Available online at www.ciencedirect.com ScienceDirect Procedia Technology (5 ) 46 44 SMART GRID Technologie, Augut 6-8, 5 A Robut Decentralized Controller Deign for Interconnected Power Sytem with Random Load Perturbation uing SDO Software Paala Gopi a, *, Potla Linga Reddy b a Reearch Scholar, Dept. of EEE, K L Univerity, Grean Field, Guntur (Dt) -55, India b Profeor, Dept. of EEE, K L Univerity, Grean Field, Guntur(Dt)-55, India Abtract Becaue of increae in load, ize and change in power ytem tructure, the repone of load frequency control problem of the interconnected power ytem i more complex. Thi paper deal with Load Frequency Control of three area interconnected Power ytem having Reheat, Non-reheat and Reheat turbine in all area repectively. The repone of the load frequency control problem in a multi-area interconnected power ytem i improved by deigning a PID controller uing different tuning technique and proved that the propoed controller, which wa deigned by Simulink Deign Optimization Software give the uperior performance than other controller for both Step load and Random load perturbation. Finally the validity and robutne of the propoed controller wa checked againt variou Step load and Random load perturbation. 5 The Author. Publihed by by Elevier Ltd. Ltd. Thi i an open acce article under the CC BY-NC-ND licene Peer-review (http://creativecommon.org/licene/by-nc-nd/4./). under reponibility of Amrita School of Engineering, Amrita Vihwa Vidyapeetham Univerity. Peer-review under reponibility of Amrita School of Engineering, Amrita Vihwa Vidyapeetham Univerity Keyword: Load Frequency Control; Interconnected Power Sytem; PID Tuning Technique; Simulink Deign Optimization(SDO) Software.. Introduction For large cale power ytem which conit of interconnected control area, load frequency then it i important to keep the frequency and inter area tie power near to the cheduled value. The input mechanical power i ued to control the frequency of the generator and the change in the frequency and tie-line power are ened, which i a meaure of the change in rotor angle. A well deigned power ytem hould be able to provide the acceptable level of power quality by keeping the frequency and voltage magnitude within tolerable limit. Change in the power * Correponding author. Tel. +985647 E-mail addre: paala.epe7@gmail.com -73 5 The Author. Publihed by Elevier Ltd. Thi i an open acce article under the CC BY-NC-ND licene (http://creativecommon.org/licene/by-nc-nd/4./). Peer-review under reponibility of Amrita School of Engineering, Amrita Vihwa Vidyapeetham Univerity doi:.6/j.protcy.5..6
Paala Gopi and Potla Linga Reddy / Procedia Technology ( 5 ) 46 44 47 ytem load affect mainly the ytem frequency, while the reactive power i le enitive to change in frequency and i mainly dependent on fluctuation of voltage magnitude. The AGC or LFC ytem olely cannot control the diturbance, it need another controller like Integral (I) or Proportional plu Integral (PI), Proportional plu Integral plu Derivative (PID) controller [8]. In [], the ame author were explained about literature review on AGC trategie in interconnected power ytem very clearly.. Mathematical Modeling of Power Sytem The main difference between Load Frequency Control of multi-area ytem and that of the ingle area ytem i, the frequency of each area of multi-area ytem hould return to it nominal value and alo the net interchange through the tie-line hould return to the cheduled value. So a compoite meaure, called area control error (ACE), i ued a the feedback variable. A decentralized controller can be tuned auming that there i no tie-line exchange power, P tiei =. In thi cae the local feedback control will be ui = -K i ()B i f i. Thu load frequency controller for each area can be tuned independently... Modeling of Power Generating Unit In power ytem, a turbine unit i ued to tranform the natural energy (like energy from team or water) into mechanical power (ΔP m ) that i upplied to the generator. In LFC model, there are three different type of commonly ued turbine; thoe are Non-Reheat, Reheat & hydraulic turbine, all of which can be modelled by tranfer function [, ]. Non-reheat turbine are firt-order unit. A time delay (T t ) occur between witching the valve and producing the turbine torque. The tranfer function of the non-reheat turbine i repreented a GNr () NUMNt() ( ST t ) DENNt() Becaue of different tage due to high and low team preure in the Reheat turbine, it wa modelled a econdorder unit. The tranfer function of reheat turbine can be repreented a Gr () SCT t r NUMt() ( ST t r)( ST lpr) DENt() Where T lpr i the low preure reheat time and C repreent the high preure tage rating, T tr i the reheat turbine time contant. The Speed Governor are ued in power ytem to ene the frequency variation (Δf) which are caued by the load change (ΔP L ) and are cancelled by varying the turbine input. P L Valve/Gate - Steam/Water Turbine P mech + Generator f P V Load Ref. Set point - + S R T g R - _ Fig. Block diagram repreentation of Speed Governing unit The block diagram repreentation of a peed governing ytem i hown in Fig., where R i the peed regulation and T g i the time contant of the Speed Governor [6]. Suppoe if there i no load reference and there are load change occur, ome part of the change may be compenated by the valve or gate etting and the remaining of the change i repreented in the form of frequency variation. The goal of Load Frequency Control (LFC) i to
