Performance Analysis of Fuzzy Logic Based Unified Power Flow Controller

Transcription

1 Internationa Journa of Eectrica and Eectronics Engineering 3: Performance Anaysis of Fuzzy Logic Based Unified Power Fow Controer Lütfü Saribuut, Mehmet Tümay, and Đyas Eker Abstract FACTS devices are used to contro the power fow, to increase the transmission capacity and to optimize the stabiity of the power system. One of the most widey used FACTS devices is Unified Power Fow Controer (UPFC). The controer used in the contro mechanism has a significanty effects on controing of the power fow and enhancing the system stabiity of UPFC. According to this, the capabiity of UPFC is observed by using different contro mechanisms based on P, PI, PID and fuzzy ogic controers (FLC) in this study. FLC was deveoped by taking consideration of Takagi- Sugeno inference system in the decision process and Sugeno s weighted average method in the defuzzification process. Case studies with different operating conditions are appied to prove the abiity of UPFC on controing the power fow and the effectiveness of controers on the performance of UPFC. PSCAD/EMTDC program is used to create the FLC and to simuate UPFC mode. T Keywords FACTS, Fuzzy Logic Controer, UPFC. I. INTRODUCTION HE growth of the power systems in the future wi rey on increasing the capabiity of existing transmission systems rather then buiding the new transmission ines and the power stations for an economica and an environmenta reasons. The requirement of the new power fow controers, which is capabe of increasing the transmission capabiity and controing the power fow through the predefined corridors, wi certainy increase due to the dereguation of the eectricity markets. Additionay, these new controers must be contro the votage eves and the fow of the rea/reactive power in the transmission ine to use fu capabiity of the system in some cases with no reduction in the system stabiity and security margins []. A new technoogy concept known as Fexibe Aternating Current Transmission Systems (FACTS) technoogy was presented in the ate of 980s [2]. FACTS devices enhance the stabiity of the power system with its fast contro characteristics and continuous compensating L. S. is with the Çukurova University, Department of Eectrica & Eectronics Engineering, Bacaı, Adana, Turkey (phone: /9; e-mai: saribuut@cu.edu.tr). M. T. is with the Çukurova University, Department of Eectrica & Eectronics Engineering, Bacaı, Adana, Turkey (phone: /48; e-mai: mtumay@cu.edu.tr). Đ. E. is with the Çukurova University, Department of Eectrica & Eectronics Engineering, Bacaı, Adana, Turkey (phone: /35; e-mai: iyas@cu.edu.tr). capabiity. The controing of the power fow and increasing the transmission capacity of the existing transmission ines are the two main objectives of FACTS technoogy [3]. Thus, the utiization of the existing power system comes into optima condition and the controabiity of the power system is increased with these objectives. Gyugyi proposed the Unified Power Fow Controer which is the new type generation of FACTS devices in 99 [4]. Unified Power Fow Controer (UPFC) is the member of FACTS device that has emerged for the controing and the optimization of power fow in the eectrica power transmission systems [5]. This device formed of the combination of two other FACTS devices namey as Static Synchronous Compensator (STATCOM) and the Static Synchronous Series Compensator (SSSC). These are connected to each other by a common DC ink, which is a typica storage capacitor. The a parameters of the power transmission ine (impedance, votage and phase ange) can be contro simutaneousy by UPFC [6]. In addition, it can perform the contro function of the transmission ine rea/reactive power fow, UPFC bus votage and the shunt-reactive-power fow contro [7]. The contro mechanism and the controer have an important effect on the performance of UPFC. In the iterature, severa contro mechanisms are used in UPFC modes. A nove fuzzy inference