e Nucleus
Th The Nucleus 57, No. 4 (2020) 112-117
The Nucleus
ISSN 0029-5698 (Print)
www.thenucleuspak.org.pk
ISSN 2306-6539 (Online)
Pa
ki sta n
Modeling of an Extraction Steam Turbine and Speed Control System Design
H.M. Abdullah1*, G. Mustafa2, K.S. Chaudri3 and M. Abid2
1
Management Information System Division, PINSTECH, Islamabad
2
Department of Electrical Engineering, Pakistan Institute of Engineering & Applied Sciences (PIEAS) Islamabad, Pakistan
3
Department of Nuclear Engineering, Pakistan Institute of Engineering & Applied Sciences (PIEAS) Islamabad, Pakistan
ARTICLE INFO ABSTRACT
Article history: Steam-driven power plants essentially convert mechanical energy into electrical energy by using steam
Received: 29 January, 2020 turbines. It is imperative to control the speed of the turbines as the frequency of the power system depends
Accepted: 19 February, 2021 on it. This paper presents a model of a steam turbine containing three steam extractions from the
Published: 28 April, 2021 intermediate-pressure section and four extractions from the low- pressure section. The underlying
methodology for modeling is the continuity equation of a steam vessel. As the input variables, the model
uses the valve opening degree of different valves, namely high-pressure valve, reheater valve,
Keywords: intermediate-pressure steam extraction valve and low-pressure steam extraction valve. The model
Steam turbine behavior is observed against each input variable. It is then subsequently used to design the speed control
Modeling system of the steam turbine using the proportional and proportional-integral controllers. The response
Speed control of the speed control system is analyzed for both types of controllers and different valve openings.
Steam extractions Simulation results demonstrate that the proposed model is suitable to study the dynamic behavior of an
extraction steam turbine and for the feedback control system design.
1. Introduction control systems [3-5]. Most of these models are developed for
Power plants using steam as the working fluid are major non-extraction steam turbines and may not be suitable to
source of electricity. Over 80% of global electric power is represent steam turbine systems where extractions are
being generated by steam-driven power plants. In steam- employed from the turbine sections for feed-water (the water
driven power plants, steam turbines are used as prime movers to be supplied to boiler from a tank or condenser for
for electrical generators which then generate electrical energy conversion into steam) heating or other heating purposes.
[1]. Steam turbines are not only used in the power sector but Preheating the feed-water not only increases the efficiency of
also for industrial purposes. As the steam is allowed to expand the plant but also improves the power plant cost economics.
through different blade stages of the turbine, the heat energy Complex turbines with multiple controlled and/or
of superheated steam is converted into rotational mechanical uncontrolled extractions are also popularly used in the process
energy. As a result, the turbine’s rotor spins at a specific industry and cogeneration plants (plants which are used to
angular speed, which in turn drives the generator for the generate electricity and useful heat at the same time) to
generation of electricity. It is, thus, crucial from the safety provide steam of different temperature levels [6].
and performance perspective of the power plant that the speed Some nonlinear models of steam turbines are also
of the turbine must be monitored and controlled to avoid any developed based on the energy balance, thermodynamic
catastrophic situation and to improve the economics of the principles and semi-empirical equations [3-5, 7, 8].
power plant. Kulkowski et al. [9] presented simplified and detailed
For the model-based speed control system design of a nonlinear models for Nuclear Power Plant (NPP) steam
turbine, we need its mathematical model which can depict the turbines, which included a static model and a dynamic model
torque characteristics of the actual steam turbine with but without steam extractions. A non-extraction steam turbine
reasonable accuracy. Furthermore, transient operational model using hybrid-thermodynamic method and neural
behavior of steam turbines can be studied through simulations network approach has been presented by Dettori et al. [10]
to face the challenges concerning efficiency, commissioning and Lu and Hogg [11] for online monitoring applications.
time, start-up time, operation, availability, safety, cost- In this paper, a steam turbine model is presented exploiting
effectiveness, etc. [2]. For this purpose, we can make use of a hybrid modeling approach to incorporate the effect of steam
the steam vessel transfer functions because the steam turbine extractions in the turbine model, developed using the
is essentially a steam vessel where steam enters from one continuity equation. The mass continuity equation, as
point and exits from another and expands in between them as described previously [3], is employed to model the steam
reported previously [3]. turbine cylinders, whereas the case study presented by
A vast collection of steam turbine models is developed to Chaibakhsh and Ghaffari [7] is used for parametric estimation
study their behavior and to analyze the stability of the speed
*Corresponding author: abdullah-24@live.com
112
Fig. 1: Steam turbine configuration and steam conditions [6].