48 Paala Gopi and Potla Linga Reddy / Procedia Technology ( 5 ) 46 44 compenate the frequency deviation due to active power load variation. Thu, the load reference et-point can be ued to adjut the valve/gate poition o that all the load change i cancelled by the power generation rather than reulting in a frequency deviation a hown in Fig.. The tranfer function can be repreented a Gg () NUMg() ( Tg) DENg() A generator convert the mechanical power developed by the turbine into electrical power. Once the load variation occur, the mechanical power (P mech ) from the turbine will not match the electrical power (P ele ) generated by the generator. The error between the mechanical (ΔP mech ) and electrical power (ΔP ele ) i integrated into the rotor peed deviation (Δω r ), which can be converted into the frequency variation (Δf) by multiplying with π. The Fig. how the block diagram of generator with load damping (D) effect. P L f i P mech - f + P tie T ij + M+D _ S f j Fig. Block diagram of generator with load damping effect The Laplace tranform of generator with load damping i Gp () ΔP mech () ΔP L () = (M + D)ΔF() Kp NUMp() (D MS) ( ST p) DENp() Fig. 3 Block diagram repreentation of tie-line link In an interconnected power ytem, different area are connected with each other with the help of tie-line. When the frequency variation in the two area are different, a power exchange occur through the tie-line between the connected two area. The block diagram repreentation of tie-line i a hown in Fig.3. The laplace tranform of tie line in Fig.3 i given by Tij( F i() F j()) Ptieij() Where ΔP tieij i tie line power exchange between area i and j, and T ij i the tie-line ynchronizing coefficient between area i and j [6]. Droop Characteritic i P d u i X Gi P Gi Load & f i Generator i Turbine i Machine i - B i + + /S T i ACE i P tiei + -.( j). + T in - f i.(j i) Fig.4 Block diagram of control area i The goal of Load Frequency Control i not only to compenate the frequency error in each area, but alo to control the tie-line power exchange according to chedule [6]. Becaue the tie-line power error i the integral of the
Paala Gopi and Potla Linga Reddy / Procedia Technology ( 5 ) 46 44 49 frequency difference between each pair of area, if we control frequency error to zero, any teady tate error in the frequency of the ytem would reult in tie-line power error. Therefore, we need to include the information of the tie-line power deviation into our control input. A a reult, an area control error (ACE) i defined a (referred to Fig.4) ACEi Ptieij Bifi j,...n, j i Where B i i the frequency bia contant for area-i and B i =/R i + D i. Thi ACE ignal i ued a the plant output of each power generating area [6]... Modeling of Power Sytem Area Let Area-I, Area-II and Area-III are the non identical interconnected power ytem with Reheat, Nonreheat and Reheat turbine in all three area repectively. The tranfer function of each area with generator drooping characteritic can be defined a NUMg() NUMt() NUMp() G p () = B DENg() DENt() DENp() NUMg() NUMt() NUMp()/R The tranfer function of all three area of interconnected power ytem are a follow (ee appendix for Turbine, Speed Governor and Power ytem parameter): For Area-, the tranfer function i G () = * (3 )*5*.587 (.8)(.3 3 )( 5) * (3 )*5/ 4 48.75 6.5 3 6 44.3 55 6.5 For Area-, the tranfer function i G () = **.45 (.8)(.3)( ) **/.4 For Area-3, the tranfer function i 3 6.5 5.88 4.46 6.5 G 3 () = * (5 )**.45 (.8)(.3 3 )( ) * (5 )*/.7 4 53.5.65 3 5.98 44.5 58.4.65 3. PID Controller Tuning For indutrial plant proce, the conventional PID controller are mot commonly ued. They are doing ome challenge to control, intrumentation and power engineer in the area of tuning of the gain of controller required for bet tranient performance and tability. There are everal precriptive rule ued for tuning of PID controller [3-5]. The parallel form of a PID controller ha tranfer function: Ki Gc() = Kp Kd Kp( Td) Ti K p = Proportional Gain contant; K i = Integral Gain contant; T i = Reet Time contant =K p /K i, K d = Derivative gain contant; T d = Rate time or derivative time contant. The tuning of the PID load frequency controller of multi-area power ytem that it ha to bring frequency of each area to it nominal value and alo the change in tie-line power hould return to the cheduled value. So the combination of both, called Area Control Error (ACE), i ued a feedback variable. For i th area, the Area Control Error (ACE) i defined a ACE i = P tiei + B i f i and Feedback control ignal for area-i i u i = - K i () AEC i. A PID load frequency controller can be tuned auming that there i no tie line power exchange i.e P tiei =. Now the feedback control ignal u i = - K i () B i f i.
