Digital Twin-Based Condition Monitoring and Health Management for Electromechanical Systems

Digital twin (DT) technology holds promise in the prognostics and health management of electromechanical systems due to its ability to reflect physical entity states, outperforming traditional condition - monitoring methods in data quantity and quality. However, ignoring actual sensor data in DT systems and using threshold - based monitoring may lead to ineffective state monitoring and "missed alarms". This paper proposes a condition - diagnosis method for equipment state monitoring based on causal graphs from DT system deviations. By defining state and health state vectors, it characterizes the DT - physical entity similarity at micro and macro levels and overall equipment health. A data - fusion approach integrates DT and sensor outputs for reliable data. Validation using an electro - hydrostatic actuator shows that this method surpasses traditional ones in monitoring state deviations and data accuracy, offering novel insights for DT application in prognostics.