Research On Generation Method For Heavy Haul Train Driving Curve Based On Neural Network

Heavy haul train has become an important transportation in the national economy by its characteristics such as high efficiency, low cost and high transport capacity. Because of the factors such as big train mass, long train marshalling, the line slope, speed limit and the nonlinear propagation and attenuation of air braking wave along the train when the heavy haul train air braking reducing the train pipe pressure, and also the nonlinear charging of train pipe when releasing, all these factors increased the difficulty for drivers to control the train when the heavy haul train running. How to take the above factors into consideration and generate the safe heavy haul train driving curve, it is an important foundation to ensure the safe running of heavy haul train. Based on the big train mass, long train marshalling, continuous long heavy down grade and nonlinear of train air braking, this thesis researched the driving curve generation method for heavy haul train and provided a technical support for safety driving of heavy haul train drivers.Based on the basic algorithm of BP neural network, the nonlinear model of the heavy haul train was established in this thesis, which was optimized by genetic algorithm. The capacity of algorithm, such as fast, random and global search ability was also improved by genetic algorithm, and the operation model of heavy haul train was built at the same time. Generation method for heavy haul train driving curve based on BP neural network which has been optimized by genetic algorithm was put forward, the simulation was verified by the actual data. The main research contents were as follows:(1) Conditions such as speed limit, continuous down grade and train marshalling were considered and the line was divided in a reasonable manner. And the noise reduction and normalized processing of data were also needed.(2) The neural network and its optimization algorithm were respectively used modeling the braking operation process of heavy haul train and then the braking parameters were obtained.(3) Based on the braking parameters, characteristics and operation requirements of heavy haul train, the heavy haul train operation parameters were calculated by the constraint conditions.(4) The actual data of ShuoHuang line heavy haul train operation direction was taken as the basic, and the scenario of train operation on continuous long heavy down grade was analyzed and researched. In the Simulation, BP neural network and its optimization algorithm were respectively used by adopting MATLAB software, then the heavy haul train driving curve was obtained and the results of simulations were compared with actual driving curves. By comparing the data between actual driving curves and simulation driving curves, the heavy haul train driving curves which was generated by the algorithm in this thesis can ensure the operation safety of train and all the results has turned out to be very effective.