基于改进BP神经网络的驼峰场车辆减速器故障诊断的研究

减速器是驼峰场控制速度的主要部分,随着当前高速铁路大发展和铁路货运量的提高,编组站解体和编组能力加大,车辆减速器使用率和故障率增加,而现场维修人员凭借经验的低效率维修已经不能满足当前的货运溜放要求,对驼峰溜放控制,钩车安全连挂提出更高的要求,因此根据现场收集的数据建立BP神经网络模型进行仿真训练,精确诊断驼峰车辆减速器TJK(Y)2,3的故障部位,仿真结果表明,故障判断准确率达到96%。

Research on Fault Diagnosis of Hump Car Retarder Based on Improved BP Neural Network

The retarder is the main part of the control speed system of the hump yard. With the development of high speed railway and the increase of railway freight volume,the capacity of break-up and make-up of the marshalling yard is increased and the utilization and fault of the retarder is increased simultaneously,while the low efficiency of the maintenance workers fails to meet the current requirements of freight humping,which addresses more on humping control and effective coupling. Therefore,based on the field data,BP neural network model is established for simulation training and the fault location of TJK( Y) 2,3 of hump retarder is diagnosed precisely. The results show that the fault diagnosis accuracy rate reaches 96%.