in SS-1. Aspen HYSYS solves HP SHPT1 and
WSC = (WHPST + WISPT + WLPST ) ηgen−(WHPP + WIPP + WLPP + WRP )
WCCGT = WGT + WSC
(20)
where ηgen is the generator efficiency, ηmech is the AC mechanical efficiency, Lgen = Win/ Win, d , where Win is the work input to the generator, and Fcu is the copper loss fraction. This completes the development of an Aspen HYSYS model for simulating the CCGT plant under off-design conditions. 2.3. Simulation procedure Consider simulating the CCGT plant operation for given mf and Δα . Aspen HYSYS simulates the GT by iterating β and P (S7) using two Adjust modules (ADJ-BETA and ADJ-TP). The GT simulation procedure is presented in Fig. 6. When ADJ-BETA and ADJ-TP converge, the GT is solved and the turbine exhaust gas flow, temperature, and composition become known. Now, the SC must be simulated for this known exhaust gas conditions. However, simulating the SC in Aspen HYSYS is challenging due to the following factors. (1) Aspen HYSYS is a sequential modular simulator, in which the SC components have to be solved in a sequential manner. However, the HP, IP, and LP steam circuits that involve mass/energy recycle in the SC require simultaneous rather than sequential solution. This poses significant challenges to Aspen HYSYS. For configuring the SC components to be solved sequentially, the Recycle module in Aspen HYSYS is needed. Nonetheless, determining how many Recycle modules should be used and where to place them are combinatorially demanding and require clever thinking. (2) The HRSG steam conditions (flow, pressure, and temperature) have to satisfy the ST characteristics, namely Eq. (15). Thus, the HRSG
Fig. 7. SC simulation procedure in Aspen HSYSY. 1220
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
3. Model evaluation on an example power plant
HP SHPT2, and computes T (S12b) and T (S36) , respectively. The pressure losses in HP SPHT 1 and HP SPHT 2 are computed by Eq. (14) in SS-1. If T (S12a) does not match T (S12b) within the specified tolerance in R3, Aspen HYSYS updates T (S12b) , and the HP evaporator simulation continues again. This process repeats until T (S12a) and T (S12b) are within the specified tolerance. If T (S36) exceeds THPSmax , ADJ-SH1 adjusts m (S26a) to prevent the HP steam from over-temperature. Now, the HPT inlet steam conditions are fully known. However, they may not match the HPST characteristics, namely Eq. (15). Hence, knowing the HPT inlet flow and temperature, the HPT expected inlet pressure [P'(S36) ] is computed from Eq. (15), and P (S30a) is back calculated in SS-1 by adding the pressure losses in HP SPHT 1 and HP SPHT2. Moreover, the pressure iteration in R9 is activated. If P (S30a) does not match P (S30b) within the specified tolerance in R9, Aspen HYSYS updates P (S30b) until the difference between P (S30a) and P (S30b) falls within the specified tolerance. Hence, when R9 converges, the HP steam conditions match the HPST characteristics. This completes the simulation of the HP circuit in a sequential manner, starting from the HP evaporator. The same simulation logic is applied to the IP/LP circuits. Finally, a Recycle module (R6) is used for the LP economizer water recirculation; an Adjust module (ADJ-RCF) adjusts the water recirculation flow to control the LP economizer feedwater temperature; an Adjust module (ADF-CDP) adjusts the condenser pressure to fully condense the water/steam mixture from the LPST. The detailed simulation procedure for the SC is presented in Fig. 7. The Recycle modules in the BFD and their stream variables are summarized in Table 1. All the variables are transferred forwards in the Recycle modules. Aspen HYSYS iterates on the stream variables systematically until the Recycle modules converge. Therefore, when all the Recycle and Adjust modules converge, the SC is solved successfully. Inlet guide vane control (IGVC) is usually employed for part-load operations in CCGT plants. IGVC simultaneously manipulates mf to achieve the desired part-load and Δα to maintain TET at its design value (TETd ) [57,58]. Two Adjust modules are used to implement IGVC in Aspen HYSYS. As shown in Fig. 2(a), ADJ-FF adjusts mf , and ADJ-IGV adjusts Δα . In ADJ-FF, m (NG) is selected as the adjusted variable, WCCGT / WCCGT , d × 100 computed in SS-1 is chosen as the target variable, and the desired percent part-load (PL%) is supplied as the specified target value. In ADJ-IGV, Δα is selected as the adjusted variable, T (S7) is chosen as the target variable, and TETd is supplied as the specified target value. Now, given a part-load (PL ), the Aspen HYSYS model can