2013 IEEE Symposium on Computers & Informatics

State Space Modeling of a Reheater System in

Power Generation Plant

Noor Ainy Harish, Razidah Ismail Tahir Ahmad

Faculty of Computer and Mathematical Sciences Faculty of Science
Universiti Teknologi MARA Universiti Teknologi Malaysia,
Shah Alam, Malaysia 81310 Skudai, Johore, MALAYSIA
[email protected] [email protected]

Abstract— Generally, the main components of a power plant are to the complexities of the system. In fact, the most important
a gas turbine, a boiler and a steam turbine. Factors such as high advantage of the state space representation is that the system
pressure and temperatures affect the operational conditions and dynamic properties are condensed in the model [12]. The
efficiency of large complex systems such as the power generation structural properties such as the stability and the dual concept
plants. Hence, it is desirable that the power plant be properly of the controllability and observability of a system play an
controlled in order to achieve optimum performance. Thus, a important role in the design of a control system. A state space
state space model is proposed. The model is mathematically model is not unique since a different choice of state variables
convenient structure for control design and analysis of leads to a different representation of the same system [3].
multivariable dynamic system. In addition, this approach is
based on time domain analysis and synthesis using state Boiler system is a multi-connected system which consists
variables. This paper focuses on the development of a state space of subsystems such as furnace, superheater, reheater, riser and
model for a reheater system, which is one of the subsystems in the drum. Even though, the state space model of furnace and
boiler of a power plant. The structural properties of the system superheater were successfully developed [10, 13], it is a need
are also analyzed. Thus, the possibility of integrating to complete the modeling to whole subsystems. This is due to
uncertainties and applying other mathematical techniques in the fact that more properties could be explored in future. Thus,
modeling the system can be further explored. the objective of this paper is to develop a state space
representation of a reheater system. In order to meet this
Keywords— Fuzzy State Space Model; Dynamic system; Fuzzy objective, the main assumptions and the mathematical
State Space Algorithm; Parameters Estimation equations used are presented. This is followed by the
discussion on the structural properties of the reheater system, in
I. INTRODUCTION particular stability, controllability and observability.
One of the most important tasks for developing an effective
control system and simulation of a physical process is II. REHEATER SYSTEM IN A COMBINED CYCLE
modeling the system. In large complex system such as the POWER PLANT
power generation plant, factors such as high pressure and A. Process Description
temperatures affect production and environmental costs [1].
Hence, in order to ensure the smooth and continuous energy Reheater is one of the important subsystems in a boiler of
flow, it is desirable that the power plant be properly controlled combined cycle power plant, which produces steam [14]. Other
in order to achieve optimum performance [2]. Generally, the subsystems are furnace, superheater, riser and drum. The
main components of a power plant are a gas turbine, a boiler process starts when water enters the boiler and goes through a
and a steam turbine where reheater system is one of the stage of preheating in the economizer, exchanging energy with
important subsystems in a boiler of a combined cycle power the combustion gases. The evaporation process occurs in the
plant [3]. However, the number of each of these components riser and the water/steam mixture circulates though the risers
depends on the plant complexity. In modeling the dynamics of passing the combustion chamber. The separation of vapor and
a boiler, some studies [4-6] consider the distributed parameter water is taken in the drum. The steam from the drum is
which is described by partial differential equations and the extracted into the superheater section and from there enters the
solution may be complex. The state space approach is based on steam turbine. The reheater section is used to reheat the steam
time–domain analysis and synthesis using state variables. It is a between the two stages of the steam turbine. In combined cycle
unified method for modeling, analyzing and designing a wide power plant, boiler is a pure heat recovery steam generator.
range of systems [7]. This approach provides a good The exhaust gases are used to heat water in a boiler to raise
approximation in modeling engineering and biological systems, steam.
for example, [8-11]. In addition, the mathematical equations The main assumptions and equations to construct the
are represented in matrix algebra and first order differential mathematical model of a reheater system are based on the
equations. Therefore, only the size of matrix changes according following assumption [14]:

