ISSN (Online) 232 24 ISSN (Print) 232 5526 Vol. 2, Iue 4, April 24 Integral Control AGC of Interconnected Power Sytem Uing Area Control Error Baed On Tie Line Power Biaing Charudatta B. Bangal Profeor, Electrical Engineering, Sinhgad Academy of Engineering, Pune, India Abtract: The following work i to propoe a new method of defining the area control error for implementing integral control AGC cheme for an interconnected power ytem. Conventionally, the area frequency deviation are biaed with a parameter Frequency Bia (B) and added to tie line power deviation to compoe the area control error. However, deciding a uitable value of B ha been a crucial and debatable iue over year. Many reearcher have propoed different way of deciding mot appropriate value of B for a given ytem. In the propoed method, the tie line power deviation i biaed with regulation R and added to frequency deviation to compoe the area control error. Exhautive imulation and invetigation have been carried out on model of interconnected power ytem with the propoed method and the reult have been compared with that of the conventional method. The preent dicuion and remark are not for challenging the uefulne of the conventional method, however, effort have been made to how that, the propoed method can give better reult and it can have ome additional advantage over the conventional method. Further, thi method i quite imple to adopt. The preent invetigation have been kept limited to the apect uch a magnitude of excurion, tranient & ettling time. The propoed method i found to give atifactory performance at variou combination of governor regulation, prevailing loading condition & imultaneou load perturbation. A imple two area thermal (non-reheat) interconnected power ytem i ued to demontrate the propoed method. Further invetigation on the propoed method for any other apect are open for the reearcher to trengthen the cope of the preent method and to explore it hidden merit. Keyword: Interconnected Power Sytem, Automatic Generation Control, Area Control Error, Area Frequency Repone Characteritic, Frequency Bia. I. NOMENCLATURE ACE Area control error B Frequency bia (pu MW/Hz) β Area frequency repone characteritic (AFRC) P Prevailing load in each area (pu MW) P r Rated capacity of each area (MW) fr Rated frequency (Hz) Rate of change of prevailing load with frequency ΔP Load perturbation in area (pu MW) ΔP 2 Load perturbation in area 2 (pu MW) Δf Frequency deviation of area (Hz) Δf 2 Frequency deviation of area 2 (Hz) ΔP tie eviation of tie line power (pu MW) Tg Governor time contant (Second) Tt Turbine time contant (Second) Tp Power ytem time contant (Second) Kp Power ytem contant (Hz/pu MW) K Integral controller gain R Governor regulation (Hz/pu MW) T Synchronizing coefficient of tie line (pu MW/rad) II. INTROUCTION The modelling procedure of interconnected power ytem with integral control AGC cheme are well i.e., B = = etablihed [-5][8][]. An interconnected power ytem with two identical thermal (non-reheat) area along with conventional integral control cheme (imulated in MATLAB) i hown in Fig...43 -.2 Ki() -.2 Ki(2) Integrator Integrator2.8+ Governor Fig. Two area power ytem with conventional control Conventionally, the area control error are compoed a; ACE = ΔP tie + B Δf ACE 2 = - ΔP tie + B 2 Δf 2 The iue of electing mot appropriate value of frequency bia parameter B ha been much hotly dicued and debated in all the pat year [2] [3] [6][7][][]. From mot of the literature related to AGC of interconnected power ytem, the value of B i conventionally taken equal to the AFRC ( ) for certain reaon [2]. = Kp R R Further, in majority of related literature and publihed paper, the analyi of interconnected ytem i done with the help of power ytem model which almot alway aume Kp = 2 and Tp = 2..4+ Turbine.4+ Turbine2.5 dp - -.43 B B2 /R /2.4 /2.4 /R2.8+ Governor2.2 dp2.77 T 2.5+ Power Sytem 2.5+ Power Sytem2 Tie Line Copyright to IJIREEICE www.ijireeice.com 396
