733 / 2022-03-31 19:11:00
On-line vacuum degree monitoring of vacuum circuit breaker based on laser-induced breakdown spectroscopy combined with random forest algorithm
vacuum circuit breakers,on-line vacuum degree monitoring,laser-induced breakdown spectroscopy,Random Forest
终稿
Feilong Zhang / Xi'an Jiaotong University
Huan Yuan / Xi'An Jiaotong University
Aijun Yang / Xi'An Jiaotong University
Xiaohua Wang / Xi'an Jiaotong University
Dingxin Liu / Xi‘an Jiaotong University
Mingzhe Rong / Xi'an Jiaotong University
As an important switchgear in power system, vacuum circuit breaker is more and more widely used in medium voltage field with its advantages of strong arc extinguishing ability, no pollution and compact structure. The traditional off-line vacuum detection means cannot meet the demand of vacuum breaker reliability monitoring in actual engineering. In this study, we propose a method for online vacuum monitoring of vacuum circuit breakers based on laser induced breakdown spectroscopy (LIBS) combined with variable importance random forest (VI-RF) algorithm. The experiments use the LIBS platform to collect the spectral data of the target material in the vacuum cavity under 9 different pressure conditions from 10-3 Pa to 10Pa, establish the random forest model for calculating the pressure level of the spectral data, select the spectral lines of the target material and four elements in the ambient gas from the original data as the data set to input the RF model for training, and use different data pre-processing methods (standard normalization, multivariate scattering correction, first-order derivative, wavelet transform) to increase the classification ability of the model, correlation coefficient (R2) and root mean square error (RMSE) as evaluation indexes, and the VI-RF algorithm was used to eliminate redundant information, which improved the operation efficiency and accuracy of the algorithm. The effects of the number of decision trees(ntree), and the number of features to be selected(mtry) on the accuracy of air pressure estimation are investigated. The results show that the accuracy of this method exceeds 99% and can reach the accuracy of traditional vacuum monitoring technology, which proves that LIBS technology combined with VI-RF algorithm is a new method that can be used for online vacuum monitoring of vacuum circuit breakers.
重要日期
  • 会议日期

    09月25日

    2022

    09月29日

    2022

  • 08月15日 2022

    提前注册日期

  • 09月10日 2022

    报告提交截止日期

  • 11月10日 2022

    注册截止日期

  • 11月30日 2022

    初稿截稿日期

  • 11月30日 2022

    终稿截稿日期

主办单位
IEEE DEIS
承办单位
Chongqing University
移动端
在手机上打开
小程序
打开微信小程序
客服
扫码或点此咨询