Robust Adaptive And Fault-tolerant Control Of High Speed Trains

High speed trains (HST) play an important role in society, economic development and even national security. The crucial precondition for such development is the safe and reliable operation of HST. Advanced control is one of the key enabling technologies for enhancing system safety. The dissertation is focused on how to enhance system operation safety and reliability by making use of active control methodology to construct multiple safe layers to prevent possible fatal accident.Based on the principle of dynamics, time-domain dynamic models of HST are established comprehensively in the dissertation. With considering non-symmetric input nonlinearities and input saturation in the dissertation, the single model of HST and multiple model of HST are enriched and improved, and these models are more practical. An modified model of HST-multiple point-mass with single-coordinate dynamic model that reflects resistive and transient impacts between the adjacent vehicles is derived in the dissertation. Multiple point-mass with single-coordinate dynamic model keeps the multi-mass feature of HST and has low dimension and simple structure. The model establishment provides the base for analysis and design of effective control strategy.There are some inherent challenges in control design of HST, for instance, randomly time-varying operation condition, uncertain transient impacts, unknown system parameters, unavailable resistance coefficients, and some possible traction/braking failures. In order to tackle these problems, based on Lyapnov theory and analysis of the corresponding model, the dissertation presents robust adaptive and fault-tolerant traction/braking control methods, which are independent of system parameters, resistance coefficients, and the precise information of failures and transient impacts. Safe and reliable operation is ensured under whether the normal condition or the possible traction/braking failure. By using the system core information, the failures and other uncertain factors are compensated effectively. Robustness and adaptivity are achieved under the developed control method, and precise fault detection and diagnosis (FDD) process is unnecessary simultaneously, and global reliable active fault tolerant control of the train is ensured. The dissertation constructs a "pre-fault active prevention"(PFAP) safe layer, which is an innovative fundamental change from "passive treatment after the failure" in safe mechanism.