摘要
弓网受流质量是制约和影响高速列车供电质量和运行速度的核心要素,其中弓网之间的接触力标准差是评价受流质量的重要指标。以受电弓参数为输入,弓网接触力标准差为输出设计6因素5水平实验方案。分别通过响应面法(RSM)及神经网络和遗传算法相结合的优化算法(BPNN-GA)构建输入-输出的关系模型,搜寻最小接触力标准差及对应的最优受电弓参数组合。结果表明,两种优化方法均具有良好的预测精度和优化效果。然后通过响应面分析了各输入变量对输出结果的影响。最后采用基于神经网络权值矩阵和Garson方程的评估方法量化了各个输入变量的相对重要程度。
Pantograph catenary current collection quality is the core factor that restricts and affects the power supply quality and operation speed of high-speed train.The standard deviation of contact force between pantograph and catenary is an important index to evaluate the current collection quality.Taking the pantograph parameters as the input and the standard deviation of the pantograph catenary contact force as the output,an experimental scheme with 6 factors and 5 levels was designed.The input-output relationship model was constructed by response surface method(RSM)and the optimization algorithm combined BP neural network with genetic algorithm(BP-GA).The minimum standard deviation and corresponding pantograph combination parameters were searched by RSM and BP-GA.The results show that the two optimization methods have good prediction accuracy and optimization effect.Then,the influence of input variables on output results was analyzed by the response surface results.Finally,the evaluation method based on neural network weight matrix and Garson equation was used to quantify the relative importance of each input variable.
作者
卢琪
LU Qi(State-Key Laboratory of Traction Power,Southwest Jiaotong University,Chengdu 610031,China)
出处
《电工技术》
2022年第15期101-105,151,共6页
Electric Engineering
关键词
受流质量
受电弓
响应面法
BP神经网络
遗传算法
current collection quality
pantograph
response surface methodology
BP neural network
genetic algorithm