Comparative Study On Modelling Of Gas Turbines In Combined Cycle Power Plants

Proceedings of the 4 th International Middle East Power Systems Conference (MEPCON 0), Cairo University, Egypt, December 9-2, 200, Paper ID 37. Comparative Study On Modelling Of Gas Turbines In Combined Cycle Power Plants H. E. M. A. Shalan M. A. Moustafa Hassan A. B. G. Bahgat El-Kureimat Power Station Electrical Power Department Electrical Power Department Ministry of Electricity Faculty of Engineering, Cairo University Faculty of Engineering, Cairo University Cairo, Egypt Giza, Egypt Giza, Egypt hanyemam980@yahoo.com mmustafa@eng.cu.edu.eg agbahgat@hotmail.com Abstract Gas turbines are important for electric power generation specially the Combined Cycle Power Plants (CCPP). For this electric power generation, the dynamics of the gas turbines become increasingly more important. In order to study such dynamics, accurate models of gas turbines are needed. Recently, several gas turbine models have been proposed with different degree of complexity and success. The purpose of this work is concerned with understanding, modelling, and analysing the behaviour of the gas turbinebased plants to investigate the power system problems. This purpose is achieved by a complementary and comparative study of different dynamic models response that published in different literature for Combined Cycle Power Plants (CCPP). Among these models, there are three models were completely simulated using Matlab/Simulink. It is easy to conclude that the obtained results via these simulations in this study are highly matched with the results presented in the related scientific articles. The study illustrates the effectiveness and accuracy of frequency dependant model as well as the detailed model of gas turbines in CCPP. Index Terms: Combined Cycle Power Plant (CCPP), Gas Turbines Modeling, Modeling and simulation, Electric Power plants. T I. 0BINTRODUCTION HE gas turbine is a main part of the current power plant, which produces a great amount of energy for its size and weight. The gas turbine has established growing service in the past 40 years in the power industry both among utilities and merchant plants as well as the petrochemical industry, and utilities throughout the world. The last 20 years has seen a noticeable growth in Gas Turbine Technology. The growth is developed by the enhancement of materials technology, new coatings and new cooling schemes. This, with the conjunction of increase in compressor pressure ratio, has increased the gas turbine thermal efficiency from about 5 to over 45 percent, which is suitable for power plants. In the past, large coal and nuclear power plants dominated the electric power generation. However, natural gas-fired turbines now dominate the field of Power generation because of their black start capabilities, higher efficiencies, lower capital costs, shorter installation times, better emission characteristics, and abundance of natural gas supplies. The construction cost of gas turbine power plants is roughly half that of comparable conventional fossil fuel steam power plants, which were the primary power plants until the early 980s. More than half of all power plants to be installed in the foreseeable future are forecast to be gas turbine or combined gas-steam turbine types. Current low prices for crude oil make fuels such as diesel, kerosene, and clean gaseous fuels such as natural gas the most desirable for gas turbines. However, these fuels will become much more expensive and will eventually run out. So, provisions must therefore be made to burn alternative fuels. Now, gas turbines are used in a wide range of applications. The two major application areas of gas turbine engines are: Aircraft propulsion. Electric power generation. For electric power generation, common uses include stationary power generation plants (electric utilities) and mobile power generation engines (ships and aircraft). The term Combined Cycle Power Plant (CCPP) describes the combination of gas turbine generator(s) (Brayton cycle) with turbine exhaust waste heat boiler(s) and steam turbine generator(s) (Rankine cycle) for