摘要
微风振动是架空线路时常出现的现象,长期振动会造成导线断股断线、金具脱落等危害。针对现有微风振动在线监测传感器测量误差较大的问题,设计了一种微风振动在线监测数字传感器,传感器在采用悬臂梁式位移计的基础上,将线性回归的方法和BP神经网络算法应用到传感器的非线性标定当中,并进行了传感器标定实验。研究表明,线性回归的标定算法比神经网络的标定算法精度更高,标定后最大相对误差为1.93%。根据研究结果可知,微风振动传感器由频率不同带来的非线性误差,可以采用线性回归的方法进行补偿,能够提高传感器的测量精度,从而能为导线状态检修提供更加可靠的参考依据。
Aeolian vibrations often occur on transmission lines,which can result in fatigue damage to the conductor and metal fittings. To solve the problem of low accuracy,a kind of digital sensor for monitoring aeolian vibration is designed in this paper. Cantilever beam is used in the sensor to measure the bend amplitude. Linear regression and BP neural network are applied to calibrate the sensor. In the end,the experiments of calibration show that linear regression algorithm has higher accuracy compared with BP neural network. The maximum relative error of the former is 1.93%. It can be concluded from the results that linear regression method can be used to compensate the error causing by varying frequency. And the sensor can provide a more reliable reference for lead state maintenance.
出处
《电网与清洁能源》
北大核心
2015年第6期1-5,共5页
Power System and Clean Energy
基金
陕西省重点科技创新团队计划项目(2014KCT-16)~~
关键词
输电线路
微风振动
传感器标定
线性回归
BP神经网络
transmission line
aeolian vibration
sensor calibration
linear regression
BP neural network
