During Damping of Low Frequency Oscillations in Power Systems with Fuzzy UPFC Controller


Published on


1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

During Damping of Low Frequency Oscillations in Power Systems with Fuzzy UPFC Controller

  1. 1. International Journal of Modern Engineering Research (IJMER) www.ijmer.com Vol.2, Issue.6, Nov-Dec. 2012 pp-4215-4219 ISSN: 2249-6645During Damping of Low Frequency Oscillations in Power Systems with Fuzzy UPFC Controller M.Bala Ankanna, 1 K.Harinath Reddy, 2 M.Tech (EPE), Department of EEE, AITS, Rajampet, A.P, Assistant Professor, Department of EEE, AITS, Rajampet, A.P,Abstract: Low Frequency Oscillations (LFO) occur in components. Devices for high power levels have been madepower systems because of lack of the damping torque in available in converters for high and even highest voltageorder to dominance to power system disturbances as an levels. The growth of the demand for electrical energy leadsexample of change in mechanical input power. In the recent to loading the transmission system near their limits. Thus,past Power System Stabilizer (PSS) was used to damp LFO. the occurrence of the LFO has increased. FACTsFACTs devices, such as Unified Power Flow Controller Controllers has capability to control network conditions(UPFC), can control power flow, reduce sub-synchronous quickly and this feature of FACTs can be used to improveresonance and increase transient stability. So UPFC may power system stability. The UPFC is a FACTS device thatbe used to damp LFO instead of PSS. UPFC damps LFO can be used to the LFO. The primarily use of UPFC is tothrough direct control of voltage and power. A control the power flow in power systems. The UPFCcomprehensive and systematic approach for mathematical consists of two voltage source converters (VSC) each ofmodelling of UPFC for steady-state and small signal them has two control parameters namely me, δe ,mb and δb(linearize) dynamic studies has been proposed. The other [3]. The UPFC used for power flow control, enhancementmodified linearize Heffron-Philips model of a power system of transient stability, mitigation of system oscillations andinstalled with UPFC is presented. For systems which are voltage regulation [3]. A comprehensive and systematicwithout power system stabilizer (PSS), excellent damping approach for mathematical modelling of UPFC for steady-can be achieved via proper controller design for UPFC state and small signal (linearized) dynamic studies has beenparameters. By designing a suitable UPFC controller, an proposed in [4-7]. The other modified linearized Heffron-effective damping can be achieved. In this research the Philips model of a power system installed with UPFC islinearize model of synchronous machine (Heffron-Philips) presented in [8] and [9]. For systems which are withoutconnected to infinite bus (Single Machine-Infinite Bus: power system stabilizer (PSS), excellent damping can beSMIB) with UPFC is used and also in order to damp LFO, achieved via proper controller design for UPFC parameters.adaptive neuro-fuzzy controller for UPFC is designed and By designing a suitable UPFC controller, an effectivesimulated. Simulation is performed for various types of damping can be achieved. It is usual that Heffron-Philipsloads and for different disturbances. Simulation results model is used in power system to study small signalshow good performance of neuro-fuzzy controller in stability. This model has been used for many yearsdamping LFO. providing reliable results [10]. In recent years, the study of UPFC control methods has attracted attentions so that different control approaches are presented for UPFC control Keywords: Neuro-Fuzzy Controller, Low Frequency such as Fuzzy control [11-14], conventional lead-lagOscillations (LFO), Unified Power Flow Controller control [15], Genetic algorithm approach [16], and Robust(UPFC), Single Machine-Infinite Bus (SMIB) control methods [17-19]. In this case study, the class of adaptive networks I. INTRODUCTION that of the same as fuzzy inference system in terms of Flexible AC Transmission Systems, called FACTS, performance is used. The controller utilized with the abovegot in the recent years a well known term for higher structure is called Adaptive Neuro Fuzzy Inference Systemcontrollability