Method for predicting residual life of rolling bearing based on dropout-SAE and Bi-LSTM

Abstract

The invention discloses a rolling bearing RUL prediction method based on dropout-SAE and Bi-LSTM, and belongs to the field of prediction of bearing operation states. The invention aims to solve the problems of relatively long model training time and relatively low prediction accuracy of an existing rolling bearing RUL prediction method. The invention provides an improved SAE, that is, dropout-SAEperforms unsupervised deep feature adaptive extraction on a rolling bearing vibration signal, the network uses a new Tan activation function to replace an original sigmoid activation function, and a dropout method is adopted to realize sparsity of the network; meanwhile, considering that the prediction method of the remaining service life of the rolling bearing generally only considers past information and ignores future information, a bidirectional long-short-term memory network is introduced to serve as a prediction model of the rolling bearing RUL. Experimental results on two bearing data sets show that the provided prediction method not only can improve the convergence rate of the model, but also has relatively high accuracy.

Description

technical field [0001] The invention relates to a method for predicting the remaining life of a rolling bearing, which belongs to the field of prediction of the running state of the bearing. Background technique [0002] Rolling bearings are the most commonly used and easily damaged key components of rotating equipment, and their operating status often directly affects the performance of the entire equipment [1] . Therefore, it is of great practical significance to predict the remaining useful life (RUL) of rolling bearings. [0003] Feature extraction is an important prerequisite for RUL prediction of rolling bearings. In recent years, deep learning has gained widespread attention for its powerful adaptive feature extraction capabilities and nonlinear function representation capabilities, and has provided a new solution for the feature extraction of rolling bearing vibration signals [2] . Literature [3] proposes an improved deep belief network, which directly takes the ...

AI Technical Summary

Problems solved by technology

[0007] In order to solve the problems of long model training time and low prediction accuracy in the existing rolling bearing RUL prediction method, the present invention further proposes a rolling bearing RUL prediction method based on dropout-SAE and Bi-LSTM

Images