Method Based on Stacking Denoising Auto-encoder for Rolle Bearing Remaining Useful Life Predtction
The traditional rolling bearings remaining useful life prediction modeling method requires experts with rich experience to select availableindicators or models to extract effective degradation feature curves,then use appropriate prediction models for remaining useful life prediction.To solve the problem of complicated experts dependence on the experience of rolling bearing remaining useful life prediction.This paper proposes a rolling bearing remaining useful life prediction method based on SDAE(stacking denoising auto-encoder)deep learning model.This method first transforms the original data through Fourier,and then calculates multiple times and indicators.Secondly,it is directly used as the input of the SDAE,the life forecast is finally performed.the experimental results shows that the preditc-tion accuracy of our proposed method is better than SAE(stacking auto-encoder)neural network,ELM(extreme learn-ing machine),and LSTM(long short and time memory)neural network models as a whole.
roll bearingstacking denoising auto-encoder(SDAE)deep learningremaining useful life prediction
万方数据
维普
