摘要： 通过分析神经网络理论和模糊度量方法的特点，提出基于神经网络预测和模糊度量方法的故障再融合诊断方法。该方法采用多个观测数据及主客观相结合的诊断方法，实现方法间的互补融合诊断，更全面、客观地辨识故障，提高故障定位能力，并给出一个液压设备故障诊断实例验证了该方法的有效性。

关键词: 神经网络预测, 模糊度量, D-S证据理论, 再融合诊断, 隶属度函数

Abstract: Traits of neural network theory and fuzzy measurement method are discussed. A method of the further fusion diagnosis using neural network and fuzzy measurement is proposed. The method uses observational data, integrates diagnosis methods with subjective and impersonal information, which can realize the complementary fusion diagnosis, identify fault more comprehensive and objective, and efficiently improve the fault location capability. A hydraulic equipment fault diagnosis example proves the effectiveness of this method.

Key words: neural network prediction, fuzzy measurement, D-S evidence theory, further fusion diagnosis, membership function

中图分类号: 

• TP139