A Novel Classical Controller for Automatic Generation Control in Thermal and Hydrothermal Systems

J. Nanda, Fellow, IEEE, S. Mishra, Senior Member, IEEE , P. G. Mishra, and K.V. Sajith

Abstract This paper presents automatic generation control (AGC) of interconnected two equal area thermal system, two equal area hydrothermal System and five unequal area thermal system. Thermal plants have single reheat turbines with generation rate constraints of 3% per minute. The hydro plant in hydrothermal system has electric governor and has generation rate constraints of 270% per minute for raising generation and 360% per minute for lowering generation. A novel classical controller based on Integral and Double Derivative control is used in AGC and its performance is compared to several other classical controllers such as Integral (I), Proportional-Integral (PI), Integral-Derivative (ID) and Proportional-IntegralDerivative (PID) controllers. Integral squared error (I.S.E) and Bacteria Foraging (BF) techniques are used for optimization. Investigations reveal on comparison that Integral (I), Proportional-Integral (PI), IntegralDerivative (ID) or ProportionalIntegral-Derivative (PID) controllers all provide more or less same response where as Integral - Double Derivative (IDD) controller provides much better response. Further, the controller action can be optimally delayed to provide the best dynamic response. Index Terms Automatic generation control, bacterial foraging technique, integral plus double derivative controller, speed regulation parameter.

KIi =

Integral gain of integral controller and of integral double derivative controller in area i KDDi = Double Derivative gain of integral double derivative controller in area i Kd, Kp, Ki Electric governor derivative, proportional and integral gains respectively. i = (Di + 1/Ri) (i.e. Frequency response characteristics of area i); = Cost = ( fi ) + ( Ptie i-j ) dt
2 2 0 T

T = fi = Pgi = Ptie i-j = Tij = p = Tw =

Simulation time (s) Incremental change in frequency of area i (Hz) Incremental generation of area i (p.u) Incremental change in tie power of tie i-j (p.u) Synchronizing coefficients. Number of parameters to be optimized Water starting time.
I.

INTRODUCTION

NOMENCLATURE i Subscript referred to area i (1, 2, 3,4,5). * Superscript denotes optimum value Pri = Rated power of area i (MW) Hi = inertia constant of area i (s) PDi = Incremental load change in area i (p.u) Di = PDi / fi (p.u/Hz) Pgi = Incremental generation change in area i (p.u) Governor Speed regulation parameter of area i. Ri = (Hz/puMW) Tgi = Steam governor time constant of area i (s) Kri = Steam turbine reheat coefficient of area i Tri = Steam turbine reheat time constant of area i (s) Tti = Steam turbine time constant of area i (s) Bi = Frequency bias of area i f = Nominal system frequency (Hz) Tpi = 2Hi / f Di (s) Kpi = 1/Di (Hz/pu)

J. Nanada, S. Mishra, P. G. Mishra and K.V .Sajith are with Department of Electrical Engineering, Indian Institute of Technology Delhi, India (esukumar@ee.iitd.ac.in and mail: janardannanda@yahoo.co.in, pg_mishra@rediffmail.com)

a large interconnected power system, generation comprises of thermal, hydro, nuclear and gas power generation. Gas power generation is best when used for meeting varying load demand and peak demands. However, as these plants form a very small percentage of total system generation, they do not play very significant role in Automatic Generation Control of a large power system. Owing to their high efficiency, Nuclear units are usually kept at base load, with no participation in AGC. Thus the natural choice for AGC falls on either thermal or hydro units. To achieve the integrated operation of a power system, an electric energy system must be maintained at a desired operating level which is delineated by nominal frequency, voltage profile and load flow configuration. The desired operating condition can be achieved by controlling the real and reactive powers generated through the controllable sources of the system. Automatic generation control (AGC) plays an important role in the power system by maintaining scheduled tie power and scheduled system frequency in normal operation and during small perturbation. Most research in the area of AGC pertains to interconnected two equal area thermal systems and not much attention has been paid to AGC of unequal multi area systems. Substantial work in the past have been done to design the governor secondary controllers, and design of governor primary control loop (i.e. selection of suitable governor droop or speed regulation parameter R)

