Fuzzy Logic Controller for Three Phase PWM AC-DC Converter

Transcription

1 Journl of Electrotechnology, Electricl Engineering nd Mngement (2017) Vol. 1, Number 1 Clusius Scientific Press, Cnd Fuzzy Logic Controller for Three Phse PWM AC-DC Converter Min Muhmmd Kml1,, Husn Ali2,b nd Bkht Muhmmd Khn3,c 1,2,3 Northwestern Polytechnicl University, Xi n, , P.R.Chin minkml@mil.nwpu.edu.cn, bhusn@mil.nwpu.edu.cn, cengineerbkht@yhoo.com Keywords: PWM converter, Fuzzy logic controller, Power fctor, Totl hrmonic distortion, Voltge regultion. Abstrct. This pper discusses control strtegy for the voltge regultion of three phse pulse width modultion (PWM) AC-DC converter. A fuzzy logic controller (FLC) is employed in the outer voltge loop. Fuzzy logic controllers re preferred over the conventionl proportionl integrl (PI) controllers becuse they re bsed on linguistic description nd do not require prior mthemticl model of the system. The objective of the presented control system is to regulte the output DC voltge during both trnsient nd stedy stte period. Also the line currents drwn from the grid should be sinusoidl nd in phse with respective phse voltges to chieve low totl hrmonic distortion (THD) nd unity power fctor. PI nd FLC bsed control systems re implemented for three phse PWM AC-DC converter in MATLAB/Simulink nd dynmic performnce of the system is tested. The results demonstrte tht compred to the conventionl PI, the FLC bsed system is more robust nd hs much better dynmic response. The low vlue of THD nd sinusoidl current response ensures the system voids polluting the utility grid. 1. Introduction The commonly employed three phse AC-DC converters in industril pplictions nd consumer products re bsed upon diode bridge circuit nd lrge storge cpcitor. Although this system hs the dvntge of being robust nd low cost, its drwbcks re unidirectionl power flow, low power fctor, nd high totl hrmonic distortion (THD) for input line current. To overcome these drwbcks pulse width modultion (PWM) AC-DC converters re used, which effectively regulte the output DC voltge nd hve low vlue of THD. The PWM AC-DC converters hve the cpbility of bi-directionl power flow, unit power fctor, low THD nd regulted DC voltge t the output [1-3]. Fig. 1 shows the topology of three phse PWM AC-DC converter. The converter is supplied by three phse voltge source connected vi L-filter. 1

2 S p S bp S cp i r i i b e L u e 0 b u b C udc e c i c u c R L S n Sbn S cn Fig. 1. Three-phse PWM AC-DC converter To regulte the opertion of three phse PWM AC-DC converters, proportionl integrl (PI) nd proportionl integrl derivtive (PID) controllers re commonly used. Although the conventionl PI nd PID controllers re reltively esy to implement, however their design is bsed upon n explicit mthemticl model of the system nd they lso lck in intelligence. This leds us to dptive controllers, i.e. fuzzy logic, neurl network, neuro fuzzy etc. These controllers do not require mthemticl model of the system prior to the control system design [4-5] nd lso hve the bility to hndle nonlinerity of the system. Among these fuzzy logic bsed controller is much robust nd lest sensitive to disturbnces [6]. In this study, fuzzy logic controller (FLC) is used to regulte the output voltge of three phse AC-DC converter. Fig. 2 shows the control system developed for three phse PWM converter which is composed of two inner current loops nd n outer DC voltge loop. The fuzzy logic controller is employed in the outer voltge loop for dc-bus voltge regultion. The FLC bsed voltge controller genertes reference current commnd fed to inner current loop. The inner loops employ PI bsed current controller. e bc i, i bc i q i q i d i d b L L i bc e, e bc e d FLC e q b L u d u q SPWM u bc bc Fig. 2. Fuzzy logic control scheme implementtion for three phse converter 2. Fuzzy Logic Controller The bsic rchitecture of FLC is shown in Fig. 3. Three min blocks of the FLC re fuzzifier, inference engine nd the defuzzifier. N C u ref u dc i r R L g 1 e v Δe v g 2 Fuzzifier Decision Rules Defuzzifier g o Output Fig. 3. Fuzzy logic controller block digrm. 2

