| 引用本文: | 熊卫红,张宏志,谢志成,韩雄辉,李正天,林湘宁.基于云理论及熵权法的变压器潜在故障风险评估方法[J].电力自动化设备,2018,(8): |
| XIONG Weihong,ZHANG Hongzhi,XIE Zhicheng,HAN Xionghui,LI Zhengtian,LIN Xiangning.Risk assessment of power transformer potential fault based oncloud theory and entropy weight method[J].Electric Power Automation Equipment,2018,(8): |
|
| 摘要: |
| 为了充分利用海量样本中所蕴含的信息对变压器的潜在故障风险进行有效识别，采用云理论方法对不同故障类型下变压器油色谱数据与故障类型进行映射，建立了不同故障类型下不同气体的云分布模型，依此构造故障云判断知识库。同时，引入熵权法对油中气体指标的权重系数进行确定，结合云分布隶属度系数，提出变压器潜在故障风险的计算方法。通过对不同训练样本数目下准确判断率的比较，证明了该方法具备对数据的学习能力。与改进的三比值法及已有云理论方法进行了对比，结果证明了该方法的有效性及优越性。 |
| 关键词: 变压器 数据挖掘 潜在故障 风险评估 云理论 熵权法 |
| DOI：10.16081/j.issn.1006-6047.2018.08.018 |
| 分类号:TM41 |
| 基金项目:国家自然科学基金面上项目(51477090)；武汉市“黄鹤英才(科技)计划”资助项目；南网广州供电局科技项目(GZHKJXM20160038) |
|
| Risk assessment of power transformer potential fault based oncloud theory and entropy weight method |
|
XIONG Weihong1, ZHANG Hongzhi2, XIE Zhicheng2, HAN Xionghui3, LI Zhengtian2, LIN Xiangning2
|
|
1.Central China Electric Power Company, Wuhan 430077, China;2.State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology, Wuhan 430074, China;3.Meizhou Power Supply Bureau of Guangdong Electric Power Company, Meizhou 514021, China
|
| Abstract: |
| In order to identify the potential fault risk of power transformer effectively by making full use of the information contained in the large amount of samples, the oil chromatographic data of power transformer and fault types are mapped by adopting the cloud theory method, and the cloud distribution models of different gases under different fault types are established, based on which the fault cloud knowledge base is set up. Meanwhile, the weight coefficients of the indicators for gases in the oil are determined by introducing the entropy weight method and the calculation method for potential fault risk of power transformer is proposed combining with the membership degree of cloud distribution. The distinguishing-positive rates under different amounts of training sample are compared, which verifies the ability of the proposed method to learn data. The validity and superiority of the proposed method are proved by comparing with the improved three-ratio method and the existing cloud theory method. |
| Key words: power transformers data mining potential faults risk assessment cloud theory entropy weight method |
