Structural Optimization of Simple Catenary Based on BP Neural Network and Genetic Algorithm
The standard deviation of contact force(CFSD)between pantograph and catenary is an important index to evaluate the current collection quality of pantograph and catenary.In order to optimize the pantograph catenary current collection quality,a pantograph-catenary coupling simulation model verified by EN 50318 standard and the simulation results in the existing literature is established.The central composite design method is used to design the test table of input parameters(dropper spacing and contact wire pre-sag).The pantograph catenary coupling simulation model is carried out after modifying the model according to the test table.Based on this,the output parameter(CFSD)is obtained.Back propagation neural network(BPNN)is used to establish the relationship model between input and output parameter,and the prediction accuracy of the model is proved to be 95.50％.Then,genetic algorithm(GA)is used to search the minimum standard deviation of contact force and the corresponding optimal combination of input parameters of the BP neural network model.The optimization results show that BPNN-GA optimization method can significantly reduce the standard deviation of pantograph catenary contact force and improve the current collection quality of pantograph-catenary.Finally,by comparing the standard deviation of contact force between the optimized catenary and the original catenary at different speeds,the improvement effect of pantograph catenary current collection of the optimized catenary is verified.The optimized catenary can provide guidance for the design,construction and maintenance of the actual catenary.
NETL
NSTL
万方数据