3 The numericl vribles, error (e v ) nd chnge in error (Δe v ) re the two inputs pplied to the fuzzifier block of the FLC. Fuzzifier converts mesured quntities into fuzzy sets which re then given s input to the inference engine. In order to convert the pplied numericl vribles into linguistic vribles, five fuzzy sets re used from negtive big (NB) to positive big (PB). Among severl membership functions to define fuzzy sets, few commonly used re tringulr, gussin nd trpezoidl etc. In this work, tringulr membership functions re used s they re more effective to chieve better trnsient response [7]. The input to FLC is the difference of output voltge nd reference voltge (V ref -V o ), which is normlized so the universe of discourse is selected to be between -1 nd 1. The tringulr membership functions used re lbelled s NB, NS, ZE, PS nd PB for Negtive Big, Negtive Smll, Zero, Positive Smll nd Positive Big respectively. The x-xis represents the error nd chnge in error, while the y-xis shows the degree of membership µ. The second block i.e. inference engine is bsed upon set of if-then rules. The set of these rules enbles FLC to tke intelligent decisions [8]. In this study the inputs nd output re relted by set of 25 rules which re shown in Tble 1. Tble 1 Fuzzy Logic Rule Set Δe v / e v NB NS ZE PS PB NB NB NS NS ZE ZE NS NB NS NS ZE PS ZE NS NS ZE PS PS PS NS ZE PS PS PB PB ZE ZE PS PS PB Finlly the defuzzifier block converts the control output generted by the inference engine into the numericl vrible. 3. Results nd Discussion The PI nd FLC bsed PWM AC-DC converter system is simulted in MATLAB/Simulink softwre. The prmeters used for the simulted system re given in Tble 2. Tble 2 Simultion Prmeters Prmeter Input AC voltge (V rms ) Source frequency (V rms ) Filter inductor (L) Output filter cpcitor (C) Lod voltge (V o ) Active Power (P) Vlue 115 [V] 400 [Hz] [mh] 3300 [µf] 400 [V] 15k [W] Both the control systems re put under test by pplying resistive lod chnge. The lod resistor s vlue is chnged form 10.67ohm to 8.89ohm t t=0.4s which results in incresed current i.e. 37.5A to 45A. Fig. 4 shows the response of both PI n FLC bsed system for the pplied resistive lod chnge. It lso shows the ctive nd rective power flow wveforms. 3

4 Fig. 4. Output voltge nd Power response to resistive lod chnge Due to the resistive lod chnge the lod demnd lso chnges from 15kW to 18kW. The vlue of zero for rective power component verifies unity power fctor opertion. It cn be clerly observed tht the FLC bsed system responds much better to the pplied chnge. It hs less undershoot nd very short settling time. It is evident from the trnsient response tht fuzzy logic controller effectively compenstes the overll trnsient effects nd quickly restores the system to stedy stte condition. Fig. 5 shows the input voltge nd current wveform for one phse of the FLC bsed system. Fig. 5. Power fctor nd THD nlysis The input current wveform is in phse with the voltge wveform required for the system to hve unity power fctor. Fig. 5 lso shows the hrmonic nlysis plot of the input current wveform where the fundmentl component t 400Hz is clerly dominnt. The vlue of THD is computed to be 3.47%. The low vlue of THD vlidtes the sinusoidl current wveform nd is well within the limit of the hrmonic stndrd [9-10]. 4. Conclusion This pper discusses the control system for three phse pulse width modultion AC-DC converter. The conventionl PI nd fuzzy logic controllers re designed nd implemented in the outer voltge loop. The performnce of the system is evluted for both controllers by pplying certin lod chnge. The results show tht fuzzy logic bsed controller is more efficient to regulte the output voltge compred to the PI bsed controller. The fuzzy logic bsed system hs sinusoidl wveform for input current which is lso in phse with the corresponding phse voltge to ensure low vlue of THD nd unity power fctor. 4

5 References [1] K. I. Hwu, nd T. J. Peng, A Novel Buck Boost Converter Combining KY nd Buck Converters, IEEE Trnsctions on Power Electronics, Vol. 27, No. 5, My [2] A. Choui, J. P. Gubert, F. Krim nd G. Chmpenois PI Controlled Three-phse Shunt Active Power Filter for Power Qulity Improvement - Electric Power Components nd Systems, 35:12, pp , [3] A. Semmh, A. Mssoum nd Ptrice Wir Improvement of PWM Rectifier Performnces Using Direct Power Control nd Fuzzy Logic Theory, Austrlin Journl of Bsic nd Applied Sciences, 7(4): , [4] S. H. Zk, Systems nd control: Oxford University Press New York, [5] Y Liyong, L.Y inghong, C Yi nd L.Zhengxi, A Novel Fuzzy Logic Controller for Indirect Vector Control Induction Motor Drive, Proceedings of the IEEE 7th World Congress on Intelligent Control nd Automtion, Chongqing, Chin, pp.24-28, June [6] J. H. Lilly, Fuzzy control nd identifiction: John Wiley & Sons, [7] D. Q. Dng, Y. S. Choi, H. H. Choi, nd J. W. Jung, Experimentl Vlidtion of Fuzzy Adptive Voltge Controller for Three-Phse PWM Inverter of Stndlone DG Unit, IEEE Trnsctions on Industril Informtics, vol. 11, pp , [8] K. Tnk nd M. Sugeno, Stbility nlysis nd design of fuzzy control systems, Fuzzy Sets Syst., vol. 45, pp , [9] F. Botterón nd H. Pinheiro, A three-phse UPS tht complies with the stndrd IEC , IEEE Trnsctions on Industril Electronics, vol. 54, pp , [10] Limits for Hrmonic Current Emissions (Equipment Input Current up to nd Including 16A Per Phse), IEC Interntionl Stndrd,