the production of electric power. During the last decades there has been continuous development of combined cycle power plants due to their increased efficiency and their low emissions, as well as reduced natural gas prices. However, in a large-scale blackout occurred in Malaysia in August 996, CCPP and gas turbine plants sequentially tripped out. The cause of this chain trip was thought to be a system frequency drop. Also, there were the blackout events in Italy, Denmark/southern Sweden and the USA/Canada, which resulted in major economic losses. Also strong crises in electric demand appeared in Egypt in summer of 200, which made more focus on CCPP. Therefore, A mathematical model of a CCPP is needed, including relevant control and protective functions. In this study, a simple model for CCPP will be developed in the Simulink environment of MatLAB. This to help in obtaining accurate models, which are highly needed. In the last decades, several gas turbine models have been proposed with different degree of complexity and success. It should be known that the complete gas turbine model consists of the 970

turbine thermodynamics, the fuel system, and the control loops, as illustrated in Figure (). Some models implement all of them together such as [, 2], whereas the other models implement a separate block for the turbine dynamics such as [3, 3]. This article is focusing on the comparative study for various gas turbines characteristics (thermodynamics) only. These characteristics are published in different literature. The fuel system and control loops will be discussed later in another research results. Furthermore, this study is considered a complementary study for different gas turbines models as well. Several gas turbine modelling are based on different models. models in recent years such as [4, 5, 0, 4, 5]. Figure (2) shows the extended model with IGV (Rowen-II). a) Rowen s Model Assumptions: As illustrated in [], Rowen-I model is based on the following assumptions: It is a heavy duty gas turbine, simple cycle, singleshaft, generator drive only. Allowable speed range is between 95 and 07 percent of rated speed. Open Inlet Guide Vanes (IGV) only, i.e. no Heat Recovery Steam Generator (HRSG). ISO conditions apply, i.e., T A = 5 o C P A = 0.325 kpa Table (): Parameters of Rowen s Model [] Parameter Description Value E CR Combustion reaction time delay 0.0 sec Figure (): Gas Turbine System Overview Some of them were discussed in details in [8]. These are listed below:. Rowen s Model. 4. GGOV Model. 2. IEEE Model. 5. CIGRE Model. 3. Frequency Dependent (FD) Model. 6. Detailed Model. The current study focuses on the important three models. The considered models are:. Rowen s Model. 2. Frequency Dependent (FD) Model. 3. Detailed Model. II. Rowen s Model E TD Turbine and exhaust delay 0.04 sec T CD T R Compressor discharge time constant Turbine rated exhaust temperature 0.2 sec 950 o F It is clear that a significant assumption modification had been adopted in Rowen-II model, where the gas turbine had been equipped with modulating IGVs, and hence assumed to be a part of a CCPP. In 983, W. I. Rowen provided a model of a gas turbine that can be used for dynamic performance studies of a power system []. Rowen s model consists of a set of algebraic equations describing the steady-state characteristics of the gas turbine thermodynamics, simple time delays, and a few related controls including the temperature control, governor, and acceleration control. The model has been utilized to investigate the impacts of governor on system operation. In 992, Rowen extended the original model as in [2], to include IGVs and their effect on the gas turbine dynamics, especially the exhaust temperature. The extended Rowen s model [2] is enabled a more accurate modelling of a gas turbine operation installed as part of a CCGT. Therefore, it is considered the starting point and backbone for development of most CCGT Figure (2): Rowen s Model with IGV [2] b) Rowen s Model Equations: There are four main function blocks in Rowen s models as follows: 97

