基于混沌神经网络的异步电主轴转速辨识

Electro-spindle speed identifier based on chaos-BP algorithm

针对BP神经网络易陷于局部最优,且学习速度慢的缺陷,研究了采用混沌优化算法改进BP神经网络并应用于异步电主轴转速辨识的方法。该方法使用新的混沌自映射函数取代一般的有限折叠次数的Logistic自映射函数,使其兼具了混沌优化算法的全局寻优与BP算法局部寻优的优点。借助MATLAB/Simulink软件对无传感器电主轴转速辨识系统进行数值仿真。仿真结果表明,采用新方法使整个系统的辨识性能更优,学习速度更快。

BP neural network is easy to plunge into local solution and has slow learning convergence speed,a new chaos optimization algorithm was investigated. A new chaotic self-map is applied to replace the Logistic chaotic self-map with finite collapse and it realized the combination of chaos＇ global search capability and BP＇s local optimize performance. The new algorithm applied in electro-spindle was carried out by simulation experiment using MATLAB/simulink. The numerical results show that speed identification system has not only the advantage of accurate identification, but also the virtue of quick learning convergence speed adopted the new algorithm.