Tool Remaining Useful Life Prediction based on Edge Data Processing and LSTM Recurrent Neural Network

Abstract

The real-time prediction of tool remaining useful life is a challenging problem. This paper proposes a Long Short-Term Memory (LSTM) recurrent neural network model, which is combined with an edge data processing method to predict the tool remaining useful life in real-time. Data cleaning and feature extraction are carried out at the edge node to reduce the transmission time, save the transmission cost and improve the real-time performance of life prediction. After further feature selection in the cloud, a simple three-layer LSTM recurrent neural network model is established. Compared with the tree model and the ordinary neural network model, the experimental results show that the LSTM model has better performance of the tool remaining useful life prediction.