simulate the triple-pressure reheat CCGT plant in Fig. 1. To converge the model smoothly, some guidelines are proposed here. First, some simple correlations for the minimum and maximum parameters in the Adjust modules are developed, as shown in Table 2. Second, based on the minimum and maximum parameters, the initial guess and step size for each Adjust module are set to 0.5 × (Minimum + Maximum) and 0.1 × (Maximum−Minimum) , respectively. Third, the Adjust modules should be activated progressively. For instance, ADJ-BETA, ADJ-IGV, ADJ-TP, ADJ-CDP, and ADJ-RCF are first activated. Then, ADJ-SH1 and ADJ-SH2 are activated one at a time. Finally, ADJ-FF is activated. In this way, the Aspen HYSYS model converges smoothly for a given PL . It is clear from the above details that developing and solving the CCGT model in Aspen HYSYS require ingenious constructs and thinking based on a full understanding of Aspen HYSYS. By giving a detailed and explicit procedure, this paper makes CCGT simulation easy for the researchers, and thus makes a significant contribution. Given the plant design data, the model requires only one input, namely the desire partload (PL ), and produces all the useful outputs, including but not limited to, the power outputs and efficiencies of the GT, SC, and CCGT plant. Moreover, it can either work stand-alone, or be easily integrated with various energy systems (e.g. CO2 capture, ORCs, fuel cells, LNG terminals, air separation, and absorption chillers). Furthermore, it can be made dynamic by Aspen HSYSY Dynamics for predicting the plant realtime behavior. Therefore, it is very useful and has wide applications.
The performance of our Aspen HYSYS model is evaluated with an example CCGT plant. Since real operational data for CCGT plants are not available in the open literature, an alternative way is to compare its predictions with those of an equivalent model built in GateCycle, a widely used commercial software in the power industry. The following data is used for evaluation. The plant is assumed to use IGVC for part-load operation. Table 3 shows the design parameters of the CCGT plant and Fig. 3 presents the AC performance map. Moreover, c1 = 0.01, c2 = 0.01 and c3 = 0.0001 in Eqs. (4)–(6) [47], λ = 2.083 in Eq. (11) [10], and Fcu = 0.48 in Eq. (17) are used in this paper. Furthermore, both THPSmax and TRHSmax are assumed to be 565 °C while TFWmin is assumed to be 50 °C. Table 4 presents the design performance of the CCGT plant in Aspen HYSYS and GateCycle. In the following, the relative deviations (RD) between the two simulation models (Aspen HYSYS and GateCycle) defined by Eq. (21) in GT, SC, and CCGT performance are evaluated.
RD
(%) =
HYSYS
Result−GateCycle GateCycle Result
Result
× 100
(21)
3.1. Gas turbine (GT) performance Fig. 8 shows the relative deviations for the key operating parameters of the AC and turbine. Clearly, nearly all are within 1.0%. Moreover, the average deviation for all operating parameters in Fig. 8 is less than 0.5%. The minor discrepancies for the operating parameters in Fig. 8 arise from the differences in gas enthalpy calculations. For calculating gas enthalpies, GateCycle uses the NASA method [59], in which gases are assumed to be ideal. Aspen HYSYS uses the Peng-Robinson equation-of-state [60], which is based on real gas experimental data. The NASA method uses two separate fourth-order (5-parameter) temperature-dependent polynomials to calculate gas enthalpies below and above 1000 K (726.85 °C), respectively. Aspen HYSYS directly calculates gas enthalpies from the Peng-Robinson equation-of-state. Hence, Aspen HYSYS predicts a higher (lower) gas enthalpy below (above) 1000K than GateCycle, as shown in Fig. 9. The differences in gas enthalpy predictions affect the complex interactions between the AC and turbine, represented by the matching between the compressor map (Fig. 3) and turbine characteristics (Eq. (10)). This results in the minor discrepancies shown in Fig. 8. Because of these minor discrepancies, Aspen HYSYS predicts a lower GT power output and efficiency than GateCycle, as shown in Fig. 10. Moreover, as the plant load decreases, the differences in gas enthalpy predictions drive the GT power output and efficiency of Aspen HYSYS farther way from those of GateCycle. As a result, the relative deviations in the GT power output and efficiency increase with decreasing plant load. However, their maximum deviations are within 3.2%, and the average deviation is less than 2.0%. Table 1 Stream variables for the Recycle modules in Aspen HYSYS. All are transferred in the forwards direction.
