基于决策树算法的断路器弹簧操动机构振动诊断技术

Vibration Diagnosis Fechnology of Technology of Spring Operating Mechanical for Circuit Breaker-based on Decision Tree Alogorithm

断路器操动机构是GIS设备的主要部件之一,其机械类故障占GIS设备总故障比例高达17%,如何对其进行可靠有效的状态检测和评估受到了越来越多的关注。文中调研并选择了GIS断路器操动机构中4种典型机械故障,基于CT26弹簧操动机构对4种故障进行了实验室模拟,利用加速度传感器建立了正常及4种典型故障情况下的振动声学指纹数据库。基于5类GIS操动机构运行状态振动声学指纹数据库,建立了基于决策树分类模型的故障诊断方法,并利用测试数据完成了故障诊断方法的验证,验证结果表明该算法行之有效,部分安装位置检测数据的故障识别准确率达到了92%。

Circuit breaker operating mechanism is one of the main equipment of GIS,Its mechanical failure take up 17%of GIS failures.How to detect and evaluate the reliable and effective of the operating mechanism is getting more and more attention.This paper investigated and selected 4 kinds of typical mechanical failures and finished laboratory simulation for these 4 failures based on CT26 spring-type operating mechanism.And it established vibration acoustic fingerprint library by capacitance acceleration sensor,which contains normal situation and failures.Based on five types of GIS operating mechanisms operating state fingerprint database,a fault diagnosis method based on the decision tree classification model is established,and the verification of the fault diagnosis method is completed with the test data.The verification results show that the algorithm is effective.The fault recognition accuracy rate reached 92%with data from sensors at certain installation location.