ISSN (Online) 232 24 ISSN (Print) 232 5526 Vol. 2, Iue 4, April 24 P By definition, Kp ; where,. fr i the rate of change of prevailing load with change in prevailing value of frequency. For example, if at a certain intant the prevailing load on power ytem i 5% of it rated capacity ( P. 5Pr ) and operating at rated frequency of 6 Hz, then any change of load on the ytem.5 (ΔP ) at thi intant will caue. 8333 pu 6 MW/Hz and value of Kp at thi intant will be K P 2 Hz/pu MW. Alo, by definition, the power 2H ytem time contant i given atp, where H = fr inertia contant (uually taken a 5 econd). For the preent example, Tp = 2 ec. Thu it i evident that, the value Kp = 2 and Tp = 2 correpond to a pecific loading condition i.e., when P. 5Pr. However, in practice, the load perturbation can occur at any random operating condition. Hence, the value of Kp and Tp olely depend on amount of prevailing load and frequency at the time of perturbation. Here we aume that, at the time of load perturbation, whatever i the amount of prevailing load, the frequency i at rated value. Auming variou value of prevailing load in the tep of %, the correponding value of Kp and Tp are hown in TABLE I. TABLE I VALUES OF KP, TP AN IN ENTIRE OPERATING RANGE P Kp Tp. Pr 6.6666.9 Pr 66.6666..5.8 Pr 2.5.3333.7 Pr 85.743 4.2857.666.6 Pr 6.6666..5 Pr 2 2.8333.4 Pr 5 25.6666.3 Pr 2 33.3333.5.2 Pr 3 5.3333. Pr 6.666 TABLE II VALUES OF B FOR R=3%, 4% & 5% P B with conventional method B = = Kp R R=3% R=4% R=5%. Pr.572222.433333.35.9 Pr.555.43666.348334.8 Pr.568888.43.346666.7 Pr.567222.428333.344999.6 Pr.565555.426666.343333.5 Pr.563888.425.34666.4 Pr.562222.423333.34.3 Pr.56555.42666.338333.2 Pr.558888.42.336666. Pr.557222.48333.334999 In conventional AGC cheme for interconnected power ytem, the optimum value of integrator gain i elected a.2 for a thermal area. In the preent analyi, thi value wa conidered for all imulation. III. PROPOSE METHO In the propoed method, area control error are defined a; ACE = R ΔP tie + Δf ACE 2 = - R ΔP tie + Δf 2 where, R i the governor regulation. Since the governor regulation (R) i a direct meaure of change in area frequency a per change in power demand, it i choen for biaing the tie line power deviation o a to define the ACE and thi i the central innovative idea behind the propoed method. In the propoed method, the value of integrator gain need to be varied lightly a per value of R to get reult better than conventional method. The optimum value of K can be decided through imulation & trial on the power ytem model. For the model under conideration it wa oberved that, optimum value of K vary with R a per a hyperbolic function. For the model under conideration, K wa found to vary from.2 to.28 when R wa varied from 3% to 5%. Hence, in the preent analyi, value of K were ued a per the K-R characteritic obtained for the given model. The two area power ytem with propoed method (a imulated in MATLAB) i hown in Fig. 2 /R /2.4.5 dp It i therefore neceary to conider appropriate value of Kp & Tp while tudying the behaviour of an interconnected power ytem under AGC cheme. A mentioned, conventionally the value of B i taken a B = = Kp R Correponding to variou loading condition, the value of B can be calculated for variou value of governor regulation a hown in TABLE II. -..8+.4+ 2.5+ Governor Turbine Power Sytem Ki() Integrator Biaing (R) T 2.4.77 Tie Line - - -..8+.4+ 2.5+ Integrator2 Ki(2) Governor2 Turbine2 Power Sytem2 /R2.2 /2.4 dp2 Fig. 2 Two area power ytem with propoed control Copyright to IJIREEICE www.ijireeice.com 397
df2 (Hz) df (Hz) df (Hz) dptie (pu MW) df2 (Hz) ISSN (Online) 232 24 ISSN (Print) 232 5526 Vol. 2, Iue 4, April 24 IV. ANALYSIS & RESULTS For different combination of loading condition (P ), area load perturbation (ΔP, ΔP 2 ) & governor regulation (R), both the method were teted. The dynamic repone of frequency deviation (Δf & Δf 2 ) & tie-line power deviation (ΔP tie ) were compared and tudied. The integrator gain wa kept fixed at.2 for conventional method, wherea, the integrator gain wa changed a per the K-R characteritic for the propoed method. The value of variou parameter ued in the analyi are given in TABLE III. TABLE III VALUES OF PARAMETERS USE Parameter Symbol Value Unit Rated capacity of each area Pr. pu Rated frequency fr 6 Hz Governor time contant Tg.8 Second Turbine time contant Tt.4 Second Synchronizing Pu T.77 coefficient of tie line MW/rad Inertia contant H 5 Second The dynamic repone of Δf, Δf 2 & ΔP tie for a few ample combination a mentioned in TABLE IV are hown in Fig. 3 to Fig.. However, many uch combination in entire operating range, with wide range of imultaneou load perturbation & regulation value varying from 3% to 5 % were teted with both the method. TABLE IV RESULTS P (pu) Load Perturbation (pu) ΔP ΔP 2 R (pu) Figure.8 Pr.5.2.3 Fig. 3, 4, 5 -.2 -.4 -.6 -.8 -. -.2 -.4 5 5 2 x -3 - -2-3 -4-5 -6 Fig. 4: P=.8 pu, dp=.5 pu, dp2=.2 pu, R=.3 pu -7 5 5 2 25 3 35 4 -.2 -.4 -.6 -.8 -. Fig. 5: P=.8 pu, dp=.5 pu, dp2=.2 pu, R=.3 pu.5 Pr.3.6.4 Fig. 6, 7, 8.6 Pr.6..5 Fig. 9,, -.2 -.4 -.6 -.8 -.2 -.4 -.6 -.8 -.2 5 5 2 -.2 -.4 -.6 -.8 Fig. 6: P=.5 pu, dp=,3 pu, dp2=.6 pu, R=.4 pu -. -. -.2 -.2 -.4 -.6 -.4 5 5 2 Fig. 3: P=.8 pu, dp=.5 pu, dp=.2 pu, R=.3 pu -.8 -.2 5 5 2 Fig. 7: P=.5 pu, dp=.3 pu, dp2=.6 pu, R=.4 pu Copyright to IJIREEICE www.ijireeice.com 398