a) Block F to calculate the exhaust temperature (T X ). b) Block F 2 to calculate the output torque of the gas turbine (TRQ D ). c) Block F 3 to calculate the flow of the exhaust gasses (W X ). d) Block F \ to calculate the exhaust temperature like F but, modified to include the impacts of changing air flow via IGV as well as ambient temperature. The governing equations for each block are described in [, 0, 4, 7, 8]. This model was simulated using MatLAB/Simulink environment [6]. Where T R Turbine rated exhaust temperature, o F. W F Mass fuel flow, pu. N Turbine rotor speed, pu. The Parameters for Rowen s turbine dynamics are presented in Table (). c) Results of Rowen s Model and Discussion: The relationships between the various inputs-outputs variables of the gas turbine are presented in this subsection. As shown in Figure (3), the output power depends mainly on the fuel flow, where it is increased (linearly as assumed) with increasing of the fuel flow. Increasing of fuel flow lead to increase in the exhaust temperature, as illustrated in Figure (4). Also, the relationship of the exhaust temperature to output power is shown in Figure (5), where it is highly matched with the obtained results in [7]. The effect of the IGV position on the exhaust gasses is shown in Figures (6), and (7). Increase of the IGV position will increase the exhaust gas flow, and the temperature of the exhaust gas will decrease accordingly. It should be known that, the IGV position data is obtained from [7]. Furthermore, the exhaust flow (W X ) is inversely proportional to the ambient temperature (T A ). This relationship between W X and T A is represented as in Figure (8). It should be noted that, this relationship is depicted at fully open IGV position, and full fuel flow rate (i.e. Ligv = pu & W F = ). W X T X (Fahrenheit) 0.92 0.9 0.7 0.75 5 0.9 0.95 Ligv Figure (6): Turbine Exhaust Flow versus Inlet Guide Vanes (IGVs) Position 600 550 500 P MG 0.4 0.2 0 0 0. 0.2 0.3 0.4 0.5 0.7 0.9 W F Figure (3): Output Power (P MG ) versus Fuel Flow (W F ) 000 W X ( pu ) 450 0.7 0.75 5 0.9 0.95 Ligv Figure (7): Exhaust Temperature versus Inlet Guide Vanes (IGV) Position.02 T X ( o F) 800 600 400 200 0 0. 0.2 0.3 0.4 0.5 0.7 0.9 W F Figure (4): Exhaust Temperature (T X ) versus Fuel Flow (W F ) T X ( o F ) 000 800 600 400 200 0 0. 0.2 0.3 0.4 0.5 0.7 0.9 P MG Figure (5): Turbine Exhaust Temperature (T X ) versus Power Output (P MG ) III. 55 60 65 70 75 80 85 T A ( o F ) Figure (8): Turbine Exhaust Flow (W X ) versus Ambient Temperature (T A ) Frequency Dependent (FD) Model Many of gas turbine models are not suitable for determining the frequency dependency of the gas turbine. To be able to analyse incidents with abnormal system frequency behaviour, the frequency dependence of the gas turbine model must be taken into account. This is the main aim of FD model which is based on the physical principles. It is developed in order to clarify the effects that shaft speed and ambient temperature have on output power [6]. 972

a) Structure of FD Model: d) FD Model Equations: As shown in Figure (), the gas turbine consists of an axial compressor, a combustion system and a turbine. The related input variables are fuel flow (W F ) and air flow (W A ). The related output variables are mechanical power output (P MG ) and the exhaust heat to the Heat Recovery Steam Generator (HRSG) as determined by the exhaust gas flow (W X ) and exhaust temperature (T X ). The fuel flow and air flow are adjusted to provide a desired power output while maintaining the desired level of exhaust temperature for efficient heat transfer to the HRSG. Air flow may be regulated by Inlet Guide Vanes (IGV) and is also a function of ambient air temperature (T A ), ambient pressure (P A ), and shaft speed (N). The structure of the FD gas turbine model is given in [6, ]. The FD model is illustrated using SIMULINK environment [6] in Figure (9). This subsection shows the thermodynamic equations representing the dynamic behaviour of the gas turbine that modelled before. It is clear that, the model could be divided into five separate blocks: i) Air flow block. ii) Compressor Pressure Ratio (CPR) block. iii) Exhaust temperature block. iv) Output