4 Paala Gopi and Potla Linga Reddy / Procedia Technology ( 5 ) 46 44 3.. Peen Integral Rule (PIR) Tuning Recently a new tuning rule for PID controller wa prepared by Ziegler-Nichol called Peen Integral Rule (PIR). The procedure for tuning a PID controller uing Peen Integral Rule (PIR) i imilar to nd method of Ziegler- Nichol PID tuning [7]. The tep for tuning a PID controller uing Peen Integral Rule i a follow:. Reduce the integrator and derivative gain to.. Increae proportional gain K p value from to ome critical value at which utained ocillation occur. 3. Note the value K cr and the correponding time period of utained ocillation, P cr. 3.. Automatic PID Tuning To tune PID controller of ingle loop control ytem having PID automatically, ue Simulink control deign PID Tuner. With PID tuner it i poible to achieve good balance between performance and robutne. The procedure for automatic PID tuning i a follow:. Create a Simulink model with a PID controller for any order and any time delay in MATLAB/Simulink.. Double click on PID controller block to open the PID controller dialog box. 3. In dialog box, click Tune, it automatically linearize the plant and deign an initial controller. The K P, K I and K D value of PID controller for all three area with Peen Integral Rule (PIR) and Automatic PID tuning method are hown in Table and Table repectively. Table. Peen Integral Rule (PIR) Tuning Area-i K P K I K D Area-I 8.795.4487.69 Area-II 3.749 9.357.567 Area-III 7.844 9.635.76 Table. Automatic PID Tuning Area-i K P K I K D Area-I 5.766 4.737.76 Area-II.954.8696.5 Area-III 5.74 4.635.4896 3.3. Integral Square Error (ISE) Optimization A meaure of ytem performance formed by integrating the quare of the ytem error over a fixed interval of time; thi performance meaure and it generalization are frequently ued in linear optimal control and etimation theory. In thi paper, the tranfer function of PID Controller with Integral Square Error Optimization technique wa obtained by the SISO tool in MATLAB/Simulink. The tranfer function for PID Controller of different area in the interconnected power ytem with Integral Square Error (ISE) Optimization technique are given a For Area-I G c () = 6.456 + 4.96/S + 4.867S, for Area-II G c () = 6.9775 + 4.9839/S + 4.9839S and for Area-II G c () = 7.4557 + 4.975/S + 4.87S. 3.4. Simulink Deign Optimization (SDO) Technique Signal Contraint block i connected in developed MATLAB/Simulink model to optimize the model repone for known input. To get optimized parameter of a Simulink model, the following tep have to follow:. Develop and open the Simulink model.. Open the Simulink deign optimization block by typing dolib at MATLAB command prompt. 3. Drag and drop the ignal contraint block in the developed MATLAB/Simulink model. 4. Connect the ignal contraint block to ignal to which we want to get pecified deign requirement. To deign the controller for interconnected power ytem uing SDO Software, it require Simulink Deign Optimization toolbox. The tranfer function for PID Controller of different area in an interconnected power ytem having different turbine unit in each area with Simulink Deign Optimization Software are given a, for Area-I G c () =.966 +.84/S + 5.376S, for Area-II G c () =.3383 + 3.743/S +.338S and for Area-III G c () = 4.3 + 5.595/S + 4.69S.
