Bond Graph modeling for fault detection and isolation of a train door mechatronic system

• Bond Graph modeling of a train door used in railway sector.
• Residuals computation for faults detection and isolation.
• Influence analysis of internal faults and external perturbations.
• Faults detection and isolation of electrical and mechanical train door components.
• Model sensitivity for detection of some external perturbations.

In this paper, monitoring and diagnosis of a new generation of train door mechatronic system is proposed. Bond Graph methodology is applied to obtain a reference model. In addition to this reference model, a global model based-FDI (Fault Diagnostic and Isolation) is developed for the generation of fault indicators and residual thresholds in presence of door failures. The ability of the proposed diagnostic approach to detect train door failures is demonstrated. The main contribution of this paper concerns the implementation of FDI procedure on a train door instead of health monitoring as it is usually performed in maintenance field.