基于机器学习的电缆故障诊断知识库设计方法

Design Method of Cable Fault Diagnosis Knowledge Base Based on Machine Learning

针对电缆故障在诊断过程中采集的故障信号量庞大,加大了故障信号的检索难度,为提高电缆故障信号的检索能力,提出了基于机器学习的电缆故障诊断知识库设计方法。根据电缆故障点的发生位置示意图,描述了电缆故障的产生,利用电缆故障行波的传播过程,确定了电缆故障的发生位置,通过计算电缆故障信号的噪声能量阈值,将电缆故障信号的突变点信息去除,通过采集电缆故障信号的小波熵,反映出故障信号的噪声变化关系,确定了电缆故障信号的噪声能量阈值,采用机器学习建立了电缆故障诊断知识库的模糊决策矩阵,通过设计电缆故障诊断的知识库结构,对电缆故障诊断知识库进行了设计。结果表明,基于机器学习的电缆故障诊断知识库对电缆故障信号的检索精度高达93.4%,在检索效率方面具有更高的性能。

In view of the huge amount of fault signals collected in the process of cable fault diagnosis,it is more difficult to retrieve fault signals.In order to improve the retrieval ability of cable fault signals,a design method of cable fault diagnosis knowledge base based on machine learning is proposed.According to the schematic diagram of cable fault location,the occurrence of cable fault is described.The location of cable fault is determined by using the propagation process of traveling wave of cable fault.By calculating the noise energy threshold of cable fault signal,the abrupt point information of cable fault signal is removed.By collecting the wavelet entropy of cable fault signal,the noise change relationship of fault signal is reflected,and the noise energy threshold of cable fault signal is determined.The fuzzy decision matrix of cable fault diagnosis knowledge base is established by machine learning Experimental results show that the cable fault diagnosis knowledge base based on machine learning has a retrieval accuracy of 93.4%,and has higher performance in retrieval efficiency.