基于“小波包—模糊理论”的整流装置故障诊断系统的研究

Study of fault diagnosis system in rectifier units based on wavelet packet and fuzzy diagnosis

针对电力、化工行业广泛使用的三相桥并联的同相逆并联整流装置,以整流变压器阀侧三相电流为诊断信号,提出了小波包-模糊推理实现故障诊断。首先利用小波包分析提取故障特征值,形成样本数据库;然后利用J散度、最大隶属原则方法实现待识别状态与样本状态的模糊识别。该方法通过实验,验证了其故障诊断的准确性。

According to the cophase counter parallel connection rectifier units widespreadly used in the fields of power and chemical industry, the paper presents a wavelet packet-fuzzy recognition method with the three phase current at rectifier transformer side as the diagnosis signal to realize fault diagnosis. Firstly, the fault eigenvalues are extracted with wavelet packet analysis to form a sample database, then the fuzzy recognition of sample status are identified by using J dispersion and maximum membership principle. The method is verified by experiments, and the accuracy is excellent.