power block. v) Maximum continuous power output block. The details for each block are given in [6]. The parameters of the gas turbine equations using FD model are illustrated in [6, 9]. However, the various parameters of the model (e.g., A o, A, and A 2, etc.) are obtained directly from the test data of actual machines [7]. e) Results and Discussion of FD Model: Figure (9): FD Model Structure Using SIMULINK b) FD Model Assumptions: As illustrated in [6, ], this model is based on the following assumptions: It is expected that the model is valid for variations in shaft speed between 95 to 05 percent and for unit loading above half load, which are similar to Rowen s model assumptions. The model is not intended for use for simulation of start-up, shutdown or detailed dynamics associated with the performance of the gas turbine combustion systems. The exhaust gas flow (W X ) is practically equal to the air flow (W A ) since the fuel flow (W F ) is much smaller than the air flow. The shaft power required to drive the compressor is constant. W F solely determines P MG and the only significant dynamics effects between their changes are time lags. This point is different with respect to Rowen model. The contribution of this model (which based on physical principles) is to clarify that, the maximum power output of a gas turbine (P MX ) depends on shaft speed (and hence the system frequency), and the ambient temperature. The model was completely simulated using Matlab/Simulink environment [6]. The obtained results are very close to those given in [6]. Changes in frequency are equivalent to changes in shaft speed and would result in a change in airflow. This change then translates firstly into a change in the pressure ratio across the compressor and secondly into a change in fuel level (in order to maintain the given firing temperature). These changes will directly affect the maximum power output, as shown in Figure (0). It is clear from the obtained results that, the performance of the gas turbine is heavily affected by ambient conditions. Any parameter affecting the mass flow of the air (W A ) entering the gas turbine will have an impact on the performance of the gas turbine. When the ambient temperature decreases, the air becomes denser, resulting in an increase in the Compressor Discharge Pressure (CDP). In other words, the gas turbine is operating more efficiently due to a reduction in ambient temperature. As shown in Figure (), an important relation is reported between the ambient temperature (T A ) and the maximum power output (P MX ). These results highly match the illustrated results in [6]. For more clarification and comparison, an actual relationship between T A and P MX is depicted as shown in Figure (2) for practical 265 MW HDGT. For more details [9] could be referred. 973

W A.02 0.92.05.04.03.02.0 (a) 0.99 0.97 0.95 Using a simplified steam turbine model, assuming that the power generation depends exclusively on the heat recovery from the gas turbine. Neglecting the pressure loss in the combustor (i.e., P 2 P 3 ).. N 4 (b) T A = 5 o C CPR(bar) 3.5 3 2.5 P MX 0.9 0.7 T A = 35 o C 2.05.04.03.02.0 N 0.99 0.97 0.95 0.95 0.97 0.99.0.02.03.04.05 N.02 (c) Figure (): Effects of Ambient Temperature on Maximum Continuous Power Output, Using FD Model P MX 0.92.05.04.03.02.0 N Figure (0): Effects of Speed Changes on different turbine parameters (a) Air Flow (b) Compressor Pressure Ratio (c) Maximum Continuous Power 0.99 0.97 0.95 IV. Detailed Model This model, as was introduced in [5], is based mainly on the modelling proposed in [3, 3] (IEEE model). Other similar models are presented in different literature, which, however, are slightly different. This model extends the gas turbine modelling by adding a simplified steam turbine model. Therefore, the Detailed Model is valid for a single-shaft combined cycle plant. The main aim for this model is to study the dynamic behaviour of the CCPP in the presence of a frequency drop. It should be noted that, Vournas [5] extends the results of Kakimoto and Baba [2, 3] by including a supervisory control of the combustion temperature. (a) a) Structure of