Process Mining Driven Modelling and Simulation for Enhanced Fault Diagnosis in Cyber-Physical Systems

<p dir="ltr">Cyber-Physical Systems (CPS) integrate computational algorithms with physical processes, creating complex interdependencies that challenge traditional fault diagnosis approaches. This paper presents a novel framework leveraging process mining techniques to automatically discover, model, and simulate CPS behavior for enhanced fault diagnosis. We introduce the Process Mining-Enhanced Fault Diagnosis (PM-EFD) algorithm, which combines-algorithm extensions, temporal conformance checking, and stochastic Petri net simulation. Mathematical proofs establish the convergence properties and fault detection guarantees of our approach. Experimental validation on three industrial CPS datasets demonstrates 94.3% fault detection accuracy with 87.6% reduction in false positives compared to conventional methods. The framework enables real-time anomaly detection, root cause analysis, and predictive maintenance through continuous event log analysis and model conformance checking. All experimental code and datasets are made available for reproducibility.</p>