Research On Evaluation And Prediction Of Electrical Life For Contactor Individual

AC contactor is a low-voltage switchgear used to frequently switch on and off electrical equipment.The electrical life of the AC contactor individual will directly affect the reliable operation of the electrical equipment,so it is particularly important to study the prediction of the electrical life of the AC contactor individual.The electrical life degradation process of the AC contactor is divided into different periods.Predicting the expected electrical life grade of the AC contactor individual based on the key initial state characteristics of the early life,it is possible to achieve a qualitative evaluation of the expected electrical life of the AC contactor individual during the offline phase.Using the performance degradation characteristics at the end of life to predict the remaining electrical life times of the AC contactor individual,it is possible to achieve a quantitative prediction of the remaining electrical life of the AC contactor individual during the online use phase.Therefore,this paper proposed a comprehensive forecasting scheme for the electrical life of the AC contactor individual,and studied the qualitative evaluation of the expected electrical life and the quantitative prediction of the remaining electrical life for the AC contactor individual.First of all,for the realization of data acquisition and prediction scheme of electrical life prediction of the AC contactor individual,the characteristic parameters and life information of the AC contactor samples were collected by the AC contactor full life test platform to provide data support for subsequent electrical life prediction research.Combining the principle of the degradation stage of the bathtub curve,considering different the characteristic information and different degradation trends in degradation stages of the AC contactor,a comprehensive prediction scheme of electrical life for the AC contactor individual was designed.Secondly,in order to extract the key initial state features of the qualitative evaluation of the expected electrical life in the comprehensive prediction scheme,the correlation between the initial state characteristic parameters and the electrical life was analyzed by Spearman rank correlation coefficients.The segmented mean value of arc energy and the segmented mean value of contact voltage drop were used as two types of initial state characteristics.Then,the best attribute set of single-class initial state characteristics was screened based on rough set theory and related calculations.After combining the best attribute sets,the key initial state characteristics of qualitative evaluation of the expected electrical life were selected by the second reduction.Thirdly,in order to realize the qualitative evaluation of expected electrical life for the AC contactor individual,a method of uncertainty reasoning of the expected electrical life based on evidence theory was proposed.First,the key initial state characteristic weight set was established by G1-grey relation weighting method,and the matching table of key initial state characteristics was established based on the attribute condition density.After using the two to build independent evidence,the evidence theory was used to decide the expected electrical life grade of AC contactor individual for completing the qualitative evaluation of the expected electrical life of the AC contactor individual,and the qualitative evaluation results of the expected electrical life under different sample proportions were compared.Finally,in view of the quantitative prediction of the remaining electrical life for the AC contactor individual,a prediction model of the remaining electrical life based on the performance inflection point+Elman neural network was proposed.First,an inflection point recognition method based on judging contact voltage drop threshold was proposed to identify the performance inflection point of AC contactor.Then,the Spearman rank correlation coefficient was used to screen the cumulative arcing energy after the inflection point as a predictive variable,and the number of remaining electrical life after the inflection point was used as the output.The remaining electrical life prediction model of the performance inflection point+Elman neural network was established to complete quantitative prediction of remaining electrical life for the AC contactor individual,compared with the conventional prediction model of non-inflection point,greatly improves the prediction accuracy of remaining electrical life after the performance inflection point.