人工神经网络故障诊断系统可靠性和聚类能力的分析

Analysis of Reliability and Assembility of Artificial Neural Network Fault Diagnosis System

介绍了将人工神经网络自组织特征映射模型———KOHONEN网络运用到电厂送风机故障诊断的思路和方法 ,并进一步讨论了影响诊断系统可靠性和聚类能力的主要因素 ,如隶属度函数和模型中参数的设置等 ,分析结果表明

This paper introduces a method, using the KOHONEN self organization reflection model of the artificial neural network system, to diagnose the faults of feed fans in the power plants, and also analyzes the primary factors that affect the reliability and assembility of the fault diagnosis system. The result shows that the diagnosis system has a high reliability and strong assembility.