摘要
基于电力公司一体化平台,利用数据挖掘技术将变压器设备管理信息、运行信息以及缺陷信息关联,建立了不同电压等级、运行年限以及不同部件的平均缺陷率参照系,提示高缺陷的设备及厂家信息。提出故障集中度指标,降低特定设备缺陷对结果的影响。提出可信度指标,警示少样本可能引起的误判。在此基础上,开发了变压器家族性缺陷预警平台,对运行在高风险年限的变压器进行预警,同时为变压器采购部门提供相关企业的质量信息。运行表明,缺陷预警平台相比传统方法具有缺陷提示迅速,人为干扰因素少等优点,具有工程适用性。
Design,materials and other factors will lead to familial defects,and early warning on familial defects could improve the reliability of the system operation. Based on the integration platform of electric power company,we used data mining technology to carry out the association rules of management information,running information and defect information on transformer equipments. And an average defect rate reference system based on different voltage levels,operating time and equipment＇s components was established. The system could point out the defective equipment and manufactory＇s information. The paper proposed fault concentration index to cut down the impact of the specific equipment defects on the results,and credibility index to warn the erroneous judgement caused by less sample. On this basis,a familial defect warning systems for transformers is developed,and it warns the transformer running on high-risk transformers years,and provides the quality information for the transformer purchasing department. Compared with traditional method,this system has the advantages of rapid reactions on defects and low human interference factors after a period of operation. The system can be applicable to engineering.
出处
《实验室研究与探索》
CAS
北大核心
2016年第6期37-41,共5页
Research and Exploration In Laboratory
关键词
变压器
家族性缺陷
可信度
预警系统
transformers
family defect
credibility
warning system