to include the steam extractions in analytical model of overall The high pressure superheated steam acts as the working
steam turbine system. The extractions from different turbine fluid and is responsible for energy flow. The superheated
sections are modeled separately downstream their respective steam enters the HP turbine section at 530 °C and 18.1 MPa
steam turbine sections. In a steam turbine system, multiple pressure, where it expands between turbine blade stages and
valves are employed to extract and control steam flow. Using energy conversion takes place. At the full load, the output
this model, the effects of different valve openings are temperature and pressure of steam from HP section are 350.3
observed on the torque and speed characteristics. The °C and 5.37 MPa, respectively.
developed model is then used to design a proportional (P) The discharged steam is passed through moisture
controller and a proportional-integral (PI) controller for the separators to remove moist content. The cold steam is, then,
speed control of steam turbine. The response of the closed- sent to reheater where it is reheated to a temperature of 530
loop speed control system is studied for different demands of °C, at 4.83 MPa and is subsequently fed to IP turbine section.
the steam. Primarily, speed control is achieved by The exhaust steam from IP turbine is further expanded in LP
manipulating the High Pressure Valve (HPV) of the steam section, whereas the input temperature and pressure of steam
turbine, which controls the steam flow entering the High for LP turbine section are 289.7 °C and 0.83 MPa,
Pressure (HP) section of the steam turbine, and ReHeater respectively.
Valve (RHV) that controls the flow of steam entering the
Intermediate-Pressure (IP) and Low-Pressure (LP) turbine
sections. A re-heater stores a large amount of steam; therefore,
the HPV control alone is not enough to limit the over speed.
The over-speed control involves fast control of the HPV and 𝑄𝑖𝑛 V: Volume of the vessel 𝑄𝑜𝑢𝑡
RHV because the RHV controls about 60% to 80% of the total
power by controlling the steam flow to IP and LP sections of
the steam turbine [5]. Simulation results depict the impact of
steam extractions on the digital control algorithms which is
identified as a research direction in future study.
2. System Description
Model simulations are an important tool in dynamic power
systems. Most advanced control methods are based on process 𝑄𝑒𝑥𝑡2 𝑄𝑒𝑥𝑡3
𝑄𝑒𝑥𝑡1
models [12]. This section describes an extraction steam
turbine system for which a model is developed in the Fig. 2: A steam vessel with three steam extractions.
subsequent section.
2.1 Steam turbine system 2.2 Assumptions
Fig. 1 shows a typical steam turbine system representing The following assumptions are made:
the turbine configuration and thermodynamic steam 1. Extractions are treated as steam vessels.
properties at input/output and steam extractions. It represents 2. Extraction flow rates are taken as a function of the inlet
a steam turbine of a 440 MW power plant with a once-through
mass flow rate.
Benson type boiler. It comprises of HP, IP, LP sections and
also includes steam extractions, feed-water heaters, moisture 3. Controlled extractions are considered, i.e., steam is
separators and the related actuators. extracted through extraction valves. Furthermore,
113
H.M. Abdullah et al. / The Nucleus 57, No. 4 (2020)
extractions of a particular turbine cylinder are controlled 3.3 Modeling of IP section with three steam extractions
by its common extraction valve, e.g., the IP Extraction The IP section has three extractions which are considered
Valve (IPEV) and the LP extraction Valve (LPEV) as to be controlled by a single common IP extraction valve as
depicted in Fig. 2. depicted in Fig. 2.
4. Leaks from valve stems and glands are not considered. Let W denotes the weight, V denotes the volume, ρ
denotes the density, Qext1 denotes the steam flow rate from
The steady-state flow rates, time constants and flow rate extraction 1, Qext2 denotes the steam flow rate from extraction
transfer functions are listed in Table1, where the symbol Qin 2, Qext3 denotes the steam flow rate from extraction 3 and Qout
denotes the input steam flow rate. denotes the output steam flow rate of IP turbine section. Then,
3. Turbine Model Development the continuity equation for this particular steam vessel is as
follows:
Different section of turbine system as shown in Fig. 1 are
modeled in this section. As can be noticed that the HP section dW dρ
has no steam extractions. Whereas, IP and LP sections have = V. = Qin (t) − Qext1 (t) − Qext2 (t) − Qext3 (t)
dt dt
three and four steam extractions from them, respectively. The − Qout,IP (t)
steam extractions parameters and corresponding transfer
functions are listed in Table 1. or by using the chain rule,
Table 1: Steam extractions parameters and their transfer functions [7]. ∂ρ dP
V = Q in (t) − Qext1 (t) − Qext2 (t) − Qext3 (t)
Turbine Extraction Time Steady state flow rate Flow rate trans-
∂P dt
section constant through extractions fer function
− Qout,IP (t)
no.
