RUL Prediction Method for Electrical Connectors With Intermittent Faults Based on an Attention-LSTM Model

The electrical connector is an essential component of all kinds of electronic equipment. It is important to conduct the remaining useful life (RUL) prediction of electrical connectors. The intermittent fault phenomenon has been observed during the degradation process of electrical connectors. This article proposes combining contact resistance features and intermittent faults to predict the RUL of electrical connectors. The evolution characteristics of the contact resistance and intermittent faults in the whole-life degradation process of electrical connectors are compared and analyzed. It is found that the precursor of intermittent fault feature comes out earlier than the rise of contact resistance, which is used to determine the first predicting time (FPT) for RUL prediction. The RUL prediction of electrical connectors is carried out by combining the time-domain evolution features of the contact resistance and intermittent faults, based on a long short-term memory (LSTM) network with an attention mechanism (attention-LSTM). Experimental results demonstrate that the proposed method has good accuracy for the RUL prediction of electrical connectors.