基于最小二乘和BP神经网络算法的转辙机测力方法探究

Research on Force Measurement Method of Switch Machine Based on Least Squares and BP Neural Network Algorithm

中文关键词:  转辙机  非接触式  实时监测  最小二乘  BP神经网络

英文关键词:switch machine  non-contact  real-time monitoring  least squares  BP neural network

中文摘要:

      目前，转辙机牵引力测量多采用销式力传感器代替动作杆与道岔连接销的方法，该方法存在较大的安全隐患，只能临时测量，不能实时测量转辙机牵引力。因此以S700K为研究对象,在前人研究的基础上，提供一种能够实现实时测量转辙机牵引力的非接触式测方法。采用高精度激光位移传感器采集转辙机动作杆速度信息，运用齿轮等效杠杆法结合所测转辙机动作杆速度信息将电机齿轮输出力转换为转辙机牵引力。为提高测量数据的准确度，通过最小二乘法对数据进行初步补偿，再通过BP神经网〖JP2〗络对数据进行二次补偿。实验证明，经双重补偿后的数据平均误差为7%，较为精确地测量了转辙机拉力。

英文摘要:

      At present,pin force sensor is used to measure the traction force of the switch machine instead of the connecting pin between the action rod and the switch.This method has great potential safety hazards and can only measure the traction force of the switch machine temporarily rather than in real time.Therefore,taking S700K as the research object,based on previous studies,a non-contact measurement method is provided,which can realize real-time measurement of traction force of switch machine.The high precision laser displacement sensor is used to collect the information of the speed of the switch mechanism lever,and the gear equivalent lever method is used to convert the output force of the motor gear into the traction force of the switch machine.In order to improve the accuracy of the measured data,the least square method is used to make preliminary compensation for the data,and then the BP neural network is used to make secondary compensation for the data.Experimental results show that the average error of the data after double compensation is 7%,and the tension of the switch machine is measured more accurately.