ROBUST FUZZY-PID CONTROL OF THREE-MOTOR DRIVE SYSTEM USING SIMULATED ANNEALING OPTIMIZATION Fawzan SALEM E.H.E. BAYOUMI Power Electronics and Energy Conversion Department, Electronics Research Institute (ERI), Cairo, Egypt E-mail: fawzan@lycos.com Abstract: Multi-motor drive system is a multi-input multi- output, nonlinear and strong-coupling system. In turn to improve synchronous performance of multi-motor system, this paper presents a synchronous control system for three-induction motors. It includes three robust Fuzzy-PID controllers based Simulated Annealing Optimization. The parameters of Fuzzy-PID controllers are adjusted for fulfilling the open loop control of multi-variable system to reduce mutual coupling effect. The Simulating Annealing Optimization technique is used to adjust the input and output scaling factors by minimizing both integral of absolute of errors and integral of time multiplied absolute of errors as performance measures. To test the robustness of the proposed system, several sudden changes are presented. The results indicate that the control system can acquire the optimal parameters of the Fuzzy-PID controllers according to different running states of the system. Results of system performance endorse the proposed technique and emphasize its feasibility. Keywords: three-motor drive system, Fuzzy-PID controllers, Simulated Annealing Optimization. 1. Introduction Due to its high precision coordinated control performance, a multi-motor system has attracted more and more attentions in the drive applications such as urban rail transit, paper making, electric vehicle drive, and steel rolling [1-5]. The AC multi- motor drive system is a multi-input multi-output (MIMO), nonlinear and strong-coupling system [6, 7]. Thus, its accurate mathematical model is hard to obtain. Meanwhile, industrial production also requires the multi-motor drive control system to decouple the speed and tension, which increases the control difficulty. Nowadays, common decoupling control methods mainly contain: improved control algorithm based on the traditional PID [8], cross-coupled control [9], feed forward control [10], optimal control [11], sliding mode control [12, 13], BP Neural Network [14], and fuzzy control [15]. To a certain degree, these methods have improved coordinated control performance. However, it should be noted that most of the existing methods depend on dynamic model of the motor and traditional single motor drive system. Fuzzy-PID control, as one of the promising intelligent control techniques, is applied for multi- motor synchronous control purposes. The authors in [16] proposed a variable gain intelligent Fuzzy reasoning control scheme in the stretch tension and synchronization control system of a wide-fabric heating-shaping machine. In [17] a synchronization control strategy of multi-motor system based on PROFIBUS network was proposed using BP Neural Network and an adaptive double mode Fuzzy-PID arithmetic. Fuzzy-PID C-means clustering algorithm [18] was applied to two-motor variable frequency speed-regulating synchronous system as a multivariable nonlinear coupling system. It is used to cluster the data of input-output according to a satisfactory performance index in order to identify speed and tension for the two-motor synchronous system based on local model networks. Two different Fuzzy-PID decoupling controllers in the speed feedback and the tension feedback [19] introduced greatly improved performance. It should be noted that most of these researches consider only two- motor synchronous drive system. There are several researches in the literature employed Simulated Annealing Optimization (SAO) in Fuzzy-PID controllers. The authors in [20] proposed a general technique for optimizing fuzzy models in fuzzy expert systems (FES’s) by simulated annealing (SA) and N-dimensional simplex method. In [21], Fuzzy-PID control is combined with BP Neural Network with a Genetic Algorithms (GA) and SAO to short-term load forecasting. An optimizing