system described in matrix form is proposed and used to improve the dynamic contro of rea and reactive power [8]. Two fuzzy ogic controers based on Mamdani type fuzzy ogic are used. One of the controers is proportiona fuzzy ogic controer (PF-UPFC) and the other is Hybrid fuzzy ogic UPFC (HF-UPFC) [3]. The seection of suitabe ocation for UPFC is studied and composite-criteriabased fuzzy ogic is used to evauate the network contingency ranking [9]. The power-feedback contro scheme is used in the contro mechanism of UPFC [0]. The power fuctuation is damped readiy and the vaue of reactive power is minimized as possibe by using severa time constants. However there is no vaue changed in the rea power. The contro method of variabe interva-fuzzy-mutua is used in the contro mechanism of UPFC []. In the simuation resuts, there is a high overshoot vaues occurred both rea power and bus votage during the three phase fauts appied. However, the rea power vaue is increased but there is no vaue changed in the reactive power. The performance of UPFC is observed by using three different controers [2]. In the simuation resuts, the variation of the rea power direction can be observed 77

2 Internationa Journa of Eectrica and Eectronics Engineering 3: easiy. However, the vaue of reactive power is kept at zero because of there is no reactive power fow in the system. The performance of Puse Width Moduation (PWM) based UPFC is observed [3]. According to resuts, the vaues of rea and reactive power are changed in arge vaues with UPFC because of the ow vaues of bus votage. The capabiity of UPFC on controing of the power fow and the effectiveness of controers on performance of UPFC in the power transmission ine are examined in two case studies by using different contro mechanisms based on PI and fuzzy controers in this paper. In the modeing of fuzzy controer, Takagi-Sugeno Inference System is used in the decision making process and Weighted Average method which is the specia case of Mamdani mode is used in the defuzzification process. The eectromagnetic transient simuation program PSCAD/EMTDC is used to create UPFC mode and to obtain the resuts of case studies [4]. II. UPFC SYSTEM The UPFC is the combination of two votage-source converters; one converter is connected to the power system through a shunt transformer, whereas the other converter is inserted into the transmission ine through a series transformer []. The converters are connected by a common DC-ink where the capacitor is couped and it aows a bi-directiona rea power fow between the output termina of shunt converter and the input terminas of series converter. The UPFC can be decouped be into two branches according to its contro and structure perspective. One of these branches is the parae branch formed by the shunt transformer, VSC and DC capacitor. It operates as a STATCOM. The other is series branch composed of the series transformer, a VSC and DC capacitor. It behaves as a SSSC. The basic system configuration of UPFC structure is shown in Fig.. (V dc ) and instant vaue (V dc ) of DC bus votage in this study. The resut is used as an ange σ in the sinus bock of contro mechanism. The other component is reactive part of current and it can be set to any desired reference eve (inductive or capacitive) within the capabiity of the converter. This component is cacuated in the same way with the other component by finding the differences between reference votage (V sendpu ) and instant vaue of bus votage (V sendpu ). The resut is used as a magnitude (IVI pu ) to mutipy with sinus bock. The cacuation of these two current components is shown in Fig. 2 and Fig. 3. Hence, the reference signas are cacuated to use in the SPWM techniques. PI controer based and fuzzy controer based contro mechanisms of shunt converter are given as separatey in the foowing graphics. V dc % Vdc pu V dc V reca V recb V recc Vdc _ set _ vaue pu V recda V recdb V recdc Ki K p + s Ki K p + s α Phase Locked Loop IVI pu Mod Function σ IVI sin( σ ) SPWM Controer Fig. 2 PI controer based contro mechanism Phase Locked Loop V senda_ref sin( ) + send Firing Puse _ of Converter V send_pu Deay _ Function α e st Fuzzy Controer K i K p + s IVI pu Mod _ Function σ IVI sin( σ ) SPWM Controer Firing Puse _ of Converter Fig. Basic scheme of UPFC system A. The Contro Mechanism of Shunt Converter The shunt converter is operated to draw a controed current from the ine. One component of this current is automaticay determined as a requirement rea power to meet the osses in the converters and to baance the rea power of the series converter [5]. However, the net rea power is used to maintain the dc votage and to provide the reactive compensation without an externa-eectric-energy source to the system independenty [6]. This component of current is cacuated by finding the differences between reference vaue Fig. 3 Fuzzy controer based contro mechanism B. The Contro Mechanism of Series Converter The main function of UPFC is actuaized by the series converter. It contros the magnitude and the ange of the votage injected in series with the ine. This votage injection is aways intended to infuence the power fow the ine [5]. The magnitude and the phase ange of series-injected votage V inj is cacuated by contro mechanism to provide the desired rea and reactive power fow in the transmission ine. In the 78

3 Internationa Journa of Eectrica and Eectronics Engineering 3: process of doing this, the series inverter wi exchange rea and reactive power with the ine. The reactive power is eectronicay provided by the series inverter and the rea power is transmitted to the dc terminas. The contro mechanism of series converter is shown in Fig. 4. In the contro mechanism, the receiving-end-generator votage transforms into d q0 (Park) transform [7]. The d q0 components of receiving-end-votage are mutipied with determined reference vaue of active/reactive powers (P ref /Q ref ). Hence, the signas desired by user V recd,q,0 are cacuated. Then, the phase ange ( ) of transmission ine is obtained by using PLL (phase ooked oop). It is used to cacuate the reference signas V reca,b,c_ref. These signas are transformed into d q0 form. The desired signas are subtracted from the reference signas and the resuts are transformed into three phase baanced system to use in the sinusoida-puse-width moduation (SPWM). Thus, the firing anges of IGBTs (insuated gate bipoar transistors) are produce from the process of SPWM technique. V reca V recb V recc Phase Locked Loop sin( ) V reca_ref Vrecb_ref V recc_ref A B C to D Q O V recd_ref V recq_ref V reco_ref each phase are measured and converted into per unit (pu.) vaue. For phase A, the error and error rate are defined as: err = V V () A sin( ), pu send, pu err = err ( n) err ( n ) (2) A A A where V sin(),pu is a PLL votage possessed the same phase with the sending end votage, V send,pu is the phase of sending end votage and n is the samping time. B. FLC The section of FLC is divided in three subsections. These subsections are given as summarized in the foowing: Fuzzification: The numeric input-variabe measurements are transformed by fuzzification part into the fuzzy inguistic variabe, which is a ceary defined boundary with a crisp. These inguistic variabes of error/error rate are shown in Fig. 5. Decision Making: The fuzzy modes are created by using Sugeno Inference System [20]. According to this system, the I th rue can be cacuated by using in the foowing equations: A B C to D Q O V recd V recq V reco P ref Q ref V recd V recq + V rec0 D Q O to A B C V ref_a V ref_b V ref_c SPWM Controer Firing Puse _ of Converter () L : If x is F and... and x is F, then (3) n y = c + c x + c x + + c x (4) n n n where F denotes fuzzy set, the output set and x.x 2 is the inputs. c is the rea coefficients, y is Fig. 4 Contro mechanism of series converter III. IMPLEMENTATION OF FLC IN UPFC FLC are formed by simpe rue based on If x and y then z. These rues are defined by taking hep from person s experience and