A robust fault estimation scheme for heavy-haul trains equipped with ECP brake systems

In this paper, the fault detection and isolation problems (FDI) of heavy-haul trains equipped with electronically controlled pneumatic (ECP) brake system are studied. A longitudinal dynamical model and the fault modes of trains are considered. A fault detection estimator bases on the nonlinear observer is designed to generate residual for detecting fault. After the fault is detected, a group of fault isolation estimators with different residuals and thresholds can be designed, each estimator corresponds to one possible actuator fault of trains. And in each fault isolation estimator, the unknown actuator fault of trains can be estimate simultaneously by introduce a learning algorithm. A simulation analysis is proposed to show the validity of the designed FDI scheme, by using the parameters from the heavy-haul trains system running on Datong-Qinhuangdao railway in China.