0.0578 This continuity equation can also be written as:
1 0.3 5.78% of Qin
0.3s + 1 dQout,IP
0.0814 Tv = Qin (t) − Qext1 (t) − Qext2 (t) − Qext3 (t)
2 0.7 8.14% of Qin dt
IP 0.7s + 1 − Qout,IP (t)
0.0622 𝜕𝜌 𝑃0
3 1.1 6.22% of Qin Where, 𝑇𝑣 = 𝑉. . denotes the time constant for the
1.1s + 1 𝜕𝑃 𝑄0
0.0335 steam vessel, 𝑃0 denotes the rated pressure of the vessel, and
4 1.5 3.35% of Qin
1.5s + 1 𝑄0 denotes the rated flow of the vessel. Taking the Laplace
0.04 transform and rearranging, the transfer function of the IP
5 1.7 4.00% of Qin turbine with extractions can be written as:
1.7s + 1
LP 0.0511 Qout,IP (s) 1 𝑘1 𝑘2 𝑘3
6 1.9 5.11% of Qin = − − −
1.9s + 1 Qin (s) 𝑇𝑣 s + 1 𝑇𝑣1 s + 1 𝑇𝑣2 s + 1 𝑇𝑣3 s + 1
0.6736
7 2.1 67.36% of Qin
2.1s + 1 1
Torque HPV RHV IPEV LPEV
Output torque and valves positions
3.1 Modeling of HP section
0.8
The HP section of the steam turbine system does not
include any steam extraction. Therefore, it can be modeled 0.6
using the conventional way of steam vessel approach as
reported in previous studies [1, 3-5]. 0.4
Q out,HP (s) 1
= 0.2
Qin (s) 0.25s + 1
Whereas the power fraction for HP turbine is taken as: 0
0 50 100 150 200 250 300
Time (sec)
FHP = 0.3.
Fig. 3: Torque characteristics of turbine for different valve openings (%).
3.2 Modeling of the reheater
For a tandem-compound single-reheat turbine, the exhaust The time constant Tv for the IP turbine is negligible as the
steam from HP section enters the reheater. A reheater has steam exits from reheaters with high pressure, therefore we
large mass storage and thermal capacity. It can be modeled as can take Tv = 0. Putting the expressions for Qext1, Qext2 and
outlined previously [1, 3-5]. Qext3 from Table1, we can get the transfer function of the IP
turbine as:
Q out,RH (s) 1
= Qout,IP (s) 0.0578 0.0814 0.0622
Qin (s) 7.5s + 1 =1− − −
Qin (s) 0.3s + 1 0.7s + 1 1.1s + 1
𝐹𝐼𝑃 = 0.4
114
3.4 Modeling of LP section with four steam extractions opening of the IP extraction valve, about 14% of the total flow
For modeling the low-pressure turbine section, the same rate is extracted by the IP turbine section. Whereas, a 40%
assumptions have been made. Applying the continuity opening of the LP extraction steam valve implies that about
equation on LP turbine and using flow rate transfer function 32% of Qin is extracted by the LP turbine section.
values for extractions from Table 1 and following the 4. Speed Control System Design for Steam Turbine
modeling procedure similar to IP turbine we can drive the
transfer function model for the LP Turbine section as follow: For safe and reliable operation of steam turbines, multiple
protection and control systems are employed, collectively
Qout,LP (s) 1 0.0335 0.04 0.0511 termed as Turbine Supervisory Instrumentation (TSI) [13].
= − − 0−
Qin (s) 0.4s + 1 1.5s + 1 1.7s + 1 1.9s + 1 TSI detects and measures the deviation from operating
0.6736 conditions and malfunctions. Whereas, control systems are
− designed to control various parameters of the steam turbine
2.1s + 1
for its safe operation. The model of a system used for control
3.5 Overall turbine system model
purposes needs to be as simple as possible, as its simplicity
The overall turbine model comprises of the HP, IP, and LP will ensure small computational complexity [14-16].
turbine sections. It also includes a model of the reheater valve.