1221
Module
Stream variable
R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11
T(S10b) T(S11b) T(S12b) T(S14b) T(S17b) P(S65b) P(S53b) P(S43b) P(S30b) P(S26b) P(S41b)
and and and and and and
T(S65b) T(S53b) T(S43b) T(S30b) T(S26b) T(S41b)
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
Table 2 Minimum and maximum parameters for the Adjust modules in Aspen HYSYS. Module
Adjusted variable
Base value
Minimuma (%)
Maximuma (%)
ADJ-BETA ADJ-FF ADJ-IGV ADJ-TP ADJ-SH1 ADJ-SH2 ADJ-RCF ADJ-CDP
β m(NG) Δα P(S7) m(S26a) m(S41a) m(S65ba) P(S59)
0.7 NG design flow 100% S7 design pressure S34 design flow S34 design flow S65b design flow S59 design pressure
0.274PL + 70.0 0.8352PL + 14.0 0.6241PL + 35.0b 0.0144PL + 97.0 0 0 −0.355PL + 128.0 0.422PL + 54.0
0.274PL + 76.0 0.8352PL + 18.0 0.6241PL + 38.0b 0.0144PL + 100.0 −0.0163PL + 2.0 −0.0066PL + 1.0 −0.355PL + 146.0 0.422PL + 60.0
a b
Maximum (Minimum) value/Base value × 100. IGV opening (100 − Δα).
Table 3 Design parameters of the CCGT plant. Parameter/variable Ambient condition Pressure (kPa) Temperature (°C) Molar fraction Fuel condition Pressure (bar) Temperature (°C) Molar fraction Gas turbine Inlet air flow (kg s−1) Inlet air pressure loss (%) Compressor pressure ratio Compressor isentropic efficiency (%) Compressor mechanical efficiency (%) Fuel flow (kg s−1) Combustor efficiency (%) Combustor pressure loss (%) Combustor exit temperature (°C) Turbine inlet temperature (°C) Turbine exhaust temperature (°C) Heat recovery steam generator (HRSG) HP/IP/LP steam temperatures (°C) HP/IP/LP pinch point temperatures (°C) HP/IP/LP approach point temperatures (°C) HP SPHT 1 steam outlet temperature (°C) RHT 1/2 steam outlet temperature (°C) HP ECON 1/2 water outlet temperature (°C) Pressure losses on gas/water/steam sides (%) Steam turbines (STs) HP/IP/LP ST inlet pressure (bar) HP/IP/LP ST isentropic efficiency (%) Condenser Pressure (kPa) Cooling water temperature (°C) Cooling water temperature rise (°C) Generator Generator efficiency (%) Shaft speed (rpm)
Table 4 Design performance of the CCGT plant in Aspen HYSYS and GateCycle. Value
Performance
Aspen HYSYS
GateCycle
101.3 15.0 77.30% N2, 20.74% O2, 1.01% H2O, 0.03% CO2, 0.92% Ar
GT power (MW) GT efficiency (%) SC power (MW) SC efficiency (%) Plant net power (MW) Plant efficiency (%)
253.2 36.17 139.8 30.73 393.0 56.14
257.2 36.78 137.8 30.33 395.0 56.49
30.0 10.0 87.08% CH4, 7.83% C2H6, 2.94% C3H8, 1.47% N2, 0.68% CO2 635.0 0.5 15.4 88.0 99.0 14.74 99.5 3.5 1405.0 1328.0 615.0 565.0/297.0/295.0 10.0/10.0/10.0 8.0/10.0/16.4 510.0 520.0/565.0 208.0/280.0 1.5/5.0/3.0
98.8/24.0/4.0 87.0/91.0/89.0 7.4 25.0 10.0 98.5 3000
3.2. Steam cycle (SC) performance Fig. 8. Relative deviations for the operating parameters of the AC (a) and turbine (b).