IN

has somehow not been given enough attention. It is known that without the secondary control, smaller the value of governor speed regulation parameter, smaller is the steady state error in frequency. In the presence of supplementary control, there is nothing to be sacrosanct to use a small governor droop (of the order of 4% used in practice) as any large but credible value of R can also guarantee zero steady state error in frequency. Few works [1], [2], [3] have been reported in the past for selection of speed regulation parameters. However, Nanda, Mishra and Saikia [4] have proposed a more elaborate and comprehensive approach for finding optimum value of R for a multi-area system using bacterial foraging (BF) technique. Several classical controllers such as integral (I), Proportional-Integral (PI), Integral-Derivative (ID) and Proportional-Integral-Derivative (PID) have been used in AGC as secondary controllers, but surprisingly there is hardly any literature that compares performances of these controllers on some AGC model to establish the best of this lot. A comprehensive comparison of the above said classical controllers have been made and whether any other new classical controller can well compete and even provide better performance than these controllers is explored in this paper. The main investigations of the present work are: A new classical controller based on Integral plus doublederivative (IDD) is proposed and its performance is compared with other classical controllers such as Integral (I), Proportional-Integral (PI), Integral-Derivative (ID), and Proportional-Integral-Derivative (PID) controllers. Bacterial foraging technique is used for simultaneous optimization of primary and secondary controller parameters for different types of classical controllers and their dynamic performances are compared. Performance of the system with delayed controller action is studied. II. BACTERIAL FORAGING OPTIMIZATION TECHNIQUE In [5-8] authors have tuned the controller gains without GRC whereas in [9-10] they have attempted the same with GRC. Besides, GA [11-12] and Particle Swarm Optimization [13] have already been applied to optimize the controller gains in AGC. A recent and more competent evolutionary computation technique called Bacterial Foraging scheme has been proposed by Passino [14] and has been applied to different problems in [4, 15 and 16]. In this technique the number of parameters that are used for searching the total solution space is much higher compared to those in GA. Here, the foraging behavior of E-Coli bacteria present in our intestine is mimicked. The decision making system of these bacteria that guides how foraging should proceed can be represented by the activities namely Chemotaxis, Swarming, Reproduction and Elimination and Dispersal.

In particular
i ( j + 1, k , l ) = i ( j , k , l ) + C ( i ) ( j )

(1)

where i ( j , k , l ) represents the ith bacterium at jth chemotactic, kth reproductive and lth elimination and dispersal step. C (i ) is the size of the step taken in the random direction specified by the tumble. B) Swarming: Bacterium which is in the optimal location provides some attractant signal so that all the bacteria swarm together to reach the optimum location. The mathematical representation for swarming can be represented by
J cc ( , P ( j , k , l )) = = +
S i =1 S i =1 S i =1 i J cc , i ( j , k , l ) p

)
(2)

d attract exp attract h repelent exp repelent

m =1 p

i ( m m ) 2

m =1

i ( m m ) 2

where J cc (, P( j , k , l )) is the cost function value to be added to the actual cost function to be minimized to present a time varying cost function. 'S' is the total number of bacteria and 'p' the number of parameters to be optimized which are present in each bacterium. d attract , attract , hrepelent , repelent are different

coefficients that are to be chosen properly C) Reproduction: Fitness of each bacterium is calculated during chemotaxis process. In reproduction the least healthy bacteria die and the other healthiest bacteria each split into two bacteria, so that the total number of bacteria in the population remains constant, which are placed in the same location. D) Elimination and Dispersal: While chemotaxis provides a basis for local search and reproduction accelerates the convergence, these two processes are not enough for global optima searching. To introduce diversity in the algorithm so that bacteria may not get stuck around the initial position or local optima, it is necessary to change the position of bacteria either gradually or suddenly. The dispersion event happens after a certain number of reproduction processes. It is chosen according to certain probability Ped to each bacterium. If it is decided to be dispersed, it will be moved to another position within the environment for the calculation of next generation. These events may prevent the local optima trapping effectively, but unexpectedly disturb the optimization process. In BF technique we assign each bacterium with a set of variables to be optimized and are assigned with random values () within the solution space, defined through upper and lower limits between which the optimum value is likely to fall. In our case they are KIi, Ri, and Bi (i=1, 2, 3, 4, 5). Each Bacterium is allowed to take all possible values within the range and the objective function which is I.S.E defined by (3) is minimized.
J = {(f i ) 2 + (Ptiei-j ) 2 }dt
0 T

(3)

A) Chemotaxis: This process is achieved through swimming where T is the simulation time. In this paper, the BF algorithm and tumbling via Flagella. Based on the rotation of Flagella in reported in [4, 16] has been adapted. each bacterium, it decides whether it should move in a predefined direction (swimming) or altogether in different III. SYSTEMS INVESTIGATED directions (tumbling), in the entire lifetime. A tumble can be The systems on which investigations have been carried out are represented by a unit length random direction, say ( j ) , which two equal area system of area1: 2000MW; area2: 2000MW, is used to define the direction of movement after a tumble. and a five unequal area system of area1: 2000MW, area2:

