基于分层DSmT 的多故障诊断方法

针对现有故障诊断方法难以诊断涵盖多种不同类型故障的问题,提出一种基于分层DSmT的多故障诊断方法.利用主元凝聚层聚类方法实现证据聚类,将辨识框架分成若干个子框架;利用证据主元将BP神经网络所生成的各种故障模式的基本概率赋值函数在不同辨识框架下重新分配;利用DSmT对子框架下的证据进行融合并得出诊断结果.仿真实验结果表明,所提出的方法能将不同类型故障从辨识框架中分离出来,提高多故障诊断结果的可靠性,减少计算量,提高诊断效率.

Method of multiple faults diagnosis based on hierarchical DSmT

Aiming at that existing fault diagnosis methods are difficult to diagnose faults that cover many different types, a method based on hierarchical DSmT is proposed. Fault elements in the identification framework are separated into different levels by hierarchical clustering based on the current evidence. Then the basic probability assignment function is reassigned, which is constructed based on the output of the BP neural network according to the principal components of evidences in different levels. Finally, DSmT is used to combine the different independent evidences in the same level and draw a conclusion. The diagnostic tests show that the proposed method not only recognizes different faults from the framework, but also reduces the calculation amount, which can effectively improve the accuracy and efficiency of multi-faults diagnosis.