the Detailed Model: Detailed model which giving the available thermal power to the gas turbine and the steam turbine is modelled by algebraic equations. These equations are relating the adiabatic compression and expansion, as well as to the heat exchange in the recovery boiler [3, 2, 3, 5, 7, 8]. The Detailed Model is represented using SIMULINK environment [6] in Figure (3). b) Detailed Model Assumptions: Some assumptions for the Detailed model are adopted, such as: In the gas turbine, the mixture of air and gas is approximately equal to airflow (i.e., W F << W A and therefore W F + W A W A ). (b) Figure (2): Actual Relationship between T A and P MX of 265 MW HDGT [8] (a) A Summer Day (b) A Winter Day c) Equations of Detailed Model: Algebraic equations that describe the CCPP model can be classified according to [2, 5] into two main areas as below: i. The gas turbine equations. The compressor. The Combustor. The turbine. ii. The steam turbine equation. The details of the turbines equations are discussed and illustrated in [2, 3, 5] for the Detailed model. 974

It should be noted that the in the IEEE model are worked out in Kelvin, whereas expressed in Fahrenheit in Rowen, and normalized in per unit values in the Detailed model. V. Main Comparative points for the GT Models The main comparative points for the different GTs models that discussed in this study are summarized and listed in [9] as well as in Table (2). However, the main important models consider IGV which gives nearer representation to practical. VI. Conclusions Figure (3): Turbine Thermodynamics in Detailed Model d) Results and Discussion of Detailed Model: The Detailed model was simulated using MatLAB/Simulink environment [6]. Figure (4) shows the effect of the increase of fuel flow on various output variables such as the firing temperature (turbine inlet temperature), the exhaust temperature, and the output power. It is well known that, the gas turbine produces approximately two-thirds of the total power output of a typical combined-cycle power plant, whereas the other one-third produced by the steam turbine. This point is indicated in Figure (5), and the obtained result is highly matched with the results presented in []. Also Gas Turbine responds quickly with any change in the combined power. P GT, t f, t e Mechanical Power 0.4 0.2 0 0.3 0.4 0.5 0.7 0.9 W F Figure (4): Effect of Fuel Flow on GT Temperatures and Output Power 0.4 0.2 P CC P ST P GT 0 0 0 20 30 40 50 60 70 80 90 00 Time (sec) Figure (5): Relationship between the GT and ST Outputs P GT t e t f The main objective of this proposed study is to present an overview of existing gas turbine models and to compare between them. To that extent, different gas turbine models are presented, discussed, and compared. The compared models are of different level of accuracy. However, they are suitable for different types of studies and have been utilized for different purposes. Among these models, there are three models were completely simulated using MatLAB/Simulink. It is easy to conclude that the obtained results via these simulations in this study are highly matched with the results presented in the related scientific articles. The study illustrates the usefulness and accuracy of frequency dependant model as well as the detailed model of gas turbines in CCPP, since these three models consider the IGV. VII. References [] W. I. Rowen, Simplified mathematical representations of heavy-duty gas turbines, ASME J. Eng. Power, Vol. 05, pp. 865 869, 983. [2] W.I. Rowen, Simplified Mathematical Representations of Single Shaft Gas Turbines in Mechanical Drive Service, Turbomachinery International, pp. 26-32., July/August 992. [3] F. P. de Mello and D. J. Ahner, Dynamic models for combined cycle plants in power system studies, IEEE Trans. Power Syst., vol. 9, pp. 698 708, Aug. 994. [4] A. Bagnasco, B. Delfino, 6. B. Denegri, and S.Massucco, Management and Dynamic Performances of Combined Cycle Power Plants During Parallel and Islanding Operation, IEEE Transactions on Energy Conversion, Vol. 3, No. 2, June 998. [5] J. H. Kim, T. W. Song, T. S. Kim, and S. T. Ro, Model development and simulation of transient behavior of heavy duty gas turbines, ASME J. Eng. Gas Turbines and Power, vol. 23, pp. 589 594, 200. [6] K. Kunitomi, A. Kurita, H. Okamoto, Y. Tada, S. Ihara, P. Pourbeik, and W. W. Price, Modeling frequency dependency of gas turbine output, presented at the IEEE Power Eng. Soc., Winter Meeting, Columbus, OH, Vol. 2, Jan. 28 Feb., 200. [7] L. M. Hajagos and G. R. Berube, Utility Experience with Gas Turbine Testing and Modeling, presented at the Power Engineering Society Winter Meeting, Columbus, OH, Vol., Jan. 28 Feb., 200. 975

