故障诊断专家系统知识获取的变精度粗集方法

VPRS Approach to Knowledge Acquisition for Fault Diagnosis Expert System

针对故障诊断专家系统知识获取的问题 ,利用变精度粗集 (VPRS)理论模型进行知识简化 ,并采用正则条件熵和互信息熵对故障特征的选择进行评价 ,实现最简诊断知识的提取 ,以建立专家系统知识库。对滚动轴承故障诊断的实验表明 ,该方法有效地弥补了传统故障诊断专家系统知识获取的不足 ,可正确地实现故障诊断功能 ,在实际系统的故障诊断中具有应用价值。

In order to solve the problem of knowledge acquisition for fault diagnosis expert system, a variable precision rough set approach is proposed. Regularly conditional entropy and mutual information entropy are employed in the evaluation of fault features. The simplest diagnostic knowledge is acquired by variable precision rough set theory to construct the ES knowledge base. Simulation experiment of fault diagnosis of rolling bearings showes the effectiveness and the superiority of the proposed method.