基于改进贝叶斯网络的车载设备故障诊断研究  

Diagnosis of On-Board Equipment Based on Improved Bayesian Network

作　　者：赵君 查园园 ZHAO Jun;ZHA Yuanyuan(College of Electric power,North China University of Water Resources and Electric Power,Zhengzhou Henan 450045,China;School of Eectronic and Information Engineering,Beijing Jiaotong University,Beijing 100044,China)

机构地区：[1]华北水利水电大学电气工程学院,河南郑州450045 [2]北京交通大学电子信息工程学院,北京100044

出　　处：《信息与电脑》2022年第24期87-91,共5页Information & Computer

摘　　要：高速铁路列控车载设备是保证列车安全运行的关键,故障特征表现为复杂性和不确定性。贝叶斯网络在处理不确定性和相关复杂性的问题时具有显著优势,该研究以CTCS3-300T型车载设备为研究对象,建立约简贝叶斯网络模型进行故障诊断。首先,通过分析典型车载设备故障处理现状,提出一种结合专家知识、故障数据集和K2算法的贝叶斯网络模型研究方法。其次,利用K2算法和最大似然估计法分别进行结构学习、参数学习,从局部到整体贝叶斯网络诊断模型,实现快速定位。最后,建立整体贝叶斯网络模型,运用粗糙集理论进行信息约简,建立约简后的贝叶斯网络模型。经实例分析,结果表明该模型能够保证故障诊断结果的准确性,进一步提高故障诊断的效率。The fault characteristics of the on-board equipment in the column control system are complexity and uncertainty, the data recording is non-textual, and the traditional diagnostic methods based on expert knowledge have low efficiency and poor accuracy. Bayesian networks have significant advantages in dealing with the problems of uncertainty and related complexity, and the study used CTCS3-300T onboard equipment to establish a Bayesian network model for fault diagnosis.Firstly, by analyzing the current situation of fault processing of typical vehicle equipment, we propose a Bayesian network model research method combining expert knowledge, fault data set and K2algorithm. Second, the K2 algorithm and the maximum likelihood estimation method are used to perform structure learning and parameter learning, respectively, from local to global Bayesian network diagnosis model, to achieve rapid localization. Finally, the overall Bayesian network model is established, and the rough set theory is used to reduce the information, and to establish a reduced Bayesian network model.After example analysis, the results show that the model can ensure the accuracy of fault diagnosis results and further improve the efficiency of fault diagnosis.

关 键 词：车载设备 专家知识 K2算法 最大似然估计算法 粗糙集 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]