Universiti Teknologi MARA Malaysia and the Ministry of Higher Education

of Malaysia

978-1-4799-0210-1/13/$31.00 ©2013 IEEE 104

 2013 IEEE Symposium on Computers & Informatics

i. The model only includes time derivatives of variables. Reheat outlet mass flow:
ii. Polynomial fits to steam tables were used to establish the
d [wro ]
relations between steam parameters such as enthalpy, = (wri − wro ) / τ rh (7)
density, temperature, pressure. dt
iii. Superheated steam and furnace exhaust gases were treated Gas-tube heat balance:
as ideal gases.
iv. The model includes only the main parts of the boiler using d
Q rs = Q rh + M r C rh ( Trh ) (8)
lumped characteristics for those parts which consists of dt
more than one section. Heat transfer tube-steam:
v. The variables are assumed to satisfy basic physical Q rh = K rh w ri0 .8 ( T rh − T r ) (9)
thermodynamic balances.
Heat balance for steam:
Heat balance for a heat exchanger in the form d
d Q rh + w ri hri = w ro h ro + v r ( p rh hro ) (10)
Qin + win hin = wou hou + v ( ρou hou ) dt
dt (1)
Mass balance in the form: III. THE STATE SPACE MODEL OF REHEATER SYSTEM
d There are two equations involve in the formulation of the
win − wou = ( ρ ou V ) (2) state space model. There are state equation and output
dt
Friction losses (momentum balance): equation. The mathematical representation of state equation
G
2 dx K K
win can be expressed as = A x ( t ) + B u ( t ) where A is the
p in − p ou = γ (3) dt
ρ in state matrix of order p, B is the input matrix of order p × n .
K
dx K K K
Wd, Tv , ? wv The vector is the time derivatives of x( t ) . x (t ) and u (t )
hw, hv , hwv
dt
represent the state and input variables. The output equations
RISER
K K K
Trh
Qir
can be expressed in a linear form y (t ) = C x (t ) + D u (t ) , where
Tet
C is the output matrix and D is the direct transmission matrix.
Tst K K K
wF
wA
Ws
SUPERHEATER
x (t ) , u (t ) and y (t ) represent the state, input and output
Qgs
hG
wG
FURNACE
wa, ha
AND
ATTEMPORATOR variables respectively. In this study it is assumed that D = 0
Wv
? pe, Ts
which imply that there is no direct transmission between the
Qes K K
Pei
Qrs
DRUM input u (t ) and output y (t ) .
Wr, x
Tei ECONOMIZER Pri REHEATER Pro
Wei
Wri
Tri
Tr
Wro
We, he
L
A. Development of State Equation
Weo
In order to determine the state equation, first, T r steam
+
Wa
temperature of reheater is solved using equation (5),
Wd
hro − href + c prTref
Tr = (11)
Fig 1. Block diagram of a Boiler c pr

B. Mathematical Modeling Solving the time-dependent equation of reheater metal tube

The mathematical model of the reheater system is based temperature, Trh :
on the first principal physical and thermodynamic laws with From equation (8),
appropriate simplifications. d
Mass balance: [T rh ] = 1 [Q rs − Q rh ]
dt M r C rh
d Substituting equation (9) into the above equation gives:
wri − wro = vr (ρ rh ) (4)
d
dt [T rh ] = 1 [Q rs − k rh w ri T rh − k rh w ri T r ] (12)
dt M r C rh
Reheat steam temperature:
hro − href
d
[T rh ] = 1 [Q rs ] − k rh w ri [T rh ] + k rh w ri [T r ] (13)
Tr = + Tref (5) dt M r C rh M s C rh M s C rh
c pr
Substitute equation (11) into equation (13)
Reheat steam pressure:
p ro = Rr ρ rh Tr (6)