[8] CIGRE Task Force C4.02.25, Modeling of Gas Turbines and Steam Turbines in Combined Cycle Power Plants, 2003. [9] L. Pereira, J. Undrill, D. Kosterev, D. Davies, and S. Patterson, "New Thermal Turbine Governor Modeling for the WECC" Modeling & Validation Work Group, Oct., 2002. [0] G. Lalor and M. O Malley, Frequency Control on an Island Power System with Increasing Proportions of Combined Cycle Gas Turbines, in Proc. IEEE Powertech Conf., Bologna, Italy, Jun. 2003. [] K. Kunitomi, A. Kurita, Y. Tada, S. Ihara, W. W. Price, L. M. Richardson, and G. Smith, Modeling combined-cycle power plant for simulation of frequency excursions, IEEE Trans. Power Syst., Vol. 8, No. 2, pp. 724 729, May 2003. [2] Kazuhiro Baba and Naoto Kakimoto, "Dynamic Behavior of a Combined Cycle Power Plant in the Presence of a Frequency Drop", Electrical Engineering in Japan, Vol. 43, No. 3, 2003. [3] Naoto Kakimoto, and Kazuhiro Baba, Performance of Gas Turbine-Based Plants During Frequency Drops, IEEE Trans. on Power Syst., Vol. 8, No. 3, August 2003. [4] Gillian Lalor, Julia Ritchie, Damian Flynn, and Mark J. O Malley, The Impact of Combined-Cycle Gas Turbine Short-Term Dynamics on Frequency Control, IEEE Trans. on Power Syst., Vol. 20, No. 3, August 2005. [5] John Mantzaris, and Costas Vournas, Modelling and Stability of a Single-Shaft Combined Cycle Power Plant, Int. J. of Thermodynamics, Vol. 0 (No. 2), pp. 7-78, June 2007. [6] MATLAB, The Language of Technical Computing, Version 7.6.0.324 (R2008a), SIMULINK, Dynamic system Simulation for Matlab, Version 7. (R2008a), 2008. [7] Siemens AG, Power Generation Group, Siemens Gas Turbines Manuals, Model V94.3A (SGT5-4000F), Windows Turbine-Generator Analysis Systems, "WIN-TS". [8] Ministry of Electricity and Energy of Egypt, official internet site, available at (Hwww.moee.gov.egH), updated June 200. [9] H. E. M. Abdelwahab, " Modeling and Control of Gas Turbine in Combined Cycle Power Plant", M.Sc Thesis, Cairo University, Giza, Egypt, End of 200. Model Point of Comparison Table (2): Main Comparative Points For The Considered GT Models [9] Rowen-I Rowen-II IEEE FD GGOV CIGRE Detailed Publishing Year 983 992 994 200 2002 2003 2007 Main Article [] [2] [3] [6] [8] [9] [5] Setup Configuration GT Representation & Control Loops GT Representation SCGT, and cannot be a part of a CCGT SCGT, but can be a part of a CCGT SCGT, but can be a part of a CCGT SCGT, but can be a part of a CCGT Valid for all thermal PP except nuclear plant. It can be used for representing gas turbines CCPP CCGT Together Together Separated Separated Together Together Separated Two main algebraic equations to Calculate T X and TRQ D & simple time delays Three main algebraic equations to Calculate T X, TRQ D, and W X & simple time delays Detailed thermodyna mic treatment Five main algebraic equations to calculate W A, CPR, T X, P M, and P MX &only one simple time constant No thermodynami c treatment, but simple linear transfer functions representation Simple block of second order transfer function Detailed thermodynami c treatment IGV Modeling Not exist Exist Exist Exist Not exist Not exist Exist Parameters Units in o F in o F in K in K, and CDP in bar pu values pu values values. (Temp. calculated first in K & o C, but finally expressed in pu values) CDP Parameter Not exist Not exist Not exist Exist Not exist Not exist Not exist Acceleration Control Loop Exist Exist Not exist Exist Exist Exist Not exist TIT Parameter Not exist Not exist Exist Not exist Not exist Not exist Exist W X Parameter Not exist Exist Exist Exist Not exist Not exist Exist 976