4000MW, area3: 8000MW, area4: 10000MW and area5: 12000MW. The thermal systems are provided with single reheat turbine and generation rate constraints of 3%/minute in each area and hydro systems are provided with generation rate constraints of 270%/min for raising generation and 360%/minute for lowering generation. Integral (I), Proportional-Integral (PI), Integral-Derivative (ID), ProportionalIntegralDerivative (PID) and Integral-Double Derivative (IDD) controllers are considered for investigation. The transfer function of IDD controller is given by (4). K (4) T.FIDD = I + K DD .s 2 s

The nominal system parameters for thermal systems are taken from [17] and for hydro systems are taken from [3] and shown in the Appendix. On their respective bases per unit values of different parameters of the unequal areas are considered to be same. Hence, while modeling interconnected areas of different capacities, a parameter a12 = - Pr1/Pr2=-1 is considered in two equal area system. Similarly in five area system the quantities a12 = - Pr1/Pr2, a13 = - Pr1/Pr3, a14 = - Pr1/Pr4, a15 = - Pr1/Pr5, a23 = Pr2/Pr3, a34 = - Pr3/Pr4 , a45 = - Pr4/Pr5, a25 = - Pr2/Pr5, a24 = Pr2/Pr4, a35 = - Pr3/Pr5 are considered where each area is connected to all other areas. The transfer function model of a five area systems with integral controller is shown in Fig.1. MATLAB version 7.01 has been used to obtain dynamic responses for a SLP (step load perturbation) of 1% in area1.

Fig.1 Transfer Function Block Diagram of Five-area Thermal System.

IV. RESULTS AND ANALYSIS A. Two Area Thermal System In a two equal area thermal system speed regulation parameters of R1 = R2 = 4 % (2.4 Hz/pu) and frequency bias parameters of B1 = B2 = = 0.425 are taken. Different types of controllers such as Integral (I), Proportional-Integral (PI), Integral-Derivative (ID), Proportional-Integral-Derivative (PID) and Integral-Double Derivative (IDD) are examined and their gains are optimized by I.S.E approach. Fig.2 shows the dynamic responses for f1 = f(t) for 1 % step load perturbation in area1. It is clearly seen that the responses for I, PI, ID and PID controllers are practically same whereas the response for IDD controller is better. Hence further studies are carried out considering only Integral and Integral-Double Derivative controllers. B. Two Area Hydrothermal systems In a two equal area hydrothermal system speed regulation parameters of R1=R2=4%(2.4 Hz/pu) and frequency bias parameters of B1=B2==0.425 are chosen. Optimum values of Integral and Integral Double Derivative controllers are evaluated by I.S.E criterion using classical approach for both the thermal and hydro areas. Dynamic responses with optimum Integral controllers for both the areas and with optimum IDD controllers for both the areas are shown in Fig.3. It is seen that the responses with IDD controller are much better than the response with Integral Controller. Further, it is observed that the effectiveness of IDD controller as compared to Integral controller in improving dynamic response is much more pronounced in hydrothermal system than in a thermal-thermal system.

Fig.2. Comparison of frequency deviation in area 1as a function of time for different classical controller in two equal area thermal systems.

Fig.3(a)
Fig.4 (c)

Fig.3(b)
Fig.3. Comparison of Integral and IDD controller performance in two area Hydrothermal system. (a) frequency deviation in area1 vs. time (b) Deviations in tie line power connecting area1 and area2 vs. time Fig.4 (d)

C. Five- areas system Dynamic responses of I and IDD controllers for five unequal areas system with 1% SLP in area1 is obtained with values of R parameters and controller gains simultaneously optimized by B.F technique. Fig.4 (a, b, c, d, e) shows the dynamic responses of the controllers.

Fig.4 (e) Fig.4 Comparison of Integral and IDD controller performance in five unequal area system (a) frequency deviation in area1 vs. time (b) frequency deviation in area2 vs. time (c) Deviation in tie line power connecting area1 and area2 vs. time (d) Deviations in tie line power connecting between area1 and area3 vs. time(e) Deviations in tine line power connecting between area1 and area4 vs. time

Fig.4 (a)

Fig.4 (b)

The optimum values of gain and R for I controller are KI1*= 0.0587, KI2*= 0.0928, KI3*=0.033, KI4*= 0.0206, KI5*= 0.0365, R1*= 9.7571 (16%), R2*= 7.169 (12%), R3*= 6.4331 (11%), R4*=7.8257 (13%), R5*= 4.6648 (8%) and for the IDD controller are KI1*= 0.094916, KI2*= 0.023028, KI3*= 0.060551, KI4*= 0.048728, KI5*= 0.089354, KDD1*= 0.0070152, KDD2*= 0.0043397, KDD3*= 0.0039915, KDD4*= 0.0080857, KDD5*= 0.0046806, R1*= 6.154(10%), R2*= 7.9191(13), R3*= 9.218(15%), R4*=7.3819(13%) and R5*= 4.625(8%). Examination of dynamic responses in Fig.4 clearly shows that Integral-Double Derivative controller gives much better

