一种证据理论方法在数控机床故障诊断中的应用

CNC Fault Diagnosis Based on Evidence Theory Method

针对证据理论组合规则在数控机床故障诊断中所出现的反直观问题,提出了一种基于证据相容度的证据组合方法。根据证据相容性函数获得证据间相容度矩阵,归一化得到一致相容性度量,在此基础上得到相容系数的偏离度来表示证据的稳定性,联合证据的一致相容性和稳定性获得证据的权系数,采用D-S组合规则完成对证据的加权平均组合。该方法应用在某型国产数控机床的故障诊断中,实验结果和相关分析表明是合理的。

For the counter-intuitive results problem when using evidence theory in the fault diagnosis ofCNC, an evidence combination method was presented based on evidence of consistency degrees. The ev-idence consistency degrees matrix was obtained from evidence compatibility function, was normalizedto give evidence consistent compatibility measure, and the deviation degree of evidence compatibilitycoefficient was defined in order to characterize the stability of evidence, then the weights of evidencewas obtained combining with consistent compatibility and stability of evidence, finally, the weighted av-erage combination of evidence was completed using the D-S combination rule. This method was appliedto the fault diagnosis of a certain type of domestic CNC, the experimental results and analysis show themethod is reasonable and effective.