D.C. DRIVE USING NEURAL NETWORK CONTROLLER

Abstract

This dissertation work is concerned with the control of speed of separately excited dc motor through application of neural network technique. The separately exited dc motor, fed from a two quadrant chopper (making use of MOSFET) is operated in PWM mode of operation through control of motor terminal voltage. The gating signals are generated through PC-AT interfaced with Timer and I/O cards. The neural network is used to identify the unknown input-output performance characteristics motor & load. The control objective is to design the software controller so that the motor follows a certain speed trajectory. The neural network is first trained by presenting it with a sequence of speed/ voltage patterns obtained from the experimental set-up and data file for NNW weights is stored for subsequent use. Experimentation is carried out to investigate to drive performance when subjected to a reference speed trajectory. The chopper duty cycle is determined in a closed loop manner using previously trained neural network. The weights of NNW are also on line trained The software is written in "C" Language.