knowedge about the system behavior. The performance of the system is improved by the correct combinations of these rues. Each of the rues defines one membership which is the function of FLC. More sensitivity is provided in the contro mechanism of FLC by increasing the numbers of membership functions [8]-[9]. In this study, the inputs of the fuzzy system are assigned by using 7 membership functions and the fuzzy system to be formed in 49 rues. Hence, the sensitivity in the contro mechanism is increased. The fuzzy contro system is divided into three main sections. These sections are expained in the foowing. A. Error Cacuation The error signa (err A ) is cacuated from the difference between the source votage vaue and the reference vaue obtained from PLL. Beside, the error rate signa ( err A ) is the differences between the variation of error at current samping and its previous samping. These signas of suppy votage for DEGREE OF MEMBERSHIP.0 c LN MN SN VS SP MP LP b a 0 a b c.0 Fig. 5 Error and error rate of fuzzy membership functions The basic if-then rue is defined as If (error is very sma and error rate is very sma) then output. The signas error and error rate are described as inguistic variabes in the FLC such as arge negative (LN), medium negative (MN), sma negative (SN), very sma (VS), sma positive (SP), medium positive (MP) and arge positive (LP). These are shown in Fig. 5. In the same way, the input vaues of the fuzzy controer are connected to the output vaues by the if-then rues. The reationship between the input and the output vaues can be achieved easiy by using Takagi-Sugeno type inference method. The output vaues are characterized by memberships and named as inguistic variabes such as negative big (NB), negative medium (NM), negative sma (NS), zero (Z), 79

4 Internationa Journa of Eectrica and Eectronics Engineering 3: positive sma (PS), positive medium (PM) and positive big (PB). The membership functions of output variabes and the decision tabes for FLC rues are seen in Tabe I. Error rate /Error TABLE I FUZZY DECISION TABLE LP MP SP VS SN MN LN LP PB PB 2 PB 3 PM 4 PM 5 PS 6 Z 7 mechanisms based on P, PI, PID and fuzzy ogic controers. The test system is shown in Fig. 6 and the parameters are given in the Appendix. Vs δ V V r MP PB 8 PB 9 PM 0 PM PS 2 Z 3 NS 4 SP PB 5 PM 6 PM 7 PS 8 Z 9 NS 20 NM 2 VS PM 22 PM 23 PS 24 Z 25 NS 26 NM 27 NM 28 SN PM 29 PS 30 Z 3 NS 32 NM 33 NM 34 NB 35 MN PS 36 Z 37 NS 38 NM 39 NM 40 NB 4 NB 42 LN Z 43 NS 44 NM 45 NM 46 NB 47 NB 48 NB 49 Defuzzification: In the defuzzification process, the controer outputs represented as inguistic abes by a fuzzy set are converted to the rea contro (anaog) signas. In the created fuzzy mode, Sugeno s Weighted Average method which is the specia case of Mamdani Mode is seected for the defuzzification process [2]. According to this mode, the defuzzification is achieved by using foowing equations: y= M = M = w y w n = (6) i= w M ( i ) F x i where w is the overa truth vaue of the rue () L, M ( x ) i F i the membership function described the meaning of the inguistic variabe F. C. Signa Processing The contro signas are produced from the output of FLC process. They are used in the generation of switching signas for converter by comparing with carrier signa. It can be shown in Fig. 2 and Fig. 3. IV. CASE STUDIES PSCAD/EMTDC program is used to simuate the modeing of UPFC and the test system. The parameters of simuated system are seected ow ratings to be enabed the impementation of system in the aboratory environment. In the simuation case studies, two generators are used and named as sending end and receiving end generators, respectivey. UPFC is constructed at the sending end bus before the ine impedance. Two case studies were carried out to test the performance of UPFC with different contro (5) is Fig. 6 Test system for case study In the first case study, the receiving end generator is deayed from the sending end generator according to severa phase anges. The vaues of the