Integrated supervision approach applied to an electric vehicle

In this paper, robust Fault Detection and Isolation (FDI) of traction system in electric vehicle with structured and unstructured uncertainties is presented. Due to the structural and multi-physical properties of the bond graph, the generation of nonlinear model and residuals with adaptive thresholds is synthesized. Then, parameter and structured uncertainties are identified, and a super twisting observer is used for both estimation of unstructured uncertainties and unknown inputs. Co-simulation results with real experimental data shows the robustness of the residuals to uncertainties and their sensitivity to faults.