in power systems by means of power or briefly ANFIS [20]. Applying neural networks has manyelectronic devices. Several FACTS-devices have been advantages such as the ability of adapting to changes, faultintroduced for various applications worldwide. A number of tolerance capability, recovery capability, High-speednew types of devices are in the stage of being introduced in processing because of parallel processing and ability topractice. In most of the applications the controllability is build a DSP chip with VLSI Technology. To showused to avoid cost intensive or landscape requiring performance of the designed adaptive neuro-fuzzyextensions of power systems, for instance like upgrades or controller, a conventional lead-lag controller that isadditions of substations and power lines. FACTS-devices designed in [15] is used and the simulation results for theprovide a better adaptation to varying operational power system including these two controllers are comparedconditions and improve the usage of existing installations. with each other.The basic applications of FACTS-devices are: (i).Powerflow control, (ii) Increase of transmission capability, (iii)Voltage control, (iv)Reactive power compensation, II. MODEL OF THE POWER SYSTEM(v)Stability improvement, (vi) Power quality improvement, INCLUDING UPFC(vii)Power conditioning, (viii)Flicker mitigation, UPFC is one of the famous FACTs devices that is(ix)Interconnection of renewable and distributed generation used to improve power system stability. Fig.1 shows aand storages. The development of FACTS-devices has single machine- infinite-bus (SMIB) system with UPFC. Itstarted with the growing capabilities of power electronic is assumed that the UPFC performance is based on pulse www.ijmer.com 4215 | Page
  2. 2. International Journal of Modern Engineering Research (IJMER) www.ijmer.com Vol.2, Issue.6, Nov-Dec. 2012 pp-4215-4219 ISSN: 2249-6645width modulation (PWM) converters. In figure 1 me, mb and each inverter can independently generate (or absorb)δe, δb are the amplitude modulation ratio and phase angle of reactive power at its own ac output terminal.the reference voltage of each voltage source converter Inverter 2 provides the main function of the UPFCrespectively. These values are the input control signals of by injecting an ac voltage Vpq with controllable magnitudethe UPFC. Vpq (0≤Vpq≤Vpqmax) and phase angle (0≤≤360), at the power frequency, in series with the line via an insertion transformer. The injected voltage is considered essentially as a synchronous voltage source. The transmission line current flows through this voltage source resulting in real and reactive power exchange between it and the ac system. The real power exchanged at the ac terminal (i.e., at the terminal of insertion transformer) is converted by the inverter into dc power that appears at the dc link as positive or negative real power demanded. The reactive power exchanged at the ac terminal is generated internally by theFig.1 A single machine connected to infinite bus with inverter. The basic function of inverter 1 is to supply orUPFC absorb the real power demanded by Inverter 2 at the common dc link. This dc link power is converted back to acAs it mentioned previously, a linearized model of the power and coupled to the transmission line via a shunt-connectedsystem is used in dynamic studies of power system. In order transformer. Inverter 1 can also generate or absorbto consider the effect of UPFC in damping of LFO, the controllable reactive power, if it is desired, and there by itdynamic model of the UPFC is employed; In this model the can provide independent shunt reactive compensation forresistance and transient of the transformers of the UPFC can the line.be ignored. The Linearized state variable equations of the It is important to note that where as there is apower system equipped with the UPFC can be represented closed “direct” path for the real power negotiated by theas [8]. action of series voltage injection through Inverters 1 and 2 back to the line, the corresponding reactive power exchanged is supplied or absorbed locally by inverter 2 and therefore it does not flow through the line. Thus, Inverter 1 can be operated at a unity power factor or be controlled to have a reactive power exchange with the line