Universiti Teknologi MARA Malaysia and the Ministry of Higher Education

of Malaysia

105
 2013 IEEE Symposium on Computers & Informatics

B. Development of Output equation

d
[Trh ] = 1 [Q rs ] − k rh wri [Trh ] +
dt M r C rh M s C rh Initially, Pro ,the superheated steam pressure, is solved
(14)
k rh wri ª h ro − h ref + c pr Tref º by substituting equation (11) and ρ rh = xrh into equation (6) ,
« » hro
M s C rh «¬ c pr »¼
Pro = Rr ρ rhTr
Rr ª hro − href + c pr Tref º
Solving the time-dependent equation of x rh : Pro = « »[x rh ] (18)
By referring to the equation (10), hro «¬ c pr »¼
d
[ρ rh hro ] = Qrh + wri hri − wro hro
dt Vrh Vrh Vrh Referring to equation (9), simplify for Tr ,
Substituting equation (9) and x rh = ρ rh hro into the above Qrh (19)
Tr = Trh −
equation will leads to the following: K rh wri
Writing equation (18) and equation (19) in matrix-vector form,
d
[xrh ] = k rh wri [Trh ] − k rh wri [Tr ] + wri hri − wro [xrh ] the output equation is
dt Vrh Vrh Vrh Vrh ρ rh
§ Pro · §T ·
(15) ¨¨ ¸¸ = C ¨¨ rh ¸¸ + L (20)
© Tr ¹ © x rh ¹
Substitute equation (11) into equation (15) § ª hro − href + c pr Tref º·
¨ »¸
ªh − h + c T º where output matrix is C = ¨ 0 Rr «
d
[x rh ] = k rh wri [Trh ] − k rh wri « ro ref pr ref » + ¨ ¬« c pr ¸
¼» ¸
dt V rh V rh ¬« c pr »¼ ©1 0 ¹
wri hri
−
wro
[x rh ] and L = − Qrh
V rh V rh ρ rh K rh wri
(16)
Therefore, equations (14) and (16) can be expressed in matrix- Thus, state space model of the reheater system is given by
vector form as, the state equation, (17) and the output equation, (20) with two
state variables, two input and two output parameters. The state
§d · variables consists of Trh (metal tube temperature in K) and
¨ [Trh ]¸ §T · §Q ·
¨ dt ¸= A¨¨ rh ¸¸ + B¨¨ rs ¸¸ + J (17) xrh (The product of reheated steam density and the outlet
¨ d [x ]¸ © x rh ¹ © hri ¹ steam specific entalphy in J/m3 ) whereas the input parameters
¨ rh ¸
© dt ¹
are Qrs (heat supplied to the reheater from the furnace model
where the state matrix is in J/s) and h ri (specific enthalpy of inlet steam in J/kg ). The
§ k rh wri · output parameters that are involved in the development of state
¨− 0 ¸
M r Crh space model of a reheater system are Pro (outlet steam pressure
A=¨ ¸
and the input matrix is
¨ krh wri wro ¸ in Pa) and Tr (reheater steam temperature in K ).
¨ − ¸
© Vrh Vrh ρ rh ¹
IV. THE STATE SPACE ANALYSIS OF REHEATER
SYSTEM
§ 1 ·
¨ 0 ¸ K §T · The structural properties that formed the essential
M C
B = ¨ r rh ¸ , the state vector, x( t ) = ¨ rh ¸ ,
¨ wri ¸ ¨x ¸ requirements in the state space analysis of the reheater are
© rh ¹
¨ 0 ¸
Vrh ¹ stability, observability and controllability. Lyapunov’s First
©
Method [7] is used to evaluate the stability of the zero-input
K §Q · response for the state space model of the reheater system. It
the input vector u( t ) = ¨ rs ¸ and,
© hri ¹ involves an examination of eigenvalues, λ of the system-
§ k w ª hro − href + c pr Tref º · description matrix A. When the eigenvalues have negative
¨ rh ri « » ¸ real parts, the system is said to be asymptotically stable. By
¨ V rh « c »¼ ¸
J =¨ ¬ pr
¸ using the data published in [15], the eigenvalues of matrix A
¨ k rh wri ª hro − href + c pr Tref º ¸ are λ1 = −9.1637 × 10 −10 and λ2 = −2.6956 × 10 −1 . Since the
¨− « »¸
¨ Vrh « c pr ¸
© ¬ ¼» ¹ eigenvalues of matrix A are both negatives value, the reheater
system is asymptotically stable. Based on the Algebraic
Controllability Theorem [15], the property of controllability of
reheater system can be examined by using the system-
description matrix A and matrix B. The system is controllable
Universiti Teknologi MARA Malaysia and the Ministry of Higher Education
of Malaysia

106
 2013 IEEE Symposium on Computers & Informatics

if the rank, r( q ) of the controllability test matrix [1] M. M. P Gonzalez, F.J.F. Garcia, I.S.Ramon & H.S.Roces,

