基于重要度的分级贝叶斯网络诊断模型研究

为了提高复杂系统故障检测和诊断的准确性,从数据的不确定性、诊断的不确定性以及特征参数选择的不确定性三方面展开研究。在贝叶斯网络解决不确定性问题优势的基础上,提出了基于重要度的分级贝叶斯网络模型。以往基于贝叶斯网络的方法无法做到系统地选择参数,建立的模型都是全模型。虽然全模型没有遗漏关键的参数,但包含了很多不必要的参数,这些多余参数不仅会对诊断造成不良影响,导致拖尾效应。仿真结果表明,提出方法可以获得更佳的故障特征,有利于提高故障识别能力,验证了模型的有效性。

Research on Hierarchical Bayesian Network Diagnosis Model Based on the Importance Degree

In order to improve complex system detect and diagnosis fault precisely, make a research from three respects: uncertainty of data, uncertainty of diagnosis and uncertainty of feature parameter selection. Based on advantages for uncertainty problem of Bayesian network, Hierarchical Bayesian for fault diagnosis is proposed. The existing algorithms are not capable of selecting variables systematically so that they generally use the full model, which may contain unnecessary variables as well as necessary variables. Ignoring this model uncertainty often gives rise to, so called, the smearing effect in solutions. Complexity and difficulty of modeling is increased. The simulation results show this method can get better fault feature, improve fault discernment, and validate the model efficiency.