Remaining Useful Life Prediction Based on Self-Calibration Convolution with Attention Module

Abstract

The prognostics and health management (PHM) are current research hotspots in various application areas, and for equipment remaining useful life (RUL) prediction technology is a critical component of PHM. RUL describes the interval between the current time and equipment failure time. Accurate prediction of safe operation time of equipment is of great significance for the operation and maintenance of equipment. Therefore, this paper proposes a new data-driven prediction model based on the combination of self-calibration network, convolutional block attention module, and long and short time memory network to accomplish the prediction of RUL of aero-engine. Simulation experiments on the C-MAPSS dataset also test the model performance, and ablation experiments demonstrate the effectiveness of each module.