Integration of Numerical and Experimental Methods to Improve the Safety of Working Machines Through Machine Structure Fault Detection and Diagnosis

This paper presents an integrated numerical and experimental methodology for process-based condition monitoring, early-stage fault detection, and diagnosis to improve the operational safety and structural integrity of remotely operated working machines. Because operators have limited perception of hazardous conditions (e.g., resonance, high vibration, and transient dynamic loads), emerging faults may remain unnoticed. The framework identifies and tracks key diagnostic parameters—especially dynamic load indicators—enabling early detection of abnormal events that can initiate damage in the load-carrying structure and other critical components. A key challenge in designing and deploying such machines is limited knowledge of the occurrence, characteristics, and frequency of dynamic loads in real operations. Underestimating these loads during design can cause unexpected failures and reduced fatigue life. The approach integrates numerical strength simulations with sensor data collected during operation, correlating process signals with complex loading scenarios and hazard states. By combining model-based assessment with experimental validation, the method supports systematic process supervision and fault diagnosis under variable operating conditions. The methodology is demonstrated on an ARE 3.0 remotely operated machine case study and shows how data-informed loading characterization and early anomaly detection can enhance safety and support fatigue-oriented durability assessment.