一种磨损故障融合诊断新方法及其应用

A New Approach to Fusion Diagnosis of Wear Faults and Its Application

提出一种基于D-S证据理论的磨损故障融合诊断新方法。该方法设定磨损故障为双重假设问题,首先利用油样检测的界限值,通过引入模糊集合思想,设计了"油样检测结果异常"的隶属度函数,并以此获取各证据的可信度;其次,建立"故障"与"油样检测参数异常"之间的规则,利用规则可信度和证据可信度的乘积得到规则综合置信度,并以此作为D-S证据理论中各证据对结果的支持程度;最后,以航空发动机滑油光谱监控和自动磨粒检测的融合诊断为例,以"滚动轴承疲劳磨损"为诊断对象,进行了实际诊断案例分析。

In this paper, a new fusion diagnosis approach to wear faults based on D-S evidence theory is put forward. In the new method, the double supposition of wear faults is supposed. First, the fuzzy set is introduced, the membership functions about ＂oil data abnormality＂ are designed, and the evidences beliefs are computed. Second, the rules about ＂fault＂ and ＂oil data abnormality＂ are established, and the product of rule belief and evidence belief is rule-integrated belief which acts as the belief of evidence in D-S decision. Finally, the fusion diagnosis of the ball bearing fatigue wear of an aero-engine based on the Spectrometric Oil Analysis （SOA） and automatic debris analysis are carried out. Practical cases are analyzed, and the results show the correctness and effectiveness of the new fusion diagnosis method.