independently of the reactive power exchanged by the Inverter 2. This means there is no continuous reactive power flow through UPFC. Operation of the UPFC from the standpoint of conventional power transmission based on reactive shunt compensation, series compensation, and phase shifting, the UPFC can fulfill these functions and thereby meet multipleWhere ΔmE, ΔmB , ΔδE and ΔδB are the deviation of input control objectives by adding the injected voltage Vpq, withcontrol signals of the UPFC. Also in this study IEEE Type- appropriate amplitude.ST1A excitation system was used. III. OPERATING PRINCIPLE OF UPFC In the presently used practical implementation,The UPFC consists of two switching converters, which inthe implementations considered are voltage source invertersusing gate turn-off (GTO) thyristor valves, as illustrated inthe Fig 2.1. These back to back converters labeled “Inverter1 and “Inverter 2” in the figure are operated from a Fig2.Principle configuration of an UPFCcommon dc link provided by a dc storage capacitor. The UPFC consists of a shunt and a series transformer, which are connected via two voltage source converters with a common DC-capacitor. The DC-circuit allows the active power exchange between shunt and series transformer to control the phase shift of the series voltage. This setup, as shown in Figure 1.21, provides the full controllability for voltage and power flow. The series Fig. Basic circuit arrangement of unified power flow converter needs to be protected with a Thyristor bridge. controller Due to the high efforts for the Voltage Source Converters and the protection, an UPFC is getting quite expensive, This arrangement functions as an ac to ac power which limits the practical applications where the voltageconverter in which the real power can freely flow in either and power flow control is required simultaneously.direction between the ac terminals of the two inverters and www.ijmer.com 4216 | Page
  3. 3. International Journal of Modern Engineering Research (IJMER) www.ijmer.com Vol.2, Issue.6, Nov-Dec. 2012 pp-4215-4219 ISSN: 2249-6645 IV. CONTROLLER DESIGN combination of linear and nonlinear parameters learningI. Lead-Lag Controller Design algorithm. Description for learning procedure can be found As mentioned before, in this study two different in [20]. This network is called adaptive by Jang and it iscontrollers have been used to damp LFO. The first one is functionally equivalent to Sugeno type of a fuzzy system. Itconventional lead-lag controller. It consists of gain block, is not a unique presentation.washout block, lead-lag compensator block. The washoutblock is considered as a high-pass filter, with the timeconstant TW. Without this block steady changes in inputwould modify the output. The value of TW is not criticaland may be in the range of 1 to 20 seconds. In this study,the parameters obtained from lead-lag controller design thatis presented in [15], were used.II. Adaptive Neuro-Fuzzy Controller Design Figure.2. A fuzzy system defined by a neural network. Another controller is adaptive neuro-fuzzycontroller. In this section, we will present the procedure of With regard to the explanations presented and with the helpdesigning of the adaptive neuro-fuzzy controller. In this of MATLAB software, adaptive neuro-fuzzy controller canresearch, the neuro fuzzy controller has 2 inputs that are Δδ be designed. The rules surface for designed controller isand Δω and it has 1 output that is f ∈ {∆𝑚E, ∆𝛿 E, ∆𝑚B, ∆𝛿 B shown in figure 3.}. For each input 20 membership functions and also 20rules in the rules base is considered. Figure 5 demonstratesthe structure of adaptive neuro-fuzzy controller for a sugenofuzzy model with 2 inputs and 20 rules [20]. Fig.3 The rules surface The membership functions for input variable ∆𝜔 are Fig.3 ANFIS architecture for a two-input Sugeno fuzzy presented in figure 4. model with 20 rules In Figure 3, a Sugeno type of fuzzy system has the rulebase with rules such as follows: 1. If Δδ is A1 and Δω is B1 then f1=p1 Δδ+q1 Δω+r1. 2. If Δδ is A2 and Δω is B2 then f2=p2 Δδ+q2 Δω+r2. 