[ ]
Experimental Thermal Behavior of Power Plant Reheater, Energy. 31
q = B AB....A k −1B is equal to the order of the system. Since 2006: 665-676.
[2] T.M.I. Mahlia, M.Z. Abdulmuin, T.M.I. Alamsyah & D. Mukhlishien
r( q ) , the order of the system is equal to 2, the reheater (2003), Dynamic Modeling and Simulation of a Palm Wastes Boiler,
system is controllable. Similarly, based on the Algebraic Renewable Energy 2003: 28: 1235-1256.
[3] K. Ogata. Modern System Engineering. 3rd ed. Upper Saddle River:
Observability Theorem [15], the property of observability of Prentice-Hall International, 2002
the system involved matrix A and matrix C. The rank r( n ) of [4] I. Benyo, Kovacs, J. Kovacs, J. Mononen, & U. Kortela. “Modelling of
Steam Temperature Dynamics of a Superheater”, Proceedings 16th
the observability test matrix n = ªC T AT C T ...... ( AT ) p −1C T º European Simulation Symposium .6: 1473-8031, 2004.
¬« »¼
[5] S. Samyuktha & P. Kanagasabapathy , Optimized Fuzzy Modelling of a
is equal to 2, the order of the system. Hence, based on r( q ) Boiler Superheater. Journal Instrum. Soc. India 2002: 32(3): 157-164.
[6] W. Zima. (2001). Numerical modelling of dynamics of steam
and r( n ) the reheater system is controllable and observable. superheaters. Energy. 26(12): 1175-1184.
[7] N.S. Nise. Control Systems Engineering. 2nd ed. Menlo Park, CA:
V. CONCLUSION Addison-Wesley, 1995.
[8] W.L. Berendrecht, A, W. Heemink, F. C. Van Geer, & J.C. Behrels,
The development of state space model of a reheater system State Space Modeling of Water Tables Fluctuations In Swiching
of combined cycle power plant has been presented. The state Regimes, Journal of Hydrology 2004: 292: 249-26.
space model of the reheater system is then analyzed to obtain [9] W.L. Berendrecht, A, W. Heemink, F. C. Van Geer, & J.C. Behrels, A
considerable insight of the system behavior. It is observed that non-linear state space approach to model groundwater fluctuations,
based on the state space analysis, the reheater system is Journal of Hydrology. Advances in Water Resources 2006: 29: 959-973.
[10] R.Ismail. Furnace Modelling using State Space Representation,
asymptotically stable, controllable and observable. Since state Scientific Research Journal, 3(1), pp 37 – 52, 2006.
space representation is a mathematically convenient structure [11] I.D.J Onsen, R.A.M. Yers & J.M. Flemming (2003), Meta- Analysis Of
for control design and analysis of multivariable dynamic Animal Movement Using State-Space Models, Ecology 2003:84(11):
system, further study will also explore the possibility of 3055-3063.
integrating with other mathematical techniques such as [12] S.G. Cao & N.W. Rees , Identification of Dynamic Fuzzy System.
uncertainties in modeling the system. Fuzzy Sets and Systems 1996:74:307-320.
[13] N.A.Harish, R. Ismail and T. Ahmad. Modeling of Superheater system
A fuzzy state space approach was presented for solving the using a state space representation. Proc. of the International conference
on Science & Technology: application in industry & education. 11-13
inverse problem in reheater system modeling, whereby the
December 2008. Malaysia: Penang
manipulation of imprecise, uncertain quantities is considered. [14] A. Ordys, A.W. Pike, M. A. Johnson, R. M. Katebi, R.M. and M. J.
The uncertain value parameters of the system are represented Grimble. Modelling and Simulation of Power Generation Plants.
by fuzzy numbers with their membership function derived from London: Springer-Verlag, 1994.
expert knowledge. In many respects, fuzzy numbers depict the [15] D. saiez, A. Cipriano and A. W. Ordys, Optimization of Industrial
physical world more realistically than single-valued numbers, Process at Supervisory Level: Application to Control Thermal Power
Plants, Springer, 2002.
as the concept takes into account the fact that all phenomena [16] B. Friedland. Control System Design: An Introduction to State Space
have a degree of uncertainty. The ability of this method to Methods. New York: Academic Press,1986.
address inverse problems in multivariable systems directly is
an outstanding advantage especially in reducing computation
APPENDIX
time and cost.
wri - reheater inlet steam mass flow (kg/s)
To facilitate the implementation of this approach, an
wro - reheater outlet steam mass flow (kg/s)
interactive interface was developed. Using this computational
tool, the users can evaluate more alternatives in less time, and v rh - reheater volume (m 3 )
at the same time, the users can obtain more information on the
performance of each of those alternatives. Since fuzzy state ρ rh - reheated steam density (kg/m 3 )
space approach is designed for solving inverse problem in any Qrs - heat flow to the reheater (from the furnace model)(J/s)
multivariable dynamic system, further study is undertaken to Qrh - heat transferred to the steam(J/s)
explore the possibility of integrating with other mathematical Mr - mass of reheater tubes (kg)
techniques in order to gain a better understanding of the Trh - reheater metal tube temperature (K)
system.
Tr - reheater steam temperature (K)
ACKNOWLEDGMENT K rh - experimental heat transfer coefficient [J/(kg*K)]
The authors would like to thank Research Management hro - outlet steam specific enthalpy (J/kg)
Institute, Universiti Teknologi MARA Malaysia and the hri - inlet steam specific enthalpy (J/kg)
Ministry of Higher Education of Malaysia for the financial
href - reference steam enthalpy condition (J/kg)
support through the Fundamental Research Grant Scheme 600-
RMI/ST/FRGS 5/3 Fst (23/2008). Their gratitude is also Tref - ideal gas reference (K)
extended to anonymous reviewers for their valuable comments Pro - outlet steam pressure (Pa)
and suggestions. - specific heat of steam at constant pressure[J/(kg*K)]
C pr
REFERENCES Crh - heat capacitance of the tubes [J/(kg*K)]
Rr - ideal gas constant[-]
Universiti Teknologi MARA Malaysia and the Ministry of Higher Education τ rh - an empirical flow time constant [-]
of Malaysia

107