DESIGN AND IMPLEMENTATION OF A NEURAL CONTROLLER FOR PHASE- CONTROLLED DC MOTOR DRIVE

Abstract

The dc motors have occupied a wide spectrum of applications for variables speed drive. The present work deals with .the . design and development of concverter fed dc motor drive using fuzzy neural network control. A 3-phase thyristorized phase controlled converter is designed and devleoped. The firing pulses for the thyristors are generated using 8031 microcontroller based card. The firing angle at which the converter is to be operated is transmitted serially from personal computer at 2400 baud rate to microcontroller card. For precise control over-wide speed range and to get improved performance of dc motor drive, closed loop operation of system is desired. For closed loop operation speed measurement and error processing at much faster rate is required. This necessitates an accurate and fast acting control of speed. A pulse tachogenerator is used for accurate measurement of speed.. The motor current is sensed 'using Hall effect current sensor. The conventional control (PID) is slow and lack of 'efficiency in handling system non-linearities. To overcome* these limitations fuzzy neural controller has designed. and tested. The basic concept behind hybrid controller is that to get the advantages of both fuzzy logic and neural network. Since the machine system is ill defined fuzzy logic is used forpattern generation and then the ability of data processing in a non-linear system of neural network is used. For generating patterns from fuzzy logic triangular membership function is used. The method of height is used for defuzzification. Then 'neural network is trained off-line using these patterns. Error back propagation algorithm is used for _training and adjusting the weights of controller. The performance of the converter fed dc motor drive is investigated in open loop and closed loop mode.