Abstract:

Through the degradation parameters of the sealed electromagnetic relay,the prediction of the product storage life is used to improve the reliability of each stageof the product.Based on the grey model and the BP neural network model,the combined prediction model is constructed by the weighted combined single prediction method,and the combined weight value is optimized by using the error optimal least square method,the reciprocal of mean square error method and the equal weight distribution method.The storage life of the sealed electromagnetic relay is predicted to be 13.17 years by the combined prediction model with the best accuracy.The results show that the combined prediction model with the optimized weight value can effectively improve the prediction accuracy,which provides a new idea for predicting and improving the reliability of such products in important fields.

Key words: storage life, combined forecasting model, weighted optimization, electromagnetic relay