Digital Twin-Driven Bearing Defect Model and Life Cycle Prediction Research

It is recommended to use digital dual drives to predict the complete life of bearing. This method consists of a numerical model with two defects and a complete life cycle prediction of two parts of the bearing. The digital twin defect modeling employs the Kalman Filter Algorithm (KFA) to fit the defect area of the physical objects, enhancing the accuracy of mapping the physical objects into the virtual space. Improved Short - and Long Term Memory Networks (LSTM) are used to predict the entire life cycle of bearings. It predicts the life cycle of physical objects based on data obtained from dual fault zones. The effectiveness of this method was confirmed using an open bearing dataset. The analysis results demonstrate its capability to effectively fit defect area sizes and predict the bearing life cycle.