基于遗传算法优化的RBF神经网络的压力传感器故障诊断  

Fault Diagnosis of Pressure Sensor Based on RBF Neural Network Optimized by Genetic Algorithm

作　　者：那文波[1] 何宁[1] 刘巍[1] 刘甜甜[1] 

机构地区：[1]中国计量大学机电工程学院,杭州310018

出　　处：《煤矿机械》2016年第7期180-183,共4页Coal Mine Machinery

摘　　要：提出了一种基于遗传算法优化的RBF神经网络故障诊断方法,此诊断方法首先利用小波包分解的方法对传感器电路的各种故障数据进行特征提取,再将所得的特征向量输入到遗传算法优化的RBF神经网络进行故障诊断。经过MATLAB仿真,此故障诊断方法不仅对传感器故障进行准确分类,而且提高了BP神经网络的收敛性,从而验证了此方法是可行的。Put forward a kind of based on genetic algorithm to optimize the RBF neural network fault diagnosis method, the method first, wavelet packet decomposition, normalized as a preprocessing to extract the fault feature vectors of sensor circuit, fault feature vector is input to the genetic algorithm to optimize the RBF neural network through network fault diagnosis. Through MATLAB simulation results show that after the RBF neural network optimized by genetic algorithm not only improves the learning convergence speed of BP neural network, but also improves the accuracy of sensor fault recognition,and prove that this method is correct and effective.

关 键 词：传感器 遗传算法 RBF神经网络 故障诊断 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]