Figs. 11–13 show the relative deviations for the operating parameters of the HPST, IPST, and LPST. Since both Aspen HYSYS and GateCycle use the ASME steam table for water and steam, the relative deviations in Figs. 11–13 are primarily from their gas models. Because
1222
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
Fig. 12. Relative deviations for the IPST operating parameters.
Fig. 9. Gas enthalpy difference between Aspen HYSYS and GateCycle.
Fig. 13. Relative deviations for the LPST operating parameters.
Fig. 10. Relative deviations for the power outputs and efficiencies of the GT, SC, and CCGT plant.
why both HPST and IPST inlet temperatures from Aspen HYSYS and GateCycle are the same. The steam pressure profiles from Aspen HYSYS and GateCycle are jointly determined by the HRSG and STs. Hence, their variations are dependent on the profiles of stream flows and temperatures. Clearly, in Figs. 11–13, the relative deviations in steam pressures and temperatures for HPST, IPST, and LPST are all less than 0.6%, and the relative deviations in steam flows and power outputs are within 2.4%. Moreover, the relative deviations in the SC power output and efficiency range between 1.2 and 2.0% as shown in Fig. 10, and the average deviation is less than 1.5%. 3.3. CCGT performance Fig. 10 shows the relative deviations for the plant power output and efficiency. Since the GT dominates the plant performance, Aspen HYSYS predicts a relatively lower power output and efficiency than GateCycle. The relative deviations in the plant power output and efficiency are less than 0.6% for 100–40% loads. The reason is that Aspen HYSYS predicts a higher SC power output, which compensates its lower GT power output. However, the two simulation models are comparable in terms of their simulation results. Our comparison is useful for anybody wanting to use Aspen HYSYS instead of GateCycle, and vice versa.
Fig. 11. Relative deviations for the HPST operating parameters.
the SC operates below 1000 K , Aspen HYSYS predicts a higher gas enthalpy than GateCycle. Hence, Aspen HYSYS predicts higher steam flows and higher ST power outputs. The higher steam flows lead to higher ST isentropic efficiencies according to Eq. (16). This enables the HPST and IPST to expand to lower temperatures in Aspen HYSYS than GateCycle. Since the LPST usually expands to two-phase (water/steam) region, both Aspen HYSYS and GateCycle predict the same LPST outlet temperature. On the other hand, HP steam and reheat steam exceed their maximum allowable temperatures under IGVC; hence, desuperheaters are activated to prevent them from over-temperature. This is
4. Conclusions In this paper, a detailed Aspen HYSYS model was presented for simulating the off-design operation of a triple-pressure reheat CCGT plant. The challenges of implementing the rigorous modeling equations 1223
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
for various plant components were addressed, and a tailored procedure was proposed for their solution. The Aspen HYSYS model captures the full off-design details of the CCGT plant including compressor map, turbine characteristics, and flow-dependent variables, such as pressure drops and heat transfer coefficients. To our knowledge, this is the first fully-detailed Aspen HYSYS model in the open literature for a CCGT plant during off-design operation. The modeling strategy and solution procedure presented in this paper can be extended to simulate any CCGT plants in Aspen HYSYS. Moreover, they are generically applicable to other commercial process simulators, such as Aspen Plus, Unisim, and Pro II. The presence of mass/energy recycle in a CCGT plant poses significant challenges to such sequential modular simulators; hence, the specially tailored and ingenious simulation procedure is the major contribution of this paper. Using the data from an example CCGT plant, it was shown that the predictions from the Aspen HYSYS model and an equivalent GateCycle model are acceptably comparable. The relative deviations for the most operating parameters of the GT and SC were within 1.0% and 0.6%, respectively. Specifically, the average deviations in the power outputs and thermal efficiencies of the GT, SC, and CCGT plant were less than 2.0%, 1.5%, and 0.6%, respectively. A thorough study was also done in this paper to analyze the key causes for these various deviations. It was found that the different procedures for computing gas enthalpies are the main factor. As the world moves to integrate CCGT plants within wider and diverse energy systems to conserve fuels and reduce CO2 emissions, more general purpose simulators such as Aspen HYSYS versus stand-alone specialized simulators such as GateCycle are becoming essential. In this context, Aspen HYSYS offers several important advantages over GateCycle such as wider and more versatile physical property packages, easy integration with other energy systems or options (e.g. CO2 capture, ORCs, fuel cells, LNG terminals, air separation, and absorption chillers), dynamic simulation, and real-time optimization.