The sum of the power fractions of the various turbine sections Speed is an important parameter of a steam turbine to be
is given as follow [3-5]. controlled which is primarily controlled by the governor
valves. Various control mechanisms are in use to drive the
𝐹𝐻𝑃 + 𝐹𝐼𝑃 + 𝐹𝐿𝑃 = 1 governor valve. We have used the mechanical-hydraulic
The inputs to the model are valve positions of different mechanism as it is a common mechanism used for speed
valves. The output of the model is the mechanical power. control. Speed relays and servomotors are taken as a mean of
speed governor system. A block diagram for the speed control
3.6 Torque characteristics of turbine model system, driving mechanism and turbine system is shown in
Torque characteristics of extraction steam turbine are Fig. 4. This control system is subjected to different types of
shown in Fig. 3, when different steam valves are opened at control techniques and load demands to analyze the control
different times. The output torque and different valve open system performance. Fig. 5 to Fig. 9 show speed
ings are shown in percentages. When HPV is opened to 80% characteristics for different valve openings. Speed and valve
at T = 10 seconds, the torque starts to develop in the HP steam positions are taken in percentage, i.e., 0 represents no speed
turbine and attains a steady value of 24%. As only HPV is (or a fully closed valve) whereas 1 indicates full speed (or a
opened and RHV is kept closed, so for this particular fully opened valve).
condition no steam is passed through the RHV. Upon opening 4.1 Speed control with a proportional control
RHV to 60% at T = 80 seconds, the torque increases further
and settles at the steady-state value of about 57%. For the proportional control (P-controller), the
proportional gain is tuned to 5.5. Fig. 5 shows the speed
When the IP extraction and the LP extraction valves are variations when the speed changer position is changed. In this
opened at T = 150 seconds and T = 230 seconds, respectively, scenario RHV, IP and LP extraction valves are kept closed;
the torque is reduced. From Fig. 1, it can be observed that however, speed reference is changed. The speed reference is
when IP and LP extraction valves are opened, steam will be changed twice, first at T = 0 second and then at T = 70
extracted from turbine cylinders. This will result in the seconds. The control system successfully attains the desired
decrease of mechanical power developed in the shaft of the speed and no overshoot is observed.
steam turbine. Also, from Table 1, it can be noted that for 70%
Fig. 4: Block diagram of speed control system of steam turbine.
115
H.M. Abdullah et al. / The Nucleus 57, No. 4 (2020)
1 1
0.8 0.8
Turbine speed (%)
Turbine speed (%)
0.6 0.6
0.4 0.4
0.2 0.2
Reference speed Actual speed Reference speed Actual speed
0 0
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
Time (sec) Time (sec)
Fig. 5: Turbine speed response with a P-controller to a step change in Fig. 7: Turbine speed response to a step change in reference speed with a PI-
reference speed. controller.
1.2 1.2
Reference speed Actual speed RHV IPEV & LPEV Reference speed Actual speed RHV IPEV & LPEV
1 1
Turbine speed (%)
Turbine speed (%)
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
Time (sec) Time (sec)
Fig. 6: Turbine speed response with a P-controller to a step change in Fig. 8: Turbine speed response to a step change in reference speed and
reference speed and different valves openings. different valves opening with a PI-controller.
Fig. 6 shows the speed characteristics of proportional correspondingly. The steady-state error is observed to be 0 for
control when speed reference, RHV, and extraction valves are PI-controller.
opened at different time instants. In this case, when the speed
Fig. 9 shows the performance comparison of P and PI-
reference is set to 100% at T = 10 seconds, the turbine rotor
controllers. At the instant when RHV is opened, both
speeds up with no overshoot. However, when RHV is opened
controllers exhibit an overshoot of about 20%. The steady-
at T = 130 seconds, the control system exhibits an overshoot
state error is zero for PI-controller, but P-controller shows a
of about 20%. P-controller compensates for this overshoot by
steady-state error of 3.5%.
limiting the inlet steam flow rate. Furthermore, the opening of
1.2
extraction valves causes the speed to drop by 8%. The P- Reference P Control RHV IPEV & LPEV PI Control
controller compensates for this decrease in speed by opening 1
the HPV valve and consequently, the speed is maintained with
Turbine speed (%)
0.8
a steady-state error of about 3.5%.
0.6
4.2 Speed control with proportional and integral control
Simulation results for proportional-integral control (PI- 0.4
controller) are shown in Fig. 7 and Fig. 8. The PI-controller 0.2
parameters are tuned to be, KP = 5.5 and KI = 0.02.
