Multi-model switching predictive control with active fault tolerance for high-speed train
高速动车组多模型切换主动容错预测控制

In the continuous high-speed operation, severe demands on reliability and disturbance-rejection are needed by the high speed trains. According to its nonlinear dynamic characteristics and operation data, we build a set of multiple models for the high-speed train by using subtractive clustering and pattern classification algorithm. To adapt to the change of object and disturbance characteristics, we use a model switching scheme to select online, from this set of models, the optimal model with smallest model accumulative error. On the basis of this optimal model, we design the active fault tolerant predictive controller to realize the secure and efficient operations of the high-speed train. Simulation example is given to show the effectiveness of this method.