Intelligent Fault Diagnosis Mechanism for Industrial Robot Actuators using Digital Twin Technology

Intelligent fault detection is a mechanism’s competency to distinguish between healthy and faulty machine signals for smart and efficient diagnosis. The modelling and analysis of the parameters that contribute to the system’s fundamental operation form the crux of the framework. A heuristic technology to enable real-time intelligent fault detection is digital twin technology. Digital twin technology allows a tandem establishment between real-world machines and the virtual domain, allowing for the inclusion of optimization and maintenance frameworks. Sparsely represented machines in the digital twin domain are linear actuators, which form essential parts of various industrial and commercial machines. Therefore, this study has modelled a data-driven and multi-physics robotic linear actuator digital twin, and integrated it with a custom designed fault detection mechanism using Naïve Bayes classifier. This architecture can autonomously be deployed in tandem to the physical machine to alarm and diagnose electrical faults as soon as they occur in the machine. As compared with conventional diagnostics this will reduce machine down-time and expedite repairs. The resultant model built on MATLAB, Simulink gave an accuracy of 96% and required minimal processing capability to operate. Widespread commercial utilization of the proposed model can pave the path for Industry 4.0 utilization of linear actuators as well as technologies including industrial robots that utilize them.