4th Power Electronics, Drive Systems & Technologies Conference (PEDSTC2013), Feb l3-14, 2013, Tehran, Iran A Hybrid Adaptive Neural-Fuzzy Tuned P.I. Controller Based Unidirectional Boost P.F.C. Converter Feeds B.L.D.C. Drive S. Hr. A. Kaholi UMPEDAC U.M. K. L., Malaysia kaboli0004@gmail.com M. Mansouri Dept. ofElec.Eng. U.M. K. L., Malaysia mh.mansouri@gmail.com Abstract- The demand for solid state A.e.-D.e. converters to enhance power quality in the expression of P.F.e. (Power Factor Correction) impels the proposal of miscellaneous topologies. This paper presents a hybrid Neural-Fuzzy Controller (N.F.e.) tuned P.I. controller based Unidirectional Boost P.F.e. converter not only for regulating the drawn current, minimizing harmonics pollutions (that deteriorating the system power quality) and enhancing the circuit eiciency but also for B.L.D.e. (Brush Less D.e.) machine speed controlling. The paper irst presents a mathematical model and P.F.e. circuit; besides it depicts the way of hybrid controlling system evolves the controlling of drives performance at both transient and steady state conditions. It is important to be concerned that an integral robust S.S.E.E. (Steady State Error Eliminator) is paralleled to the system developing the controlling process wholly. Minimizing A.e. supply current distortion as close as to unity P.F. over a wide range of rotational speed is one of the advantages of this enhanced system. The presented modeled system contains Hybrid Neural-Fuzzy Controlling units, P.F.e. unit, unidirectional Boost Converter and B.L.D.e. machine; that simulated MA TLAB model is proposed and discussed in details. Simulation results signiicantly depict the impressiveness of the proposed controlling system in T.H.D. reduction. Kywords-Hybrid Neural-Fuzzy; P.l; Total Harmonic Distortion; Power Factor Correction; Brushless D.. motors;Boost Inverter. I. INTRODUCTION Servo robotic applications, dynamic actuators and machine tools are some conventional domination ield of B.L.D.C. motors. The motor preferred mechanical and elecrical characteristics; rising eiciency and minimized inertia momentum are the major factors make B.L.D.C. implementation such widespread [I], [3], [4]. Nowadays, emphasizing on the power quality improvement has been increased gradually causes realizing of presenting and enhancing new clean power converter topologies, meanwhile there is an increasing demand for high eiciency implementation of motor drives results in users looking forward a favorable performance of conrolling which should be provided meanwhile loads and motors parameter are varying during unction. By considering what mentioned above, it is inferred that there is an intense atractiveness in enhancing adaptive P.F.C. conroller that results in proposing various non- 978-1-4673-4484-5/13/$31.00 ©2013 IEEE 176 J. Selvaraj Dept. ofElec.Eng. UM. K. L., Malaysia N.B.A. Rahim Dept. ofElec.Eng. UM. K. L., Malaysia linear model based adaptive P.F.C. conrolling schemes dedicated for B.L.D.C. motors [I]. Generally, there are conventional unidirectional converters as buck, boost, buck-boost and multilevel [9] that supply variable D.C. power demand, although these converters endure poor P.F., utility injected hannonics and D.C. voltage link luctuation regarding to supply requency and voltage variation. A proper solution for direct A.C.-D.C. conversion to meet high density of power is prepared by implementing a single phase active power factor controller. The applied unidirectional boost converter is constituted rom a closed loop curent controller, is able to meet electricity authorities and agencies srict power factor penalty limits besides eliminating input line current hmonics distortion by applying active conrolling techniques as P.I.D., Fuzzy Logic, Neural Network, Adaptive Neural-Fuzzy or hybrid corolling method that lead to realizing power factor corection [8], [10]. P.I. (Proportional-Integral) controllers are one of the well known techniques applied in industrial conrolling process regarding to their ease of implementation and robust performance over a wide range of operational conditions. Gains of Proportional and Integral are the factors directly affect the design of the system; it is why designers spend a lot of time for optimizing the conroller parameters [8]. In nonlinear Boost Converter P.F.C. Inverter Fed B.L.D.C. system, linear conroller as P.I. is practically insuicient especially when a noble ransient perfonnance is desirable under all operational conditions. A dynamic gain regulation for Proportional and Integral is a solution that can solve the mentioned problem thoroughly by transfering controller to new operating points. The main obstacle for implementing this dynamic solution is providing a dynamic tuning process respectively [8]. One of the well-known tuning methods is the Ziegler-Nichols tuning formula. The technique is quite simple but is not always proper when parameters are varying besides this technique may not meet quick on-line tuning. It is also proved that B.L.D.C. machines are sensitive to parameters variations and load disturbances, those predominate hmul impacts in B.L.D.C. conrolling systems. A.N.F.I.S. (Adaptive Neural-Fuzzy Inference System) is named for Neural-uzzy network-based systems. This system is a type of Neural-network based on Takagi-Sugeno uzzy