5 response than the integral controller. It is to be appreciated that for the IDD controller fifteen parameters could be optimized simultaneously very efficaciously and cogently by the powerful Bacterial Foraging technique. It may also be observed from the optimum values of R* for the 5 areas that all the areas have much larger values of governor droop (R) than is used in practice (R of the order of 4%). Higher values of R* for the areas are to be preferred for cheaper realization of governors [5], [9]. D. Switching of Controller Fig.5 shows the dynamic responses with Integral Controller (throughout the operation) and delayed operation of Integral controller through switching (10 secs, 20 secs and 30 secs after the occurrence of the step load perturbation). It is seen from the responses that the response with 10 secs delay may be considered to provide better response than responses with other delays from the view points of settling time and peak deviation. However, the optimum delay time of the controller is to be investigated. It will be of interest to investigate up to what time the controller action can be delayed, so as to provide nearly the best response from the view points of peak deviation and settling time so that we can reduce the wear and tear on the controller.

Fig.6(a)

Fig.6(b) Fig.6. Comparison of dynamic responses with IDD controller being switched in at different instances. V. CONCLUSION The performances for integral (I), proportional-integral (PI), integral-derivative (ID), and proportional-integral-derivative (PID) controllers in AGC are practically the same from the viewpoint of dynamic responses. However, the proposed new controller based on Integral and Double Derivative (IDD) provides distinctly better response than the Integral Controller. Simultaneous optimization of speed regulation parameters (R) and controller gains in a five unequal area thermal system using powerful BF technique explores use of much higher (to the tune of even 3 to 4 times) values of R for the areas than the value of R (of the order of 4%) used in practice. Higher values of R are preferred for easier and cheaper realization of governors. Investigations reveal that the controller action can be optimally delayed to provide best response. Delayed controller action reduces wear and tear of the controller and increases its life. APPENDIX Nominal parameters of the system investigated f=60Hz, Tgi=0.08 s; Tri = 10 s; Hi = 5 s; Tti = 0.3 s; Kr = 0.5; Ptie,max = 200MW; Di = 0.00833 p.u.MW/Hz; Tpi=20s, Kpi=120Hz/p.uMW; loading = 50%, Tij=0.544; Tw= 1.0 s; Kp= 1.0; Kd= 4.0; Ki= 5.0.

Fig.5(a)

Fig.5(b) Fig.5. Comparison of dynamic responses with integral controller being switched in at different instances. Fig.6. shows the responses with IDD and delayed IDD (delayed by 10 secs, 20 secs, 30 secs). In the present case it is seen that a 10 secs delay of the IDD controller from the instant of perturbation provides more or less the best dynamic response. However, the optimum delay time for the controller is to be evaluated.

