信息熵在变压器故障诊断与状态检测中的应用
发布时间：2018-10-04 23:03
【摘要】：随着电网规模的不断扩大,电气设备数量不断增加,传统的以定期检修为主的检修模式已经无法满足电网发展的要求。近些年来状态检修得到了越来越多的关注,这种方法通过在线监测数据及时发现设备异常,并对故障部位进行有针对性的维修。但变压器结构复杂,现场数据繁多,在分析数据进行变压器状态评估与故障诊断的理论研究方面仍存在很多不足。 本文以信息熵理论为基础,结合成熟的故障树模型以及油色谱诊断技术,对变压器检修方案优化进行研究。通过计算故障树信息熵值,得出国网检修导则中实验项目的信息熵值；根据每种检测方法的信息熵值,对变压器检修方案进行优化,建立基于信息熵理论的变压器状态检修策略优化模型。在综合考虑测试结果、检测方法难易程度和故障概率等因素的情况下,对变压器检测项目进行排序。提高检修效率的同时,最大限度的减少检修次数降低检修成本。 根据变压器故障历史统计数据,以变压器油中溶解气体参数为特征向量,在传统的灰色关联度理论的基础上结合信息熵得到一种新的变压器故障诊断方法。通过对每种特征气体信息熵值的计算,对传统的特征气体参考序列因子权重值进行修订。这种新的基于事故概率的参考因子,避免的以往灰色理论诊断中的人为因素影响,有效提高了故障诊断的客观性和准确性。
[Abstract]:With the continuous expansion of the scale of power grid and the increasing number of electrical equipment, the traditional maintenance mode based on periodic maintenance has been unable to meet the requirements of the development of power grid. In recent years, more and more attention has been paid to condition-based maintenance (CBM). This method detects equipment anomalies in time through on-line monitoring data, and makes targeted maintenance of fault sites. However, the structure of transformer is complex and the field data is various, so there are still many deficiencies in the theoretical research of transformer condition evaluation and fault diagnosis by analyzing the data. Based on the information entropy theory and the mature fault tree model and oil chromatographic diagnosis technology, this paper studies the optimization of transformer overhaul scheme. By calculating the information entropy value of the fault tree, the information entropy value of the experimental items in the national network maintenance guidelines is obtained, and the transformer maintenance scheme is optimized according to the information entropy value of each detection method. The optimal model of transformer condition maintenance strategy based on information entropy theory is established. In the case of considering the test results, the degree of difficulty and the probability of failure, the items of transformer detection are sorted. At the same time, improve the efficiency of maintenance, minimize the number of maintenance and reduce the cost of maintenance. Based on the historical statistical data of transformer faults and taking dissolved gas parameters in transformer oil as eigenvector, a new method of transformer fault diagnosis is obtained based on the traditional grey correlation degree theory and information entropy. By calculating the entropy value of each characteristic gas information, the weight of the traditional characteristic gas reference series factor is revised. This new reference factor based on accident probability avoids the influence of human factors in previous grey theory diagnosis and effectively improves the objectivity and accuracy of fault diagnosis.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM407
本文编号：2252223
[Abstract]:With the continuous expansion of the scale of power grid and the increasing number of electrical equipment, the traditional maintenance mode based on periodic maintenance has been unable to meet the requirements of the development of power grid. In recent years, more and more attention has been paid to condition-based maintenance (CBM). This method detects equipment anomalies in time through on-line monitoring data, and makes targeted maintenance of fault sites. However, the structure of transformer is complex and the field data is various, so there are still many deficiencies in the theoretical research of transformer condition evaluation and fault diagnosis by analyzing the data. Based on the information entropy theory and the mature fault tree model and oil chromatographic diagnosis technology, this paper studies the optimization of transformer overhaul scheme. By calculating the information entropy value of the fault tree, the information entropy value of the experimental items in the national network maintenance guidelines is obtained, and the transformer maintenance scheme is optimized according to the information entropy value of each detection method. The optimal model of transformer condition maintenance strategy based on information entropy theory is established. In the case of considering the test results, the degree of difficulty and the probability of failure, the items of transformer detection are sorted. At the same time, improve the efficiency of maintenance, minimize the number of maintenance and reduce the cost of maintenance. Based on the historical statistical data of transformer faults and taking dissolved gas parameters in transformer oil as eigenvector, a new method of transformer fault diagnosis is obtained based on the traditional grey correlation degree theory and information entropy. By calculating the entropy value of each characteristic gas information, the weight of the traditional characteristic gas reference series factor is revised. This new reference factor based on accident probability avoids the influence of human factors in previous grey theory diagnosis and effectively improves the objectivity and accuracy of fault diagnosis.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM407
【参考文献】
相关期刊论文 前3条
1 何雷;李丽平;;基于云模型的电力变压器状态综合评判[J];华北电力大学学报(自然科学版);2009年03期
2 朱永利;宫政;武中利;薛磊;;正态云模型在电力变压器状态评估中的应用[J];华北电力大学学报(自然科学版);2010年05期
3 王国胤,于洪,杨大春;基于条件信息熵的决策表约简[J];计算机学报;2002年07期
，本文编号：2252223
本文链接：https://www.wllwen.com/kejilunwen/dianlilw/2252223.html
