RESPONSE OF AGC TO A FUZZY CONTROLLER IN AN INTER-CONNECTED SYSTEM AFTER DEREGULATION

Academic Open Internet Journal

www.acadjournal.com

Volume 15, 2005

RESPONSE OF AGC TO A FUZZY CONTROLLER IN AN INTER-CONNECTED SYSTEM AFTER DEREGULATION

1.Research Scholar, Coimbatore Institute of Technology, Coimbatore, Tamilnadu, India

2. Prof. & Head of Department, Coimbatore Institute of Technology, Coimbatore, Tamilnadu, India.

3. Prof. & Dean of Department, (4 & 5) Kumaraguru College of Technology, Coimbatore, Tamilnadu, India.

ABSTRACT: In this paper the traditional AGC for three area system with fuzzy logic controller is modified to take into the account the effect of bilateral contract on dynamic in the restructured power system. The concept of Disco Participation Matrix to simulate these bilateral contracts is introduced and reflected in the 3 area block diagrams. The frequency response of the GENCOs at steady state is studied.

INDEX TERM: Automatic generation control -AGC, Bilateral contracts, Deregulation, Power System Control, Frequency Response, Fuzzy logic controller, Proportional- Integral controller, DISCO, GENCO &TRANSCO.

1. Introduction: In traditional power system industry vertically integrated system (VIS) exists. Generally government or any one agency owns generation, transmission and distribution system. This is to maintain the cost of unit power at minimum by reducing the profit margin on the huge initial investment. But the theory- healthy competitions will bring down the cost to its most optimal value gives rise to the school of deregulation of power industries. The vertically integrated utility is now being converted into horizontal integrated utility. This is the deregulation of power system. The system is divided into 3 parts namely Generation, Transmission and Distribution system. Competition is allowed in each part. There are named Generation Companies (GENCO), Distribution Companies (DISCO) and Transmission Companies (TRANCO) respectively. As there are several GENCOs and DISCOs in the deregulated structure, a DISCO has the freedom to have a contract with any GENCO of any area for transaction of power. Such transactions of power are called bilateral transactions. All transaction is cleared by impartial entity called as Independent System Operator (ISO). The ISO has to control a number of so-called "ancillary services" one of which is AGC. The concept of Disco Participation Matrix (DPM) is proposed which helps the visualization and implementation of the contracts.

1.1 Objectives When load in the system increases turbine speed drops before the governor can adjust the input. As the change in the value of speed decreases the error signal becomes smaller and the positions of governor valve get close to the required position, to maintain constant speed. However the constant speed will not be the set point and there will be an offset, to over come this problem an integrator is added, which will automatically adjust the generation to restore the frequency to its nominal value. This scheme is called automatic generation control (AGC). The role of AGC is to divide the loads among the system, station and generator to achieve maximum economy and accurate control of the scheduled interchanges of tie-line power while maintaining a reasonability uniform frequency. During large transient disturbance and emergencies AGC is by passed and other emergency control takes over. [Book 1-5] The synchronization of different system to interconnected system depends upon (i) voltage magnitude (2) frequency and (3) phase sequence. Any wide deviation from the nominal value of frequency or voltage will lead the system to total collapse. Hence AGC has gained importance with the growth of interconnected systems and with rise in size of interconnected system automation of the control system have aroused. A number of control strategies exist to achieve better performance. Due to non-linearities of power system, system parameters are linearized around an operating point. PI controller is generally used. The disadvantage of PI controller is that the mathematical model of the control process may not exist or may be too expensive in terms of computer processing powers and memory. In PI control much stress is laid on the precision of the input, the intermediate steps that process them and model of the system is questioned. While designing a PID controller following hindrance are to be crossed viz. (a) operating interfacing, (b) smooth switching operation of the components, (c) transient parameter changes, (d) effects of non-linear actuators, (e) maximum & minimum selecting and (f) build up of integral terms i.e. heuristics plays an important part. A system based on empirical rules will be more effective. Once the system behavior is thoroughly studied, then with the help of intelligent trail and error science the heuristics rules that describe the equation can be formulated. ANFIS has proven to be excellent function approximation tool. It implements a first order Sugeno style fuzzy system with 2 input and 7 membership, 49 rules & one output. ANFIS is a fuzzy inference system formulated as a feed forward neural network. Hence the advantages of a fuzzy system can be combined with a learning algorithm. [Book 6-9] For the study in this paper we have replaced the traditional PI controller with the fuzzy controller. The inputs to the fuzzy controller generally are error and change in error of frequency [paper 9-11] but here we have used change in frequency and change in tie line power i.e. the f and Ptie of each area.

