基于贝叶斯网络的电力变压器局部放电故障检测  

Partial Discharge Fault Detection of Power Transformer Based on Bayesian Network

作　　者：白国政[1] BAI Guozheng(Shaanxi Polytechnic Institute,Xianyang 712000,China)

机构地区：[1]陕西工业职业技术学院,陕西咸阳712000

出　　处：《计算机测量与控制》2023年第9期90-94,共5页Computer Measurement &Control

摘　　要：针对传统电力变压器故障检测方法对电力系统中潜藏的故障问题检测水平不足,准确率较低,无法及时准确的发现异常隐患等问题,提出基于贝叶斯网络的变压器局部放电故障检测方法;首先通过传感器获取电力变压器不同状态下运行过程中的参数数据,对局部放电故障发生的概率和范围进行合理性评估,提取评估概率数据综合为样本数据集,构建贝叶斯网络故障树;根据逻辑规则转化为贝叶斯网络,推演计算故障节点之间的算例关系,利用贝叶斯原理抽取故障特征指标与异常概率之间的关联关系,利用模糊描述方法构建故障特征关联函数,计算可得故障特征模糊函数动态变化关系,实现对变压器故障发生的概率与位置信息的判断与确定;实验结果表明,利用贝叶斯网络对电力变压器局部放电故障检测准确率达到85%以上,最高可达96%,说明该方法具有较高的检测准确率,能够有效提高电力变压器放电故障检测的有效性。Aiming at the problems of insufficient detection level of hidden faults in power system,low accuracy,and inability to timely and accurately detect abnormal hidden dangers by traditional power transformer fault detection methods,a partial discharge fault detection method for transformers based on Bayesian network is proposed.Firstly,the parameter data during operation in different states of power transformers are obtained through the sensors,the probability and range of the partial discharge faults are evaluated rationally,and the evaluation probability data is extracted and synthesized into a sample dataset to construct a Bayesian network fault tree.According to the logical rules,it is transformed into Bayesian network,the calculation relationship between fault nodes is deduced,the correlation between fault characteristic indicators and abnormal probability is extracted by Bayesian principle,the fuzzy description method is used to construct the fault feature correlation function,cauculate the dynamic change relationship of fuzzy functions for the obtained fault features,and realize the judgment and determination of the probability and location information of transformer faults.The experimental results show that the partial discharge fault detection accuracy of power transformers by the Bayesian network reaches over 85%and up to 96%,indicating that this method has high detection accuracy and can effectively improve the effectiveness of power transformer discharge fault detection.

关 键 词：贝叶斯网络 电力变压器 故障树 模糊描述 故障关联知识 

分 类 号：TN624[电子电信—电路与系统]