0
Here, again the control system is subjected to different 0 100 200 300 400 500 600
Time (sec)
scenarios. Fig. 7 shows the response when the speed changer
position is changed. Fig. 8 shows the situation when different Fig. 9: Comparison of turbine speed responses with P and PI-controllers.
valves are opened at different time instants. When the speed
reference is changed at T = 10 seconds, the control system 5. Conclusions
achieves the desired speed. Then, the opening of RHV results Steam turbine modeling and speed control system design
in an overshoot of around 20% in the turbine’s speed. is difficult because it is part of a complex interconnected
Whereas, opening of IP and LP extraction valves to 60% system. Most practical steam turbines at generation plants
causes a drop of around 8%. The PI-controller compensates comprise of steam extractions to increase the overall plant
for this disturbance by manipulating the high pressure valve efficiency. The dynamic response of a steam turbine can be
116
related in terms of changes in steam valves opening (HPV [6] X. Luo, B. Zhang, Y. Chen and S.P. Mo, “Modeling and optimization
of a utility system containing multiple extractions steam turbines”,
position and RHV position) and also the steam extraction Energy, vol. 36, pp. 3501-3512, 2011.
valves openings. [7] A. Chaibakhsh and A. Ghaffari, “Steam turbine model”, Simul. Model.
Pract. Theor., vol. 16, pp. 1145–1162, 2008.
In this paper, based on the continuity equation for steam
vessel and steam turbine empirical relations, a model of an [8] H. Rusinowski and M. Plis, “Mathematical model of a steam turbine for
the thermal diagnostics system”, Proc. Int. Carpath. Ctrl. Conf. (ICCC),
extraction steam turbine comprising HP, IP and LP turbine Tatranska Lomnica, Slovakia, pp. 630-634, 2016.
sections is developed. This model is then used for the speed [9] K. Kulkowski, M. Grochowski, K. Duzinkiewicz and A. Kobylarz,
controller design for the steam turbine. A steam turbine’s “Nuclear power plant steam turbine: Modeling for model based control
speed response depends on several factors, e.g., RHV purposes”, Appl. Math. Model., vol. 48, pp. 491-515, 2017.
position, HPV position, speed reference change and [10] S. Dettori, V. Colla, G. Salerno and A. Signorini, “Steam turbine
models for monitoring purposes”, Energy Procedia, vol. 105, pp. 524-
extraction valves positions. However, good speed stability 529, 2017.
can be achieved by a suitable controller design. [11] S. Lu and B.W. Hogg, “Dynamic non-linear modeling of power plant
by physical principles and neural networks”, Int. J. Elect. Pow. Energy
References Syst., vol. 22, pp. 67-78, 2000.
[1] M. Bennauer, E.G. Egener, R. Schlehuber, H. Werthes and G. Zimmer, [12] E.J. Farmer and B.G. Liptak, Process Control and Optimization: Steam
“Automation and control of electric power generation and distribution turbine controls”, 4th ed., vol. 2, Washington D.C., CRC Press, pp. 2137-
system: Steam turbines”, Control Syst. Robot. Autom., vol. 18, pp. 261- 2151, 2006.
271, 2009.
[13] W. Aleite, K. Ara, J. Berggren, A. Faya and C.G. Forstner, “Modern
[2] G. Zimmer, “Modelling and simulation of steam turbine processes: Instrum. and Control for Nuclear Power Plants: A guidebook”,
Individual models for individual tasks”, Math. Comp. Model. Dyn. International Atomic Energy Agency (IAEA), Vienna, Technical
Syst., vol. 14, pp. 469-493, 2008. Reports Series, no. 387, 1999.
[3] R.T. Byerly, O. Aanstad, D.H. Berry, R.D. Dunlop, D.N. Ewart, B.M. [14] K. Kulkowski, A. Kobylarz, M. Grochowski and K. Duzinkiewicz,
Fox, L.H. Johnson and D.W. Tschappat, “Dynamic models for steam “Dynamic model of nuclear power plant steam turbine”, Archive
and hydro turbines in power system studies”, IEEE Trans. Pow. App. Control Sci., vol. 25, pp. 65-86, 2015.
Syst., vol. 6, pp. 1904-1915, 1973.
[15] B. Vahidi, M.R.B. Tavakoli and W. Gawlik, “Determining parameters
[4] M. Dulau and D. Bica, “Mathematical modelling and simulation of the of turbine’s model using heat balance data of steam power unit for
behaviour of the steam turbine”, Procedia Technol., vol. 12, pp. 723– educational purposes”, IEEE Trans. Pow. Syst., vol. 22, pp. 1547–1553,
729, 2014. 2007.
[5] M. Dulau and D. Bica, “Simulation of speed steam turbine control [16] R. Rajan, P.M. Salih and N.A. Kumar, “Speed controller design for
system”, Procedia Technol., vol. 12, pp. 716 – 722, 2014. steam turbine”, Int. J. Adv. Res. Electr. Electron. Instrum. Eng., vol. 2,
pp. 4400-4409, 2013.
117