[11] [12] [13] [14]
[15] [16] [17]
[18] [19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27] [28]
Acknowledgement [29]
Zuming Liu acknowledges ACTSYS Process Management Consultancy Company, Singapore for hosting his industrial internship under a ring-fenced Graduate Research Scholarship from the National University of Singapore. The authors thank Mr Norman Lee, MD of ACTSYS for inspiring them to work on GT modeling. They further thank Mr Norman Lee, Dr Yu Liu, and Mr Weiping Zhang of ACTSYS for several enlightening discussions and preliminary information on the GT operation in a CCGT plant. They acknowledge the support from the National University of Singapore via a seed grant R261-508-001-646/ 733 for CENGas (Center of Excellence for Natural Gas). They also acknowledge the use of Aspen HYSYS and GateCycle under academic licenses.
[30]
[31]
[32]
[33]
[34]
References [35] [1] IEA. World Energy Outlook; 2017. [2] Rao AD. Combined cycle systems for near-zero emission power generation, natural gas-fired combined cycle (NGCC) systems; 2012. [3] EMA Singapore; 2017. < https://www.ema.gov.sg/Statistics.aspx > . [4] Oyedepo SO, Fagbenle RO, Adefila SS, Adavbiele SA. Performance evaluation and economic analysis of a gas turbine power plant in Nigeria. Energy Convers Manage 2014;79:431–40. [5] Zhang N, Cai R. Analytical solutions and typical characteristics of part-load performances of single shaft gas turbine and its cogeneration. Energy Convers Manage 2002;43:1323–2337. [6] Aklilu BT, Gilani SI. Mathematical modeling and simulation of a cogeneration plant. Appl Therm Eng 2010;30:2545–54. [7] Al-Hamdan QZ, Ebaid MSY. Modeling and simulation of a gas turbine engine for power generation. J Eng Gas Turbines Power 2006;128:302. [8] MathWorks; 2016. < https://www.mathworks.com/ > . [9] Zhang G, Zheng J, Yang Y, Liu W. Thermodynamic performance simulation and concise formulas for triple-pressure reheat HRSG of gas–steam combined cycle under off-design condition. Energy Convers Manage 2016;122:372–85. [10] Zhang G, Zheng J, Xie A, Yang Y, Liu W. Thermodynamic analysis of combined
[36]
[37] [38]
[39] [40] [41]
[42]
1224
cycle under design/off-design conditions for its efficient design and operation. Energy Convers Manage 2016;126:76–88. GateCycle 6.1. General Electric Company; 2013. Ebsilon Professional; 2016. < http://www.sofbid.com > . Thermoflow; 2016. < https://www.thermoflow.com/ > . Silva JAM, Venturini OJ, Lora EES, Pinho AF, Santos JJCS. Thermodynamic information system for diagnosis and prognosis of power plant operation condition. Energy 2011;36:4072–9. Lee JH, Kim TS. Kim E-h. Prediction of power generation capacity of a gas turbine combined cycle cogeneration plant. Energy 2017;124:187–97. Liu Z, Karimi IA. Simulation and operation of a combined cycle gas turbine power plant for part-load operation. Chem Eng Res Des 2018;131:29–40. Aminov Z, Nakagoshi N, Xuan TD, Higashi O, Alikulov K. Evaluation of the energy efficiency of combined cycle gas turbine. Case study of Tashkent thermal power plant, Uzbekistan. Appl Therm Eng 2016;103:501–9. Oh S-Y, Binns M, Cho H, Kim J-K. Energy minimization of MEA-based CO2 capture process. Appl Energy 2016;169:353–62. Soltani SM, Fennell PS, Mac Dowell N. A parametric study of CO2 capture from gasfired power plants using monoethanolamine (MEA). Int J Greenhouse Gas Control 2017;63:321–8. Esquivel-Patiño GG, Serna-González M, Nápoles-Rivera F. Thermal integration of natural gas combined cycle power plants with CO2 capture systems and organic Rankine cycles. Energy Convers Manage 2017;151:334–42. Budisulistyo D, Wong CS, Krumdieck S. Lifetime design strategy for binary geothermal plants considering degradation of geothermal resource productivity. Energy Convers Manage 2017;132:1–13. Kim IS, Kim TS, Lee JJ. Off-design