基于PNN的舵机故障诊断方法研究  

Method for Fault Diagnosis of Steering Gear Based on Probabilistic Neural Networks

作　　者：周晶[1] 宋辉[1] 刘喜作[1] 

机构地区：[1]海军大连舰艇学院训练部模拟训练中心,辽宁大连116018

出　　处：《现代电子技术》2011年第18期80-82,共3页Modern Electronics Technique

摘　　要：随着科技的发展,舵机在不断提升性能的同时,其复杂性也大幅度提高,故障现象与故障本身并非简单的映射关系,设备故障诊断实质上是典型的复杂非线性分类问题。提出一种利用概率神经网络对舵机故障分类的方法,将从舵机振动信号中的提取特征值作为PNN的输入参数,构造出基于概率神经网络的舵机故障诊断方法。仿真结果显示,该网络工作稳定,运算速度快,对舵机故障分类准确率较高。Steering gear is one of the important parts of ship＇s control system, and has a crucial effect on safety. With the development of science and technology, while steering gear is constantly upgrading its performance, its complexity is wtstly increased. Since its fault phenomenon and fault itself are not simple mapping relation, the equipment fault diagnosis is the typical complex nonlinear classification problem in essence. A method of steering gear fault diagnosis based on probabilistic neural network （PNN） is proposed in this paper. The feature values extracted from steering gear vibration signals are taken as input parameters for PNN. Simulation results show that it works stably and quickly, and can perform high-accuracy classification of steering gear faults.

关 键 词：概率神经网络 故障诊断 贝叶斯网络 径向基函数 

分 类 号：TN911-34[电子电信—通信与信息系统]