作　　者： ; ;

机构地区： 国网四川省电力公司

出　　处： 《四川电力技术》 2013年第6期75-77,共3页

摘　　要： 缺陷预测是设备管理中的重要内容。变电设备由于运行环境复杂,设备缺陷发生的随机性较大。基于支持向量机理论,采用数据挖掘技术,通过对设备运行环境和缺陷发生率的统计分析,利用支持向量机建立设备缺陷平均发生率与设备运行环境的回归函数,回归结果与实际情况较为吻合。对给定运行环境下设备缺陷平均发生率进行预测,预测误差小于10%,对设备的运行维护管理具有较高的参考价值。 Defect prediction is the important content of the equipment management. Due to the complex operating condition of transmission and distribution equipment, the randomness of the occurrence of defects is high. Based on the theory of support vector machine, the operating condition and defect occurrence rate of the equipment are analysis using data mining technology, and the regression functions for the average defect occurrence rate and the operating condition are established using support vector machine, whose regression results coincide with the actual situation. The prediction for the average occurrence rate of e- quipment defects under the given operating condition is carried out and the prediction error is less than 10%, which is of valu- able reference for the management of the operation and maintenance of the equipment.

关 键 词： 支持向量机 设备缺陷管理 发生率 预测

领　　域： [电气工程]