Cross-Domain Remaining Useful Life Prediction Based on Adversarial Training

Remaining useful life prediction can assess the time to failure of degradation systems. Currently, numerous neural network-based methods have been proposed by researchers. However, most work contains an implicit prerequisite: network training and testing data same operating conditions. To solve this problem, adversarial discriminative domain adaption method based on is improve accuracy cross-domain under different working First, LSTM feature extraction constructed mine source target for deep representation. Subsequently, parameters are adjusted idea achieve invariant mining. The scheme experimented a publicly available dataset achieves state-of-the-art performance compared recent unsupervised adaptation methods.