𝜇 𝐴𝑖  and 𝜇 𝐵 𝑖 are the membership functions of fuzzy sets Aiand Bi for i=1,…,20. In evaluating the rules, we chooseproduct for T-norm (logical and). Then controller could bedesigned in following steps:1. Evaluating the rule premises: 𝑤 𝑖 = 𝜇 𝐴 𝑖 (Δ𝛿) 𝜇 𝐵 𝑖 (∆𝜔), i= 1.....20 (2) Fig.4 The membership functions for input variable ∆𝜔2.Evaluating the implication and the rule consequences: 𝑤 1 ∆𝛿,∆𝜔 𝑓1 (∆𝛿,∆𝜔 )+⋯+𝑤 20 (∆𝛿 ,∆𝜔 )𝑓 20 (∆𝛿 ,∆𝜔 ) One of the advantages of using neuro-fuzzy controller isf(∆𝛿, ∆𝜔) = 𝑤 1(∆𝛿 ,∆𝜔 ) +⋯+𝑤 20(∆𝛿 ,∆𝜔 ) that we can utilize one of the designed controllers forOr leaving the arguments out instance Δme controller in place of the other controllers. 𝑤 1 𝑓1 +⋯+𝑤 20 𝑓20 While if we use conventional lead-lag controllers, for each f = 𝑤 1 +⋯+𝑤 20 (3) control parameters, a controller must be designed.This can be separated to phases by first defining 𝑤𝑖 V. SIMULATION RESULTS w¯i = (4) In this research, two different cases are studied. 𝑤 1 +⋯+𝑤 20 In the first case mechanical power and in the second caseThese are called normalized firing strengths. Then f can be reference voltage has step change and deviation in ω (Δω)written as and deviation in rotor angle δ (Δδ) is observed. The f = w¯1f1+......+ w20f20 (5) parameter values of system are gathered in Appendix. In The above relation is linear with respect to pi , qi, ri first case, step change in mechanical input power is studied.and i=1,…,20. So parameters can be categorized into 2 sets: Simulations are performed when mechanical input powerset of linear parameters and set of nonlinear parameters. has 10% increase (ΔPm=0.1 pu) at t=1 s. Simulation resultsNow Hybrid learning algorithm can be applied to obtain for different types of loads and controllers ( ΔmE, ΔδE,values of parameters. Hybrid learning algorithm is www.ijmer.com 4217 | Page
  4. 4. International Journal of Modern Engineering Research (IJMER) www.ijmer.com Vol.2, Issue.6, Nov-Dec. 2012 pp-4215-4219 ISSN: 2249-6645ΔmB, ΔδB ) and step change in mechanical input power are demonstrates simulation result for step change in referenceshown in figures 5 to 9. voltage, under nominal load and for δb Controller.Generator Parameters :M = 2H = 8.0MJ /MVA , D = 0.0, Td0’ = 5.044SXd = 1.0pu , Xq = 0.6 , Xd „ = 0.3Exciter (IEEE Type ST1): 𝑘 𝐴 = 100, TA = 0.01SReactances: XIE = 0.1 pu , XE = XB =0.1 pu, XBv = 0.3 pu, XE = 0.5 puOperation Condition: Pe = 0.8 pu, Vt = Vb=1.0 puUPFC parameters: mE = 0.4013, mB = 0.0789,δE = -85.3478° , δB = -78.2174°DC link: Vdc = 2 pu, Cdc = 1 pu Fig. 10 Response of angular velocity for 5% step change in reference voltage in the case of nominal load (δb Controller) Consequently simulation results show that neuro- fuzzy controller successfully increases damping rate and decreases the amplitude of low frequency oscillations. Results comparison between conventional lead-lag compensator and the proposed neuro-fuzzy controller for the UPFC indicates that the proposed neuro-fuzzy controller has less settling time and less overshoot in comparison with Fig.7 Angular velocity deviation during step change in the conventional lead-lag compensator. mechanical input power for nominal load (δe Controller) VI. CONCLUSIONS With regard to UPFC capability in transient stability improvement and damping LFO of power systems, an adaptive neuro-fuzzy controller for UPFC was presented in this paper. The controller was designed for a single machine infinite bus system. Then simulation results for the system including neurofuzzy controller were compared with simulation results for the system including conventional lead-lag controller. Simulations were performed for different kinds of loads. Comparison. showed that the proposed adaptive neuro-fuzzy controller has good ability Fig.8 Angular velocity deviation during step change in to reduce settling time and reduce amplitude of LFO. Also mechanical input power for nominal load (mb Controller) we can utilize advantages of neural networks such as the ability of adapting to changes, fault tolerance capability, recovery capability, High-speed processing because of parallel processing and ability to build a DSP chip with VLSI Technology. Fuzzy logic is a convenient way to map an input space to an output space. Mapping input to output is the starting point for everything. REFERENCES [1] N. G. Hingorani and L. Gyugyi, Understanding FACTS: Concepts and Technology of Flexible AC Transmission System, IEEE Press, 2000. Fig.9 Angular velocity deviation during step change in [2] H.F.Wang, F.J.Swift," A