performance analysis of organic Rankine cycle using real operation data from a heat source plant. Energy Convers Manage 2017;133:284–91. Baccioli A, Antonelli M. Organic flash cycles: off-design behavior and control strategies of two different cycle architectures for waste heat recovery applications. Energy Convers Manage 2018;157:176–85. Bakalis DP, Stamatis AG. Full and part load exergetic analysis of a hybrid micro gas turbine fuel cell system based on existing components. Energy Convers Manage 2012;64:213–21. Lee YD, Ahn KY, Morosuk T, Tsatsaronis G. Exergetic and exergoeconomic evaluation of a solid-oxide fuel-cell-based combined heat and power generation system. Energy Convers Manage 2014;85:154–64. Sharifzadeh M, Meghdari M, Rashtchian D. Multi-objective design and operation of solid oxide fuel cell (SOFC) triple combined-cycle power generation systems: integrating energy efficiency and operational safety. Appl Energy 2017;185:345–61. Fahmy MFM, Nabih HI, El-Rasoul TA. Optimization and comparative analysis of LNG regasification processes. Energy. 2015;91:371–85. Xue F, Chen Y, Ju Y. Design and optimization of a novel cryogenic Rankine power generation system employing binary and ternary mixtures as working fluids based on the cold exergy utilization of liquefied natural gas (LNG). Energy 2017;138:706–20. Mehrpooya M, Sharifzadeh MMM, Zonouz MJ, Rosen MA. Cost and economic potential analysis of a cascading power cycle with liquefied natural gas regasification. Energy Convers Manage 2018;156:68–83. Querol E, Gonzalez-Regueral B, Ramos A, Perez-Benedito JL. Novel application for exergy and thermoeconomic analysis of processes simulated with Aspen Plus®. Energy 2011;36:964–74. Esfilar R, Mehrpooya M, Moosavian SMA. Thermodynamic assessment of an integrated biomass and coal co-gasification, cryogenic air separation unit with power generation cycles based on LNG vaporization. Energy Convers Manage 2018;157:438–51. Deng Z, Jin B, Zhao Y, Gao H, Huang Y, Luo X, et al. Process simulation and thermodynamic evaluation for chemical looping air separation using fluidized bed reactors. Energy Convers Manage 2018;160:289–301. Somers C, Mortazavi A, Hwang Y, Radermacher R, Rodgers P, Al-Hashimi S. Modeling water/lithium bromide absorption chillers in ASPEN Plus. Appl Energy 2011;88:4197–205. Mansouri R, Boukholda I, Bourouis M, Bellagi A. Modelling and testing the performance of a commercial ammonia/water absorption chiller using Aspen-Plus platform. Energy 2015;93:2374–83. Ochoa AAV, Dutra JCC, Henríquez JRG, dos Santos CAC. Dynamic study of a single effect absorption chiller using the pair LiBr/H2O. Energy Convers Manage 2016;108:30–42. Nord LO, Anantharaman R, Bolland O. Design and off-design analyses of a precombustion CO2 capture process in a natural gas combined cycle power plant. Int J Greenhouse Gas Control 2009;3:385–92. Karimi M, Hillestad M, Svendsen HF. Natural gas combined cycle power plant integrated to capture plant. Energy Fuels 2012;26:1805–13. Lee JJ, Kim YS, Cha KS, Kim TS, Sohn JL, Joo YJ. Influence of system integration options on the performance of an integrated gasification combined cycle power plant. Appl Energy 2009;86:1788–96. Aspen HYSYS V9; 2017. < www.aspentech.com > . Taimoor AA, Muhammad A, Saleem W, Zain-ul-abdein M. Humidified exhaust recirculation for efficient combined cycle gas turbines. Energy 2016;106:356–66. Mehrpooya M, Zonouz MJ. Analysis of an integrated cryogenic air separation unit, oxy-combustion carbon dioxide power cycle and liquefied natural gas regasification process by exergoeconomic method. Energy Convers Manage 2017;139:245–59. Wang Y, Tang Q, Wang M, Feng X. Thermodynamic performance comparison between ORC and Kalina cycles for multi-stream waste heat recovery. Energy Convers Manage 2017;143:482–92.
