CAJ | 학술논문

针对高压断路器故障样本难以获取等问题,在研究了其机械特性的基础上,本文提出了一种基于实数阴性选择法(RNS)和支持向量机(SVM)的两级分类器诊断方法. 建立分类器的数学模型. 通过实数阴性选择算法产生检测器,进行第一次分类,对不属于检测器的数据导入二次分类器,进行二级分类.实验表明,基于实数阴性算法和支持向量机算法相结合的断路器故障诊断算法,对断路器的多种故障都能够有效地分类,提高了断路器故障诊断的准确率和速度.
침대고압단로기고장양본난이획취등문제,재연구료기궤계특성적기출상,본문제출료일충기우실수음성선택법(RNS)화지지향량궤(SVM)적량급분류기진단방법. 건립분류기적수학모형. 통과실수음성선택산법산생검측기,진행제일차분류,대불속우검측기적수거도입이차분류기,진행이급분류.실험표명,기우실수음성산법화지지향량궤산법상결합적단로기고장진단산법,대단로기적다충고장도능구유효지분류,제고료단로기고장진단적준학솔화속도.
For high-voltage circuit breaker fault samples are difficult to obtain, based on the mechanical characteristics of circuit breakers, this paper presents a two-stage diagnostic classification method based on real negative selection method (RNS) and support vector machine (SVM). Establishing classifier mathematical model. Through real negative selection algorithm to generate detector for the first time classification, which are not part of the detector data into two classifiers, for secondary classification. Experimental results show that the circuit breaker fault diagnosis algorithm based on negative real number algorithms and support vector machine algorithm are able to effectively classify ,improving the accuracy and speed on breaker fault diagnosis.