An energy-based nonlinear pressure observer for electronically controlled pneumatic brake fault diagnosis

As widely used in the heavy-haul combined trains, electronically controlled pneumatic(ECP) brake is a highly nonlinear system. Heavy-haul combined train consists of variant freight trucks with different types and different load weights. When operating with brake mode, inconsistent braking forces will aggravate impacts of collisions between adjacent freight cars, especially uncertainty faults occur and the trucks in rigid connection mode. Therefore, it is necessary to conduct online fault diagnosis for ECP brakes. In this paper, an Lyapunov-based globally stable pressure observer for ECP brake is proposed. Firstly, according to the output value of the observer, a residual generator is designed and residual signals are evaluated with the Generalized Likelihood Ratio(GLR) method. Secondly, fault diagnosis can be achieved by GLR when system's faults happened. Finally, experimental results demonstrate and validate effectiveness of the proposed observer and algorithm.