Neural Network Pid Control In The Automatic Test System For The Automobile Generator Performance

Auto-alternator is one of the main electrical installations in cars, and all its' performance parameters to be qualified is very important to cars' safety. Performance test is the last and the most important working procedure on the alternator's product line. Products' quality and production's efficiency are connected with the level of performance test.With the enlargement of the production capacity, the development of the technology, the enhancement of the type of products, and clients' requirement on products' quality, many shortages are found in the forepassed servomechanisms on test precision, speed and versatility. Because when made, they were restricted by many factors such as computer's small capability and slow speed, measuring device's few interface function,controller's no self-adaption.All those become obstacles of the development of production's scale and products' quality. Therefor, Auto-Alternator Performance Automatic Testing Equipment is developed.In this paper, based on analysis of operating principle of the alternator, the whole design proposal is introduced. To design a good kind of closed loop controller is the main idea of this paper. To strengthen self-adaption of the system, an adaptive controller is designed by combining neural network with normal PID control, instead of the single normal PID control. Firstly,application of BP neural network to PID control is presented. Secondly, a PID controller is proposed based on improved BP algorithm. Finally, application of the proposed PID controller to the system is investigated.Around speed and precision, the system is designed successfully, for its advanced hardware, convenient operation, perfect function and strong interaction. And now it runs stably on the alternator's product line.