Unified Model for the mechanical input power for nominal load (δb Controller) Analysis of FACTS Devices in Damping Power System Oscillations Part I: Single-machine Infinite- As it can be seen from figures 5 to 9, lead-lag bus Power Systems", IEEE Transactions on Powercontroller response is not as good as neuro-fuzzy controller Delivery, Vol. 12, No. 2, April 1997, pp.941-946.response and also neuro-fuzzy controller decreases settling [3] L. Gyugyi, C.D. Schauder, S.L. Williams,time. In addition maximum overshoot has decreased in T.R.Rietman, D.R. Torgerson, A. Edris, "The Unifiedcomparison with lead-lag controller response. In second Power Flow Controller: A New Approach tocase, simulations were performed when reference voltage PowerTransmission Control", IEEE Trans., 1995, pp.has 5% increase (ΔVref = 0.05 pu) at t=1 s. Figure 10 1085-1097. www.ijmer.com 4218 | Page
  5. 5. International Journal of Modern Engineering Research (IJMER) www.ijmer.com Vol.2, Issue.6, Nov-Dec. 2012 pp-4215-4219 ISSN: 2249-6645[4] Wolanki, F. D. Galiana, D. McGillis and G. Joos, 13] A. Oudalov, R. Cherkaoui, A.J. Germond, “Mid-Point Sitting of FACTS Devices in "Application of fuzzy logic techniques for the Transmission Lines,” IEEE Transactions on Power coordinated power flow control by multiple series Delivery, vol. 12, No. 4, 1997, pp. 1717-1722. FACTS devices", IEEE Power Engineering Society[5] M. Noroozian, L. Angquist, M. Ghandari, and G. International Conference, 2001, pp. 74 – 80. Anderson, “Use of UPFC for optimal power flow [14] Karbalaye Zadeh, M. Khatami, V. Lesani, H. Ravaghi, control”, IEEE Trans. on Power Systems, vol. 12, no. H. A., " A fuzzy control strategy to damp multi mode 4, 1997, pp. 1629–1634. oscillations of power system considering UPFC", 8th[6] A Nabavi-Niaki and M R Iravani. „Steady-state and International Conference on Advances in Power Dynamic Models of Unified Power Flow Controller System Control, Operation and Management, 2009, (UPFC) for Power System Studies.‟ IEEE p.1. Transactions on Power Systems, vol 11, 1996, p 1937. [15] N. Tambey, M.L. Kothari, Damping of power system[7] K S Smith, L Ran, J Penman. „Dynamic Modelling of oscillations with unified power flow controller a Unified Power Flow Controller.‟ IEE Proceedings- (UPFC), IEE Proc.-Gener. Transm. Distrib., Vol. 150, C, vol 144, 1997, pp.7. No. 2, March 200.[8] H F Wang. „Damping Function of Unified Power [16] Khan, L. Ahmed, N. Lozano, C., "GA neuro-fuzzy Flow Controller.‟ IEE Proceedings-C, vol 146, no 1, damping control system for UPFC to enhance power January 1999, p 81. system transient stability", 7th International Multi[9] H. F. Wang, F. J. Swift, “A Unified Model for the Topic Conf., 2003., pp. 276 – 282. Analysis of FACTS Devices in Damping Power [17] Ruban Deva Prakash, T. Kesavan Nair, N., "A Robust System Oscillations Part I: Single-machine Infinite- Control Strategy for UPFC to Improve Transient bus Power Systems,” IEEE Transactions on Power Stability Using Fuzzy Bang-Bang Control", Delivery, Vol. 12, No. 2, April, 1997, pp. 941-946. International Conference on Computational[10] P.Kundur,"Power System Stability and Control”, Intelligence and Multimedia Applications, 2007, pp. McGraw-Hill. 352 – 356.[11] M. Banejad, A. M. Dejamkhooy , N. Talebi, " Fuzzy [18] Jang-Cheol Seo Seung-Il Moon Jong-Keun Park Jong- Logic Based UPFC Controller for Damping Low Woong Choe, " Design of a robust UPFC controller Frequency Oscillations of Power Systems", 2nd IEEE for enhancing the small signal stability in the multi- International Conference on Power and Energy, 2008, machine power systems", IEEE Power Engineering pp. 85-88. Society Winter Meeting, 2001, pp. 352-356.[12] P.K., Dash, S. Mishra, G. Panda, , "Damping [19] Rahim, A.H.M.A. Al-Baiyat, S.A., "A robust damping multimodal power system oscillation using a hybrid controller design for a unified power flow controller", fuzzy controller for series connected FACTS devices", 39th International Universities Power Engineering IEEE Trans. on Power Systems, Vol. 15, 2000, pp. Conference, 2004, pp. 265 – 269. 1360 -1366. [20] Jyh-Shing Roger Jang, Chuen-Tsai Sun, Eiji Mizutani, " Neuro-fuzzy and soft computing: a computational approach to learning and machine intelligence", Prentice-Hall, 1997. . www.ijmer.com 4219 | Page