6 REFERENCES
[1] [2] J Nanda and B.L.Kaul, Automatic generation control of an interconnected power system, IEE proceedings vol. 125, No.5, pp385390, May 1978. L. Hari, M.L. Kothari, J. Nanda, Optimum selection of speed regulation parameters for automatic generation control in discrete mode considering generation rate constraints, IEE Proc. C, vol.138, No.5, Sep 1991. J. Nanda, A. Mangla,and S. Suri Some new findings on automatic generation control of an Interconnected Hydrothermal system with conventional controllers , IEEE transaction on energy conversion Volume 21, No1, March 2006 Page(s):187-194. J. Nanda, S. Mishra, Lalit Chandra Saikia, Maiden application of bacterial foraging based optimization technique in multiarea automatic generation control, IEEE transaction on power systems ,Vol.24 ,No.2 , pp 602-609, May 2009 C. Concordia and L. K. Kirchmayer, Tie-line power frequency control of electric power system, AIEE Trans, III-A, vol. 72, pp. 562572, June. 1953. C. Concordia and L. K. Kirchmayer, Tie-line power frequency control of electric power system: Part II, AISE Trans, III-A, vol. 73, pp. 133 146, Apr. 1954. O. I. Elgerd and C. E. Fosha, Optimum megawatt-frequency control of multiarea electric energy systems, IEEE Trans. Power Apparat. Syst., vol. PAS-89, no. 4, pp. 556563, Apr. 1970. C. E. Fosha and O. I. Elgerd, The megawatt-frequency control problem-A new approach via optimal control theory, IEEE Trans. Power Apparat. Syst., vol. PAS-89, no. 4, pp. 563-577, Apr. 1970. M.L.Kothari, P.S.Satsangi and J. Nanda, Sampled data automatic generation control of interconnected reheat thermal systems considering generation rate constraints, IEEE Trans. On Power Apparatus and systems, Vol.PAS-100, No5, pp 2334-2342, May1981. J. Nanda, M.L.Kothari and P.S.Satsangi, Automatic Generation Control of an interconnected hydrothermal system in continuous and discrete modes considering generation rate constraints, IEE Proc., vol.130, pt.D, No1, Jan.1983, pp17-27. S.P. Ghoshal and S. K. Goswami, Application of GA based optimal integral gains in fuzzy based active power-frequency control of nonreheat and reheat thermal generating systems, EPSR 67 (2003)79-88 S. P. Ghoshal, Application of GA/GA-SA based fuzzy automatic generation control of a multi-area thermal generating system, EPSR 70 (2004) Pages 115-127. M. A. Abido, Optimal design of power-system stabilizers using particle swarm optimization, IEEE Trans. Energy Convers., vol. 17, no. 3, pp. 406413, Sep. 2002. K. M. Passino, Biomimicry of bacterial foraging for distributed optimization and control, Control Systems Magazine, IEEE Volume 22, Issue 3, June 2002 Page(s):52 67. S. Mishra; A hybrid least square-fuzzy bacterial foraging strategy for harmonic estimation. Evolutionary Computation, IEEE Transactions on Volume 9, Issue 1, Feb 2005 Page(s):61 73. S. Mishra, C. N. Bhende, Bacterial Foraging Technique-Based Optimized Active Power Filter for Load Compensation, Power Delivery, IEEE Transactions on Volume 22, Issue 1, Jan. 2007 Page(s):457 465. O. I. Elgerd, "Electric Energy Systems Theory An introduction" Second edition; 1983,Tata McGraw-Hill;New Delhi.

BIOGRAPHY
Janardan Nanda (F06) He is at present an INSA(Indian National Science Academy) honorary Scientist in the Department of Electrical Engineering, Indian Institute of Technology, Delhi, India. He has served as Head of Electrical Engineering Department, Dean (UGS) and Member, Board of Governors at Indian Institute of Technology, New Delhi during 1984-92. He is a Fellow of Indian National Science Academy, a Fellow of Indian National Academy of Engineering, a Fellow of Third World academy of Sciences(Italy) and a Fellow of IEEE (USA). He has served as Chairman/ Member of many National Expert Committees and has many prestigious National Awards and Life Time Achievement Awards to his credit . He has served as Consultant to many Power Industries/ Utilities in the country and abroad. He was till recently a Member of National Board of Accreditation and Chairman of Engineering Accreditation Committee. His field of interest comprises Power System analysis, Stability, Control, Optimization and application of Computational Intelligence Techniques to Power System problems. S.Mishra (M97SM04) received the B.E. degree from University College of Engineering, Burla, Orissa, India, and the M.E. and Ph.D. degrees from Regional Engineering College, Rourkela, Orissa, India, in 1990, 1992, and 2000, respectively. In 1992, he joined the Department of Electrical Engineering, University College of Engineering Burla as a Lecturer, and subsequently became a Reader in 2001. Presently, he is an Associate Professor with the Department of Electrical Engineering, Indian Institute of Technology Delhi, India. Dr. Mishra has been honored with many prestigious awards such as the INSA Young Scientist Medal in 2002, the INAE Young Engineers Award in 2002, and recognition as the DST Young Scientist in 2001 to 2002, etc. He is a Fellow of Indian National Academy of Engineering and Institute of Electronics and Communication Engineering. His interests are in soft computing applications to power system control and power quality and renewable energy.

[3]

[4]

[5] [6] [7] [8] [9]

[10]

[11] [12] [13] [14] [15] [16]

P. G. Mishra received the B. Tech degree from Kurukshetra University, Haryana, India and M.E. degree from University College of Engineering, Burla, Orissa, India. Presently she is pursuing PhD from the Department of Electrical Engineering, Indian Institute of Technology, Delhi, India. Her field of interests are Power System Stability, Control and Renewable Energy.

[17]

K.V.Sajith received the BTech Degree in Electrical and Electronics Engineering from Govt College Of Engineering , Kannur, Kerala, India in 2008.Presently he is pursuing MTech degree in Power System from Indian Institute Of Technology, Delhi,I ndia. His interests are in Power system stability , control and renewable energy.