dptie (pu MW) df2 (Hz) df (Hz) dptie (pu MW) ISSN (Online) 232 24 ISSN (Print) 232 5526 Vol. 2, Iue 4, April 24 8 x -3 7 6 5 4 3 2 5 5 2 25 3 -.5 -. -.5 -.2 -.2 -.4 -.6 -.8 -.2 -.4 -.6 -.8 Fig. 8: P=.5, dp=.3 pu, dp2=.6 pu, R=.4 pu 5 5 2 -. -.2 Fig. 9: P=.6 pu, dp=.6 pu, dp2=. pu, R=.5 pu -.22 5 5 2 -.5 -. Fig. : P=.6 pu, dp=.6 pu, dp2=. pu, R=.5 pu -.5 5 5 2 25 3 Time (econd) Fig. : P=.6 pu, dp=.6 pu, dp2=. pu, R=.5 pu V. OBSERVATIONS AN REMARKS Exhautive imulation and trial were carried out on the power ytem model under conideration with conventional a well a propoed method for tudying the dynamic repone of Δf, Δf 2 and ΔP tie under different combination of loading condition (P ), area load perturbation (ΔP & ΔP 2 ) and governor regulation (R). The dynamic repone of Δf, Δf 2 and ΔP tie for both cae were tudied for variou apect like magnitude of excurion, time to ettle to zero (or cloe to zero upto an accuracy of about -4 ), preence of tranient etc. The ample cae are hown in Fig. 3 to Fig.. It i evident that, the preent method can give better performance than the conventional method. A few advantage of propoed method over conventional method are: ) In reality, the power ytem i alway ubjected to hift in loading condition with time and hence i alway changing. In conventional method, the ACE are dependent on value of B. Further, B depend on & R. Although R i practically not changed o often, till B ha to be changed in real time according to change in. Thu there i a need of alway monitoring the prevailing loading condition and adjuting the value of B accordingly for a proper control. In the propoed method, the ACE depend only on R and hence they are practically independent of loading condition. 2) In the propoed method, even if the feedback from the tie line i miing due to ome reaon, all the interconnected area would till continue to control the frequency deviation effectively becaue under thee circumtance; ACE = Δf ACE 2 = Δf 2 Hence all the area would continue to get control ignal a if they were iolated ingle area ytem. On the other hand, in conventional method, miing the feedback from tie line would give ACE a: ACE = B Δf ACE 2 = B 2 Δf 2 which may not offer proper control to curb frequency excurion atifactorily. Further, it wa oberved that, under thee circumtance, the teady tate error in tie line power i quite high in the conventional method. 3) It wa found that, to have better control than the conventional method, the value of integrator gain in propoed method need to be varied marginally a per value of regulation. With the propoed method, better reult are obtained at K =.2 for R = 3% and at K =.28 for R = 5%. It wa alo found that, for the intermediate value of R, the optimum value of K vary a per the hyperbolic K-R characteritic, which can be eaily determined through imulation. Thu, if K-R characteritic i obtained for the ytem under tudy, the appropriate value of K can be preet accordingly. It hould be noted that, the governor regulation i almot fixed and it i not changed frequently in a given ytem. Thu, for a practical ytem, the value of K can be preet according to preet value of R and there i no need to change K Copyright to IJIREEICE www.ijireeice.com 399
