Research on Storage Life Rolling Prediction of Aerospace Relay Based on Data Fusion

Abstract

The traditional prediction of aerospace relay life is based on the failure data, ignoring the test data during the degradation process. In this paper, four failure-free degradation parameters under the four kinds of temperature stress in the accelerated life test were collected and the comprehensive correlation analysis method was carried out to integrate the parameters. The storage life of the relay under each temperature gradient was obtained by using the dynamic neural network of time series to predict the fused pressure drop parameters. This method fully considers the close relationship between the degradation parameters and the dynamic changes of the parameters, which improves the accuracy of the storage life prediction.