Step 1: -First step is to determine the input variables and determine the degree to which they belong to each of the appropriate fuzzy sets in a membership function. The input s are always a crisp numerical valve limited to the universe of discourse of the input variable and the output is a fuzzy degree of membership in the qualifying linguistic set. [Book6-9] For our study we have the f and Ptie of each area as the input variable. Between the range -0.5 to +0.5, 7 Membership function were defined for f for and in the range of -1 to +1, 7 Membership function were defined for Ptie

Step 2: - Once the inputs have been fuzzified we know the degree to which each part of the antecedent has been satisfied for each rule. If the antecedent of a given rule has more than one part, the fuzzy operator is applied to obtain one number that represents the result of the antecedent for that rule. This number will then be applied to the output function. For the study ''OR' operator supported by 'Maximum' and 'Probabilistic' method that is know as the algebraic sum was applied while framing the rules. [Book6-9]Totally 49 rules were framed for the study.

Step 3: - Every rule has a weight (a number between 0 and 1) that is applied to the number given by the antecedent. Generally this weight is 1, so it has no effect at all on the implication process. [Book6-9]

Step 4: -Aggregation is the process by which fuzzy sets that represents the output of each rule are combined into a single fuzzy set. The input of the aggregation process is the list of the truncated output function returned by the implication process for each rule. The three methods used for study are (a) Maximum, (b) Probabilistic and (c) Sum. [Book6-9]

Step 5: -The final desired output each variable is generally a simple number. However the aggregate of a fuzzy set encompass3s a range of output value, which is defuzzified in order to resolve a single output valve from the set. The five different methods are (a) Centroid, (b) Bisector (c) Middle of Maximum (d) Largest of maximum and (e) smallest of Maximum. [Book6-9]

This fuzzy controller are tested and tuned with controller designed by ANFIS method for fine tuning.

3. Designing of a Fuzzy Controller by ANFIS method. For Adaptive-Neuro-Fuzzy Inference system controller back propagation form of the steepest descent method for membership function parameters was applied. Training error is the difference between training data output valve and input of the fuzzy inference system corresponding to the same training data input value. The output of a PI controller is trained by neural network and converted into fuzzy model. The training for all possible combination was taken for study i.e. 300 different data's was trained for an error tolerance of 0.0 and in 100 epochs by back -propagation.

4. Disco Participation Matrix: In restructured environment GENCOs sells power towards DISCOs at competitive price. The Discos have the liberty to choose the GENCOs for contracts for supply the power of their demand. The contracts may be the either with the GENCOs of their area or of another area. This makes the various combinations of GENCOs -DISCOs contracts possible in practice. Hence the concept of Disco Participation Matrix (DPM) is introduced to make the visualization of contract easier. [Paper 7] DPM is a matrix with the number of rows equal to the number of GENCOs and the number of column equal to the number of DISCOs in the each system. Each entry in this matrix\x is a fraction of total load contract by the DISCO (Column) toward GENCOs (row). The sum of all entries in the column is unity. The DPM matrix is as shown in Fig 1. where "CPF" means Contract Participation Matrix .

5. Formation of the system block diagram for simulation studies. When the change in load demand by a DISCOs, other than contracted load demand, it is reflected as a local load in the area to which this DICO belongs. This corresponds to local loads PL1, PL2 & PL3 which should be reflected in the deregulation AGC system block diagram at the point of input to the power system block. ACE signals has to be distributed among them in proposition to their participation in AGC, coefficient that distribute to ACE to several GENCOs termed as ACE Participation factor (apfj). apfj =1 where j=1to m, m is the number of GENCOs. Unlike the traditional AGC, DISCOs ask/demand a particular GENCO/ GENCOs for load power. They should be reflected in the dynamic s of the system. Turbine and generators should respond to this power demand. Thus a particular set of GENCOs are supposed to follow the load demanded by DISCOs, information signals must flow from a DISCO to a particular GENCO specifying corresponding demands. The scheduled steady state power flow on the tie line is given as Ptie1-2 scheduled = [demand of DISCOs in area II from GENCOs in area I] - [demand of DISCOs in area I from GENCOs in area I]. At any given time, the tie line power error Ptie1-2 error = Ptie1-2 actual - Ptie1-2 scheduled. In steady state Ptie1-2 error vanishes as the actual tie-line power flow reaches the scheduled power flow. The block diagram for AGC is shown in Fig 2.

