Research on optimization of belief rule-based on Naive Bayes theory Provisionally Accepted

Hong Qian¹ Yutong Pan^1* Xuehua Wang² Zhenpeng Li²

• ¹Shanghai University of Electric Power, China
• ²China Nuclear Power Engineering Co Ltd, China

The belief rule-based is crucial in expert systems for intelligent diagnosis of equipment. However, in the belief rule-based for fault diagnosis, multiple antecedent attributes in belief rule-based are often initially determined by domain experts. Actually, multiple fault symptoms related to multiple antecedent attributes are different when actual fault happening. This leads to results of multiple antecedent attributes matching with multiple fault symptoms non simultaneously, thereby result of fault diagnosis lack of timeliness and accuracy. To address this issue, this paper proposes a method for belief rule-based optimization based on Naive Bayes theory. Firstly, taking fault sample in a long enough window, dividing into several intervals samples, making the analysis samples approximate the overall samples. Secondly, using Gaussian Mixture Clustering and Naive Bayes optimization, Iterate over the threshold and limit values of fault symptoms in belief rule-based, based on the requirements of the timeliness and accuracy of fault diagnosis results. Finally, the belief rule-based is optimized. By using fault samples from high-pressure heaters and condensers, it is demonstrated that a significant improvement in the timeliness and accuracy of fault diagnosis with the optimal belief rule-based.