一种基于RBF神经网络的转台分系统故障诊断方法  

Turntable subsystem fault-diagnosing approach based on RBF neural networks

作　　者：付强[1] 

机构地区：[1]哈尔滨工业大学控制科学与工程系,黑龙江哈尔滨150001

出　　处：《传感器与微系统》2007年第6期26-28,32,共4页Transducer and Microsystem Technologies

摘　　要：针对三轴精密测试转台各分系统故障诊断的需要,提出了一种基于径向基函数(RBF)神经网络的局部故障诊断方法。首先,给出了相应的RBF神经网络的结构,以及一种基于递归最小二乘法的改进学习算法;然后,将其应用到转台控制分系统的局部故障诊断中。根据控制分系统的常见故障及其特征信息,建立起基于RBF神经网络的故障诊断模型;最后,仿真实验结果验证了该方法的有效性。A local fault diagnosing approach based on RBF neural networks is presented for the requirement of subsystem fault diagnosis of 3-axis precision test turntable. First, the corresponding structure of the RBF neural networks is given, together with a modified learning algorithm based on recursive LSM. Then it is applied to the local fault diagnosing of turntable control subsystem. Fault diagnosing model based on RBF neural networks is built according to common faults and its characteristic information in control subsystem. Last, computer simulation results verify the effectiveness of this approach.

关 键 词：局部故障诊断 径向基函数神经网络 递推最小二乘法 转台分系统 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]