Paala Gopi and Potla Linga Reddy / Procedia Technology ( 5 ) 46 44 4 4. Simulation and Reult Analyi Let Area-I, Area-II and Area-III are non identical, i.e. all area of interconnected power ytem incorporating Reheat, Non-reheat and Reheat turbine repectively. The parameter of all three area are collected from variou team power tation in India and are hown in Appendix-A. The Robutne of the deigned controller can be etimated by the following two illutration. 4. Illutration-I: To etimate the performance of the deigned decentralized PID controller, different tep load perturbation are aumed a dp L =.pu, dp L =.pu and dp L3 =.5pu i applied to Area-I, Area-II and Area-III repectively at t = ec. The repone of Frequency variation (pu) and Tie line power variation (pu) of the ytem are hown in Fig. 5 and Fig.6 for nominal value of ytem parameter.. Load frequency variation of Area-I for naminal value ofparameter 3 x -3 Tie Line power variation of Area-I for nomina lvalue of parameter.5.5.5 Frequency variation(pu) -.5 -. Tie Line power variation(pu).5 -.5 -.5 - -.5 -. 5 5 (a) Area-I Frequency variation (df ) - 5 5 Time (Sec) (a) Area-I Tie Line Power variation (dp tie). Load Frequency Variation of Area-II for nominal value of parameter 3 x -3 Tie Line power Variation of Area-II for nominal value of parameter.5 Frequency Variation (pu) -.5 -. -.5 Tie Line power Variation (pu) - - -3 -. -4 -.5-5 -.3 5 5 (b) Area-II Frequency variation (df ) -6 5 5 (b) Area-II Tie Line Power variation (dp tie). Load Frequency Variation of Area-III for nominal value of parameter.5 x -3 Tie line power variation of Area-III for nominal value of parameter.5.5 Frequency Variation (pu) -.5 Tie line power variation(pu).5 -. -.5 -.5 5 5 (c) Area-III Frequency variation(df 3) - 5 5 (c) Area-III Tie Line Power variation(dp tie3) Fig.5 Frequency deviation of Three area interconnected power Fig. 6 Tie Line Power deviation of Three area interconnected ytem ytem The performance of the deigned decentralized PID controller i alo checked by applying variou tep load perturbation (like dp Li =.pu and dp Li =.pu) in all three area at t = ec. The ummary of performance of all controller for variou tep load perturbation i a hown in Table 3.
4 Paala Gopi and Potla Linga Reddy / Procedia Technology ( 5 ) 46 44 Tuning Method t Peak ( -3 ) pu Table.3 Summary of Illutration-I Area-I (dp L=.pu) Area-II (dp L=.pu) Area-III (dp L3=.pu) t Peak t Peak ( -3 ) pu ( -3 ) pu PIR -.6 55-9.5 9-9.56 88 PID Tuner -.5.5-4.7 6.5-3.75 ISE Optimization -6.4 5 -.3 5. -5. 6.83 SDO Software -6.56 5-3.4 3.5-5. Area-I (dp L=.pu) Area-II (dp L=.pu) Area-III (dp L3=.5pu) Tuning Method t Peak ( -3 ) pu t Peak ( -3 ) pu t Peak ( -3 ) pu PIR -3 > -7.8 > -4. > PID Tuner -5.3.55-6.7 9.4-4.9.5 ISE Optimization -3.5 3.57-3. 4.5-7.8.55 SDO Software -6.9.55-6.7 8.45-7.5.3 Area-I (dp L=.pu) Area-II (dp L=.pu) Area-III (dp L3=.pu) Tuning Method t Peak ( -3 ) pu t Peak ( -3 ) pu t Peak ( -3 ) pu PIR -9. 56-3.95 78-3.75 8 PID Tuner -8. 7-3.6 6. -4.7 6. ISE Optimization -6. 5-36.8 5. -3 7 SDO Software -6.4.5-8.5 9. -.4 3.93 4. Illutration-II: In thi ection, the performance of the deigned decentralized PID controller wa alo etimated by applying Random load perturbation a dp L =.pu, dp L =.pu and dp L3 =.5pu to Area-I, Area-II and Area-III repectively at t = ec. The repone of Frequency variation (pu) and Tie line power variation (pu) of the ytem are hown in Fig. 7 and Fig.8 for nominal value of ytem parameter for certain time period..5 Frequency variation (pu) of Area-I for Random load perturbation (b) Area-II Frequency variation (df )..5 3 x -3 Tie-line Power Varation (pu) of Area-I with Random Load Variation. Change in frequency(pu).5 -.5 -. -.5 Tie-line Power Varation (pu) - -. - -.5 7 75 8 85 9 95 Time(Sec).5..5. (a) Area-I Frequency variation (df ) Frequency variation (pu) in Area-II for Ramdom Load Perturbation -3 7 75 8 85 9 95 (a) Area-I Tie Line Power variation (dp tie) 4 x -3 Tie-Line power variation (pu) of Area-II for Random Load Varation 3 Change in Frequency (pu).5 -.5 -. Tie-Line power variation (pu) - - -.5 -. -3 -.5 7 75 8 85 9 95 Time (Sec) -4-5 7 75 8 85 9 95