Energy Conversion and Management 171 (2018) 1213–1225
Z. Liu, I.A. Karimi
[43] Alobaid F, Starkloff R, Pfeiffer S, Karner K, Epple B, Kim H-G. A comparative study of different dynamic process simulation codes for combined cycle power plants – Part A: part loads and off-design operation. Fuel 2015;153:692–706. [44] Alobaid F, Starkloff R, Pfeiffer S, Karner K, Epple B, Kim H-G. A comparative study of different dynamic process simulation codes for combined cycle power plants – Part B: start-up procedure. Fuel 2015;153:707–16. [45] Palmer CA, Erbes MR, Pechtl PA. GateCycle performance analysis of the LM2500 gas turbines utilizing low heating value fuels. ASME Cogen Turbo Power, ASME IGTI 1993;8:69. [46] Smooth C 8.3. Preparing compressor maps for gas turbine performance modeling; 2015. < http://www.gasturb.de/compressor-maps.html > . [47] Kurzke J. GasTurb 12 - design and off-design performance of gas turbines; 2015. < http://wwwgasturbde/ > . [48] Palmer CA, Erbes MR. Simulation methods used to analyze the performance of the GE PG6541B gas turbines utilizing low heating value fuels. ASME IGTI CogenTurbo; 1994. [49] Reynolds WC, Perkins HC. Engineering thermodynamics. McGraw-Hill; 1977. [50] Kim YS, Lee JJ, Kim TS, Sohn JL. Effects of syngas type on the operation and performance of a gas turbine in integrated gasification combined cycle. Energy Convers Manage 2011;52:2262–71. [51] Kim YS, Lee JJ, Kim TS, Sohn JL, Joo YJ. Performance analysis of a syngas-fed gas
[52] [53] [54]
[55] [56] [57] [58]
[59]
[60]
1225
turbine considering the operating limitations of its components. Appl Energy 2010;87:1602–11. Lu S, Lin R. Gas turbine steady-state design and off-design characteristic general model. Inst Eng Thermophys 1996:404–7. Holman JP. Heat transfer. McGraw-Hill; 1997. Sanchez Fernandez E, Sanchez del Rio M, Chalmers H, Khakharia P, Goetheer ELV, Gibbins J, et al. Operational flexibility options in power plants with integrated postcombustion capture. Int J Greenhouse Gas Control 2016;48:275–89. Csanady GT. Theory of turbomachines. New York: McGraw-Hill; 1964. Cooke DH. On prediction of off-design multistage turbine pressures by Stodola’s ellipse. J Eng Gas Turbines Power 1985;107:596–606. Haglind F. Variable geometry gas turbines for improving the part-load performance of marine combined cycles – gas turbine performance. Energy 2010;35:562–70. Domachowski Z, Dzida M. Influence of inlet guide vane control on combined cycle power plant transients. In: Proceedings of ASME turbo expo, Amsterdam, Netherlands; 2002. McBride BJ, Gordon S, Reno MA. Coefficients for calculating thermodynamic and transport properties of individual species. NASA Technical Memorandum, 4513; 1993. Aspen HYSYS V9. Aspen HYSYS V9 Help. Peng-Robinson calculation methods; 2017.