6. Simulation Study. A three-area system is used to illustrate the behavior of proposed AGC. All areas are assumed to be identical. Case: 1 Consider a case where the GENCOS in each area participate equally in AGC. Therefore ACE participation factors are apf1 = apf2= apf3= apf4= apf5= apf6=0.5. Assume that DISCOS demands their loads in their areas only i.e. Disco 1 and 2 belongs to are 1 and their demands are meet by GENCOS 1 and 2 of their area. The DPM matrix can be calculated from demand - generation table given below.

DISCOs Demands.1MW .1MW .1MW .1MW .1MW .1MW Total generation by Individual GENCO

Case: 2 Consider a case where the GENCOS in each area participate unequally in AGC. Therefore ACE participation factors are apf1=0.6, apf2=0.4, apf3= 0.3 apf4= 0.7, apf5=. 2 and apf6=0.8. Assume that DISCOS demands their loads in their areas only i.e. Disco 1 and 2 belongs to area 1 and their demands are meet by GENCOS 1 and 2 of their area. The DPM matrix is same as in case 1.

7. The Results: The cases 1 and 2 are simulated using MatLab simulation tool box and the results for the cases are given in result1 and resut2 respectively.

The important of AGC will continue in restructured power system. Bilateral contracts can exist between DISCOs and GENCOs from same area or different area. The use of DPM facilitates the simulation of bilateral contracts. The fuzzy controller so designed is as good as the conventional PI controller.

1. Prabha Kundur, "Power System Stability And Control" Mc Graw Hill, Inc, USA, 1994

3. Thirry Van Certsem And Costas Vournas - "Voltage Stability Of Electric Power Systems" Kluwer Academic Publishes - 1998.

4. B.M.Weedy And B.J.Cary "Electric Power System" Fourth Edition, John Wiley And Sons - 1999.

5. Le.K.Kirchmayer, "Economical Control of Interconnected Systems" John Wiley & Sons Ltd, 1959.

6. Timothy.J.Ross, "Fuzzy Logic With Engineering Applications" Mcgraw-Hill, Inc, International Edition, 1997.

7. D.Drain Kov, H.Hellendoorn, M.Rein Frnak, "An Introduction To Fuzzy Control", Narosa Publishing House, New Delhi, 1996.

8. Statmatros.V.Kartalopoulos, "Understanding Neural Networks And Fuzzy Logic" Prentice Hall Of India Pvt.Ltd, New Delhi, 2002.

9. George.J.Klir, Bo Yuan, "Fuzzy Sets And Fuzzy Logic Theory And Applications", Prentice Hall Of India, Pvt. Ltd, New Delhi, 110001.

1. Baker, George C, " The wave of deregulation operational and design challenge" IEEE power Engg. Review Vol. 19(11) 1999 pp15-16.

2. Hyman, Leonard S "Transmission congestion, pricing and incentives" IEEE power Engg. Review Vol. 19(5) 1999 pp4-10.

3. Mountford, John D and Austria Ricardo R "Keeping the light on" IEEE spectrum vol.36 (6) 1999 pp34-39.

4. Taylor Carson W "Improving grid behavior" IEEE spectrum vol.36 (6) 1999 pp40-45.

5. Jayanth Kumar, Kah-Hoe_NG_Gerald Sheble, "AGC simulation for Price-based operation" Part 1, IEEE transaction on power systrem Vol 12 No 2 May 1997 pp 527-532.

6. Jayanth Kumar , Kah-Hoe_NG_Gerald Sheble, "AGC simulation for Price-based operation" Part 2, IEEE transaction on power systrem Vol 12 No 2 May 1997 pp 533-538

7. Richard D Christie, Anjan Bose "Load frequency control issue in power system operation after deregulation" IEEE transaction on power systrem Vol 11 No 3 August 1996 pp 1191-1200

8. Nanda, J. And Kaul, B.L., "Automatic Generation Control Of An Interconnected Power System" IEE Proc.Vol.125. No 5 May 1978, Pp 385 - 390.

9. Pan, C.T, and Liaw C.M., "An Adaptive Controller for Power System Load Frequency Control", IEEE Trans, On Power System Vol4 No.1 Feb 1989, Pp122-128.

10. Jyh - Shing Roger Jang & Chuen- Tsai Son, "Neuro Fuzzy Modeling And Control", IEEE Proceedings Vol. 83 No3 March1995 Pp 378-406.

11. Lixin Wang and Chenwei, "Approximation Accuracy of Some Neuro-Fuzzy Approaches" IEEE Trans. In Fuzzy System Vol8. No. 4 August 2000, Pp 470-478.