Paala Gopi and Potla Linga Reddy / Procedia Technology ( 5 ) 46 44 43 (b) Area-II Tie Line Power variation (dp tie).5 x -3 Tie-line power varation (pu) of Area-III for Random load variation. Frequency variation (pu) of Area-III for Random Load Perturbation Change in frequency (pu).5..5 -.5 With PIDTuner Tie-line power varation (pu).5.5 -.5 - -. -.5 -.5-7 75 8 85 9 95 Time -. 7 75 8 85 9 95 (c) Area-III Frequency variation (df 3) Fig.7 Frequency deviation of Three area interconnected power ytem (c) Area-III Tie Line Power variation (dp tie3) Fig. 8 Tie Line Power deviation of Three area interconnected ytem From the above illutration, it can be oberved that the controller deigned by Simulink Deign Optimization oftware give better performance and more robut than other controller for variou value of tep load and Random load perturbation. 5. Concluion The decentralized PID controller were deigned for three area interconnected power ytem with different turbine in repective area uing variou PID tuning technique. From the reult, the Load Frequency Controller deigned by Simulink Deign Optimization Software give the effective and uperior performance than other controller for variou tep load and random load perturbation in all three area. Alo the Simulink Deign Optimization Software tuned PID Load Frequency Controller i more robut than other controller for both variou tep load and random load perturbation in all three area. Appendix A. The nominal parameter of Reheat and Non-Reheat Turbine are collected from variou Thermal power plant in India and are a hown below [9]: Parameter Area-i Area-I Area-II Area-III Speed Governor Time contant.8.8.8 Speed Governor Regulation.7.4 Power Sytem gain contant 5 Power Sytem Time contant 5 Parameter Area-i Area-I Area-II Area-III Turbine Time contant.3.3.3 Coefficient of re-heat team turbine (High Preure).3 -.5 Re-heater time contant (Low Preure) - Rated capacity = MW; P tiemax = MW; (δ -δ ) = 3 ; Frequency f = 6Hz; D i = 8.33x -3 ; Syn. Co-efficient T ij =.545. Reference [] Paala Gopi and Dr. Polta Linga Reddy, A Critical review on AGC trategie in interconnected power ytem, Proceeding of IET Int. Conf. on Sutainable Energy and Intelligent, Dec 3, p 99-35. [] Wen Tan Unified Tuning of PID Load Frequency Controller for Power Sytem via IMC IEEE Tran. on Power Sytem, Vol. 5, No., p. 34-35, Feb. [3] Sigurd Skogetad, Simple analytic rule for model reduction and PID controller tuning, Jrnl of Proce Control, Vol.3, p. 9 39, 3. [4] S.Ohba, H.Ohnihi & S.Iwamoto,,An Advanced LFC Deign Conidering Parameter Uncertaintie in Power Sytem, Proceeding of IEEE conf. on Power Sympoium, p. 63 635, Sep. 7. [5] K. P. Singh Parmar S. Majhi, D. P. Kothari, Optimal Load Frequency Control of an Interconnected Power Sytem, MIT Int. Journal of Electrical & Int. Engineering Vol., No., p-5, Jan. [6] P. Kundur, Power Sytem Stability and Control, New York: McGraw-Hill, 3. [7] Brian R Copeland,The Deign of PID Controller uing Ziegler Nichol Tuning, IJETE, p.-4, Mar 8.
44 Paala Gopi and Potla Linga Reddy / Procedia Technology ( 5 ) 46 44 [8] R. N. Patel, S. K. Sinha, R. Praad, Deign of a Robut Controller for AGC with Combined Intelligence Technique, Journal of World Academy of Science, Engineering and Technology, p. 687-693, 8.