Health Index Construction and Remaining Useful Life Prediction of Mechanical Axis Based on Action Cycle Similarity

Abstract

Aiming at the low efficiency of manual detection in the axis health management of the industrial robot, a health index (HI) construction method based on action cycle similarity measurement is proposed. Furthermore, the remaining useful life prediction is carried out by Long Short-Term Memory (LSTM) network. Firstly, MPdist is used to calculate the comparison distance, and then the health index is constructed. Secondly, the long short-term memory network model is trained by the health index set. Finally, the MPdist-LSTM model is used to calculate the remaining useful life (RUL) automatically. The experimental results show that the monotonicity and trend of HI constructed by the MPdist algorithm improve by 0.07 and 0.13 compared with Dynamic Time Warping (DTW), Euclidean Distance (ED), and Time Domain Eigenvalue (TDE). The R-square of RUL prediction based on the MPdist-LSTM model reaches 0.96, which is higher than MPdist-RNN, DTW-LSTM, and TDE-LSTM model.