rea/reactive power resuts in the ine are taken by consideration of using the different controers in the contro mechanism separatey and compared as with and without UPFC in the ine. The resuts are given in Tabe II. TABLE II POWER FLOW OF THE LINE FOR PHASE VARIATION Phase ange of receiving end generator ( 0 ) without UPFC P Controer with UPFC PI Controer with UPFC P(KW) Q(KVAR) P(KW) Q(KVAR) P(KW) Q(KVAR) PID Controer P(KW) with UPFC Q(KVAR) FUZZY P(KW) Controer with UPFC Q(KVAR) The test system for second case study is given in Fig. 7. In this case study, three transmission ines, which have same impedance parameters, are used and UPFC is constructed on transmission ine 3 after the oad. The reactive power and the bus votage, the connection point which UPFC is connected to the ine, are expected to be restored to their nomina vaues by UPFC. In the second case, the three phase faut is appied to the transmission ine. It is started at the.4 sec and it is continued 0.2 sec. The faut is sensed by contro mechanism of series converter, the eectronic bypass is immediatey activated to protect the series converter. The eectronic bypass is removed by the contro mechanism automaticay after the faut. The PI and fuzzy controer s resuts vaues of reactive 720

5 Internationa Journa of Eectrica and Eectronics Engineering 3: power and bus votage of receiving end are compared as graphicay during the faut to be considered with and without UPFC in the ine. V s δ Fig. 7 Test system for case study 2 The variations on the receiving end reactive power (Q re ) and the receiving end votage in pu. (Vpu.re ) are iustrated as graphicay in Fig. 8 and Fig Qre (with UPFC) Qre (without UPFC) V r V. CONCLUSION In this paper, the performance of UPFC on controing of the rea/reactive power fow in the ine is examined with different contro mechanisms based on P, PI, PID and fuzzy controers. For the first case study, the resuts show that P controer does not sufficient on controing the power fow whie FLC shows the best performance on controing the power fow. In the second case study, three phase faut is appied to the test system. The resuts are taken to be consideration of PI and fuzzy controers because of the taken same resuts of these controers in the first case study. The PI and fuzzy controers show neary same resuts but there is a ow overshoot occurred during the faut in the fuzzy controers resuts. According to resuts that UPFC improves the system performance under the transient and the norma conditions. However, it can contro the power fow in the transmission ine, effectivey. Beside, the fuzzy controer based contro mechanism showed better performance than P, PI, and PID controers based contro mechanisms. PSCAD/EMTDC program is used for modeing UPFC and taking the simuation resuts from the test system. In this study extent P, PID and fuzzy controers are added as a new component in main ibrary of PSCAD/EMTDC Qre (with UPFC) Qre (without UPFC) y (KVAr) y (K V A r) Vre (with UPFC) Vre (without UPFC).20 Vre (with UPFC) Vre (without UPFC) y (pu) y (p u ) time Fig. 8 PI controer resuts at faut condition 0.00 time Fig. 9 Fuzzy controer resuts at faut condition 72

6 Internationa Journa of Eectrica and Eectronics Engineering 3: APPENDIX The technica detais of sending end and receiving end generators: Based MVA (3-phase) : 0.0 [MVA] Base votage (L-L) : [kv] Base frequency : 50.0 [Hz] Phase : 0.0 [ ] The technica detais of shunt converter transformers: Based MVA (3-phase) Winding # votage (L-L) Winding #2 votage (L-L) Base operation frequency : [MVA] : [kv] : [kv] : 50.0 [Hz] The technica detais of series converter transformers: Based MVA (3-phase) Winding # votage (L-L) Winding #2 votage (L-L) Base operation frequency The system parameters: V dc C dc R r L r V dc votage controer PI parameters: : [MVA] : 0.0 [kv] : 0.0 [kv] : 50.0 [Hz] : [kv] : 20 mf : 4 Ω : 0 mh K p : 4.6 K i : Bus votage controer PI parameters: K p : 4.05 K i : Fixed oad parameters: Rated rea power per phase : [MW] Rated reactive power per phase : 0.00 [MVAR] Rated oad votage (rms L-G) : [kv] Fundamenta Frequency : 50 [Hz] REFERENCES [] E. Uzunovic, C. A. Canizares, J. Reeve, EMTP Studies of UPFC Power Osciation Damping, North American Power Symposium (NAPS), San Luis Obispo, Caifornia, October 999. [2] Y. H. Song, A. T. Jons, Fexibe AC Transmission Systems (FACTS), IEE Power and Energy Series 30, 999. [3] S. Hongbo, D.C Yu. Luo Chunei, A nove method of power fow anaysis with unified power fow controer (UPFC), Power Engineering Society Winter Meeting, IEEE Voume 4, Jan Page(s): vo.4 Digita Object Identifier 0.09/PESW [4] Narain G. Hingorani, Laszo Gyugyi Understanding FACTS: Concepts and Technoogy of Fexibe AC Transmission Systems, Power Eectronics Sponsored By, 2000 by the institute of Eectrica and Eectronics Engineers, Inc. 3 Park Avenue, New York, NY [5] A. A. Edamaty, S. O. Faried, S. Aboreshaid, Damping Power System Osciations Using a Fuzzy Logic Based Unified Power Fow Controer, Eectrica and Computer Engineering, 2005, pp.: [6] R.P Kayani, G.K. Venayagamoorthy, M. Crow, Neuroidentification of system parameters for the shunt & series branch contro of UPFC, Power Engineering Society Genera Meeting, 2003, IEEE Voume 4, 3-7 Juy 2003 Page(s): Digita Object Identifier 0.09/PES [7] S. Kannan, S. Jayaram, M. M. A. Saama, Rea and Reactive Power Coordination for a Unified Power Fow Controer, IEEE Transactions on Power Systems, Voume 9, Issue 3, 2004, pp [8] R. Orizondo, R. Aves, UPFC Simuation and Contro Using the ATP/EMTP and MATLAB/Simuink Programs, Transmission & Distribution Conference and Exposition, 2006, IEEE/PES, pp. 7. [9] D. Thukaram, L. Jenkins, K. Visakha, Improvement of System Security with Unified Power Fow Controer at Suitabe Locations under Network Contingencies of Interconnected Systems, IEE Proceedings Generation, Transmission and Distribution, Vo. 52, Issue 5, 2005, pp [0] H. Fujita, Y. Watanabe, H. Akagi, Contro and Anaysis of a Unified Power Fow Controer, IEEE Transactions on Power Eectronics, Voume 4, Issue 6, 999, pp [] B. Lu, L. Hou, B. Li, Y. Liu, A New Unified Power Fow Fuzzy Contro Method, Innovative Computing, Information and Contro, Second Internationa Conference, 2007, pp [2] Y. Qing, L. Norum, T. Undeand, S. Round, Investigation of Dynamic Controers for a Unified Power Fow Controer, Industria Eectronics, Contro, and Instrumentation, Vo. 3, 996, pp [3] M. W. Mustafa, A. A. Zin, A.F. Kadir, Steady State Anaysis of Power Transmission Using Unified Power Fow Controer, Transmission and Distribution Conference and Exhibition, Vo. 3, pp [4] Visua Power System Simuation, Web site avaiabe at [5] C.D. Schauder, L. Gyugyi, M.R. Lund, D.M. Hamai, T.R. Rietman, D.R. Torgerson, A. Edris, Operation of the unified power fow controer (UPFC) under practica constraints, Power Deivery, IEEE Transactions on Pubication Date: Apr 998 Voume: 3, Issue: 2 On page(s): ISSN: [6] L. Liu, P. Zhu, Y. Kang, J. Chen, Design and Dynamic Performance Anaysis of a Unified Power Fow Controer, Industria Eectronics Society, IECON st Annua Conference of IEEE Pubication Date: 6-0 Nov On page(s): 6 pp. - ISBN: INSPEC Accession Number: [7] S. Leva, A.P. Movando, Park s equations for distributed constants ine, Pubication Date: 5-8 Sep 999 Voume: 2, On page(s): vo.2 Meeting Date: 09/05/999-09/08/999 Location: Pafos, Cyprus ISBN: References Cited: 7. [8] T. S. Chung, Y. Xiaodong, D. Z. Fang, C. Y. Chung, Deveopment of Adaptive UPFC Suppementary Fuzzy Controer for Power System Stabiity Enhancement, Eectric Utiity Dereguation, Restructuring and Power Technoogies, Vo., 2004, pp [9] G. L. Sheng, W. Y. Lin, L. Sheng, Research on Fexibe Power Suppy System for Arc Furnace Based on UPFC, Industria Eectronics and Appications, 2007, pp [20] T. Takagi, M. Sugeno, Fuzzy Identification of Systems and Its Appications to Modeing and Contro. IEEE Transaction on Systems, Man and Sybern, Vo. 985, pp. : [2] Timothy J. Ross, Fuzzy Logic with engineering appication, McGraw- Hi, Inc