Condition Monitoring of High-Speed Rotating Machinery using Digital Twins & Machine Learning Techniques

In the era of Industry 4.0, maintenance of factory equipment is changing with new maintenance system using predictive or prescriptive methods. These methods leverage condition monitoring through digital twins, AI, & machine learning techniques to obtain early signs of faults, type of faults, locations of fault etc. Bearings and Gears are one of the most common components and cracking/misalignment/rubbing/bowing are the most common failure modes in high-speed rotating machinery. In the present work, an end-to-end automated Machine Learning based condition monitoring algorithm is developed for predicting and classifying internal gear and bearing faults using external vibration sensors. A digital twin model of the entire rotating system consisting of the gears, bearings, shafts, and housing was developed as a co-simulation between MSC ADAMS and MATLAB. The gears and bearing dynamic models were developed in MATLAB and shaft and housing models were developed in MSC ADAMS flexible body. The co-simulation was achieved through S-function exported from ADAMS/Controls wherein the forces from the MATLAB model were inputs to the ADAMS model and displacements from the ADAMS model were output to the MATLAB model. Autoregression and Spectral Kurtosis based signal processing methods were implemented to perform signal denoising operation on the raw vibration signal. Empirical Mode Decomposition (EMD) on vibration signals was performed for feature extraction studies. Subsequently an Artificial Neural Network (ANN) based Machine Learning model was trained for gear and bearing fault classification. Finally, the effectiveness of proposed algorithm was verified by implementation on one case of rotating machinery at various operating conditions.