交流变频调速系统神经网络控制器的研究

Study on Neural Network Controller for Frequency Conversion Alternative System

论述了基于神经网络控制技术的交流变频调速控制器的设计原理和实现方法。由计算机、采集板卡、D/A转换卡以及编码器等构成计算机硬件控制系统,由visualC++实现神经网络控制算法。利用神经网络的非线性映射能力、并行计算能力、自学习能力以及强鲁棒性等优点实现高品质高阶非线性交流调速的闭环响应特性。经实验验证,系统具有良好的动态跟踪能力,阶跃响应超调量小,鲁棒性好,稳定时间短。

This paper put forward a design principle and realization method for frequency conversion alternative motor controller based on neural networks control technology. The system composed with computer, collector,D/A converter, and encoder. Using VC++ language realize the control arithmetic. Non-linear mapping, parallel calculating, self-learning and strong robot capability of NN result in good close loop responses for highorder non-linear alternative motor. Test verified out the system has good dynamic tracing ability, its step respond get small magnitude, strong robustness and short steady time.