A Hybrid Neuro-Fuzzy - P.I. Speed Controller for B.L.D.C. Enriched with an Integral Steady State Error Eliminator Mahdi Mansouri University of Malaya Kuala Lumpur, Malaysia mh.mansouri@gmail.com S. Hr. Aghay Kaboli University of Malaya Kuala Lumpur, Malaysia kaboli0004@gmail.com Jalil Ahmadian University of Malaya Kuala Lumpur, Malaysia ahmadian.jalil@gmail.com Jeyraj Selvaraj University of Malaya Kuala Lumpur, Malaysia jeyraj@um.edu.my Abstract—This paper is subject to present a hybrid Neuro- Fuzzy (N.F.) - P.I. fed Controller for controlling the speed of B.L.D.C. (Brush Less D.C.) motors to evolve the drives controlling performance at both transient and steady state conditions by considering a paralleled robust integral S.S.E.E. (Steady State Error Eliminator) to enrich the whole controlling process. In the presented hybrid system, P.I. N.F.C. is the main controller loop while the paralleled integral S.S.E.E. controller reimburses the steady state errors. The presented B.L.D.C. drive contains the capabilities of quick tracking, small steady state error and high stability despite of all load and parameter variations. MATLAB simulation results depict the impressiveness of the presented controlling system. Index Terms—Nonlinear Control, Hybrid Neuro-Fuzzy, B.L.D.C. Speed Control, Integral Steady State Error Elimination. I. INTRODUCTION A B.L.D.C. machine is a synchronous machine accompanying with a P.M. (Permanent Magnet) in the rotor circuit. The windings of the armature are mounted on the stator switched electronically respecting to the rotor position. The B.L.D.C. motors are widely applied in robotic servo implementation, machine tools and dynamic actuators, based on their preferred electrical and mechanical features, efficiency increasing feature and the ability of lowering the inertia momentum [3]-[5] and [13]. The high degree of precision is not something imperious for the most electrical drives, although, a favorable controlling performance has to be prepared in high performance drive application even when the motor and load parameters are changing during the functioning. The constant gain controllers are the most conventional system implemented in high performance variable speed drives. Their prominent drawback is the poor performance when the load is nonlinear, parameters are changing and there are uncertainties. Consequently, the strategy of control in high performance electrical drives has to be robust and adaptive. When there is a demand for high performance implementation of drive, a favorable controlling performance has to be served while the motor and load parameters are changing during operating. It can be resulted that there is a vivid interest in developing adaptive controlling systems how different schemes of adaptive control for B.L.D.C. motors are proposed based on nonlinear models [6], and [7]. A.N.F.I.S. (Adaptive Neuro-Fuzzy Inference System) is a given name for neural fuzzy network-based systems. The aforementioned system has become the most desirable especially in the area of controlling nonlinear systems [8]. Neuro-Fuzzy controllers are conventionally improved with a P.I. [9], P.D. [10] or an adaptive controller [11] how a paralleled S.S.E.E. can be accompanied with to improve the system performance and efficiency. In this hybrid controlling system, a P.I. (Proportional Integral) N.F.C. is the main follower controller paralleled with an Integral S.S.E.E., which are used to track the exact control law. The in-use N.F.C. training algorithm in a direct adaptive control scheme is back-propagation algorithm and the trained N.F.C. is used considering the “Fuzzy Set Theory Based Control on a Phase Controlled Converter D.C. Machine Drive” [12]; although a smooth and simple activation mechanism is applied for the Integral S.S.E.E. to modify the law of control adaptively [1]. The performance of the represented speed control is evaluated under varying parameters and loads to show the efficiency and performance of the applied controlling system. II. A HYBRID N.F.- P.I. SPEED CONTROLLER FOR B.L.D.C. DRIVES The B.L.D.C. machine can be modelled as V=E+(R+jωl)I, how V and E are sinusoidal at the frequency of ω, R is the phase resistance and ωL is phase inductance (Fig. 1). Due to E = jωλ m when λ m stands for the linkage flux of windings of the stator per phase respecting to the permanent magnet. Still by assuming  اܮ and implementing a position feedback that keeps V and E (and henceforward I) in the same phase, the equation of the voltage can be simplified in the form of algebraic as: ൌ ܧ൅ ܫ(1) When substituting relations of ~ܧ ௥ and~ܫ, we result in: Fig. 1. Steady state per phase equivalent circuit of B.L.D.C. motors 2012 IEEE International Conference on Control System, Computing and Engineering, 23 - 25 Nov. 2012, Penang, Malaysia 978-1-4673-3143-2/12/$31.00 ©2012 IEEE 234