Automatic Driving Control Method Of Heavy Haul Train With Multi Locomotive Traction

As two important strategic pivots to realize railway modernization in China,heavy-haul transportation and high-speed railway are developing vigorously in freight and passenger transportation respectively.Due to the large number of tractive vehicles,the forces on the internal vehicles in the marshalling of heavy-haul trains during their operation are more complicated than those on high-speed trains,and the workshop forces are mainly expressed in the form of coupler forces,the numerical value has a direct influence on whether the train can run safely,so it must be controlled in a certain range to prevent the occurrence of coupler breaking,decoupling,derailment and other accidents.The traditional heavy-haul train driving mostly adopts driver’s manual control mode,which often causes the problems such as: too big impact of coupler and train running late,because of the difference of driver’s labor level.In order to better achieve the strategic goal of “The transportation volume is larger and the transportation speed is faster” in Chinese railway freight transport,to achieve the great leap of reducing the drivers’ labor intensity while increasing the volume of railway freight transport.It is urgent to research and develop Automatic Train Operation(ATO),which can meet the needs of freight transportation in China,and to apply it to busy mainlines in our country.In this paper,the multi-particle longitudinal dynamic models of heavy-haul trains are established.The Proportional Integral Controller theory(PI)and Model Predictive Control(MPC),which have been widely used in industrial production,are used to design the ATO controllers of heavy-haul trains,the train can track the given speed target curve,which is improved by Back Propagation Neural Network(BPNN)+ PI control theory and Explicit MPC(EMPC)respectively,finally use MATLAB / SIMULINK software for simulation test.The main contents are as follows:(1)On the basis of fully considering the forces acting on the train in the marshalling yard,the operating mechanism and characteristics of the train are analyzed,and the state space equations are established.Based on the state space model,the controller is designed by determining the specific marshalling situation of heavy-haul train and selecting the type of train and vehicle as well as the actual road information.(2)The PI controller is designed for heavy-haul train to track the speed target curve.In order to improve the performance of train operation,BP neural network algorithm is introduced to improve it,and a heavy-haul ATO system based on BPNN+PI control algorithm is designed.The simulation results show that the control effect of BPNN+PI controller is better.Compared with the traditional PI controller,the performance of BPNN+PI controller is greatly improved,and it can meet the actual driving requirements.(3)In order to better consider the constraint conditions of the heavy-haul train,the speed target curve can be tracked and several performance indexes can be achieved.Compared with PI and BPNN+PI control algorithms,this algorithm is more suitable for MPC algorithm of heavy-haul train,which is a complex large-scale system with constraints.In order to improve the real-time control effect,the idea of EMPC control algorithm is introduced,and the ATO system based on EMPC is designed.The simulation results show that the train running with EMPC controller is more stable,the maximum coupler force is smaller and the energy consumption is lower.