ISSN (Online) 232 24 ISSN (Print) 232 5526 Vol. 2, Iue 4, April 24 thereafter for any operating condition from zero to rated output. Following remark need to be made for inviting further invetigation on the propoed method. ) Although the imulation model ued in thi tudy i relatively implified (i.e., without involving nonlinearitie or other iue uch a generation rate contraint, ue of reheat turbine, governor dead band etc.), the propoed method can be teted under uch circumtance. 2) The propoed method can be teted for comparative tudie / invetigation on interconnected ytem involving area with different characteritic. The optimum value of integrator gain (The K-R relationhip) can be obtained for hydro or other type of prime mover. 3) The tability tudie can be carried out on variou type of power ytem model with the propoed method. Certain comment about election of conventional frequency bia parameter (B) from the report of AGC tak force of the IEEE/PES/PSE/ytem Control Subcommittee, Tranaction on Power Sytem [9] are tated below: i) The ytem natural repone coefficient ( ), i not a contant, neither it i accurately obtainable nor predictable. It depend on the current tatu and governor repone characteritic of the preently online unit and the enitivity of load. epending on the magnitude of upet from the prevailing pre-diturbance frequency, the variable number of governor coming out of deadband caue to be highly enitive to upet ize. Moreover, the obervation or meaurement of natural repone can be obcured by normal ytem activitie. E.g. generating unit may be actively reponding to prior control ignal and, of coure, individual ytem load are contantly and arbitrarily changing. ii) If every area ued an underetimated value of B, operation of the interconnection would tend to how characteritic imilar to thoe aociated with contant net interchange control. On the other hand, indicriminate ue of over-etimated value for B would tend to yield interarea generation ocillation. From the above comment it i evident that, in the conventional integral control method the iue of determining the appropriate value of B ha till remained crucial and debatable. Hence the propoed method of defining the ACE (biaing of tie line power deviation with governor regulation) need to be invetigated further for it uefulne in a broad ene and other hidden merit. REFERENCES [] O. I. Elgerd, Electric Energy Sytem Theory, Mc-Graw Hill, Newyork (983). [2] N. Cohn, Some Apect of Tie Line Bia Control on Interconnected Power Sytem, Amer. Int. Elect. Eng. Tran., vol, pp 45-436, Feb. 957. [3] C. Foha, O. I. Elgerd, The megawatt frequency control problem: A new approach via optimal control theory, IEEE Tran. Power App. Syt., vol. PAS-89, no. 4, pp. 563 577, Apr. 97. [4] IEEE Committee Report, IEEE Tran. Power App. Syt.,, vol. PAS-89, Jul./Aug. 97. Standard definition of term for automatic generation control on electric power ytem. [5] IEEE PES Committee Report, IEEE Tran. Power App. Syt.,, vol. PAS-92, Nov. 973. ynamic model for team and hydroturbine in power ytem tudie. [6] IEEE PES Committee Report, IEEE Tran. Power App. Syt.,, vol. PAS-98, Jan./Feb. 979. Current operating problem aociated with automatic generation control. [7] B. Oni, H. Graham, and L. Walker, Invetigation of Non-linear Tie Line Bia Control of Interconnected Power Sytem, IEEE Tran. Power App. Syt., vol. PAS-, no. 5, pp. 235-2356, May 98. [8] IEEE PES Working Group, Hydraulic turbine and turbine control model for ytem dynamic Studie, IEEE Tran. Power Syt., vol. PWRS-7, no., pp. 67 74, Feb. 992. [9] Naer Jaleeli, Loui S. VanSlyck, onald N. Ewart, Leter H. Fink, Undertanding Automatic Generation Control, a report of the AGC tak force of the IEEE/PES/PSE/ytem Control Subcommittee, Tranaction on Power Sytem, Vol. 7, No. 3, Augut 992) [] P. Kumar and Ibraheem, AGC trategie: A comprehenive review, Int. J. Power Energy Syt., vol. 6, no., pp. 37 376, 996. [] Ibraheem, Prabhat Kumar, and.p. Kothari, Recent Philoophie of Automatic Generation Control Strategie in Power Sytem, IEEE Tran. on Power Sytem, Vol. 2, no., Feb. 25, pp. 346-357. BIOGRAPHY Charudatta B. Bangal received the degree of B.E. (Electrical) & M.E. (Electrical Power Sytem) from College of Engineering, Pune (COEP) (Univerity of Pune, India) and Ph.. in Electrical Engineering from Bharath Univerity, Chennai, India. He ha teaching experience of 29 year in variou Engineering college in India. He i preently working a Profeor of Electrical Engineering in Sinhgad Academy of Engineering, Pune (India) and preently uperviing 2 Ph.. reearch cholar of Univerity of Pune in the area of power ytem. He ha authored 3 text book in Electrical Engineering ubject. He ha reviewed a few reearch paper for IEEE Tranaction of Power Sytem and Power elivery. Copyright to IJIREEICE www.ijireeice.com 4