Neural Network Adaptive PID Control Based On High-Gain Observer And Its Application In Semi-Active Control Of Pantograph-Catenary System

The pantograph-catenary system of high speed trains is a nonlinear coupled system with complex structure and constraints.When the train runs at high speed,the pantograph-catenary system will be influenced by many factors.The simple linear model cannot provide the valid information about the system.Furthermore,the design of the controller cannot be too heavily dependent on the model.In fact,it is not very useful that a controller relies on the model or the internal states of the system too much,and there is only a theoretical significance.Therefore,a dynamic mathematical model close to the real scene and a controller independent on the internal state of the system are needed to establish,so as to the current collection quality of pantograph-catenary system can be further improved.In this paper,the nonlinear coupling model of two freedoms pantograph-catenary system is established by analyzing different types of pantograph and catenary.Considering the unavailable internal state of the system,the neural network adaptive PID control based on high-gain observer is designed.The main work of this article is as follows:(1)The research status of pantograph control system in the three directions of passive control,active control and semi-active control at home and abroad is described.And the necessity of semi-active control in pantograph catenary system is pointed out.(2)The theory of neural network,especially the characteristics of nonlinear approximation of neural network,is introduced,and then the theory of uniform ultimate boundedness,the form of Brunowsky norm and the theory of high gain observer are expounded separately,which lay the foundation for the theoretical derivation.(3)The working principle of pantograph-catenary system,the evaluation system of current collection quality,the models of catenary and pantograph and their coupling models are expounded,and then the two-element nonlinear coupling model of pantograph-catenary system is established.The model takes into account the flexibility of the catenary and pantograph,and is more close to the actual situation than the commonly used model.(4)For the general nonlinear system model and the unavailable internal states of the system,the high gain observer is used to estimate the system states using,and the neural network approximates the nonlinear part of system,which reduce the difficulty of application for controller.The adaptive parameter is introduced to design the neural network adaptive PID controller based on high gain observer.The stability is proved by selecting the appropriate Lyapunov function.(5)By introducing the error state variables,two elements nonlinear model of the pantograph catenary system is transformed into the Brunowsky norm form,and the neural network adaptive PID controller based on the high gain observer is constructed.The semi-active control algorithm is further introduced to complete the semi-active control of pantograph catenary system.Finally,the feasibility and superiority of the controller designed in this paper are verified by simulation.