[1]贾茹宾,高金峰.基于ARIMA模型的变压器油中溶解气体含量时间序列预测方法[J].郑州大学学报(工学版),2020,41(02):70-75.
 Time Series Prediction Method of Dissolved Gas Content in Transformer Oilased on ARIMA Model[J].Journal of Zhengzhou University (Engineering Science),2020,41(02):70-75.
点击复制

基于ARIMA模型的变压器油中溶解气体含量时间序列预测方法()
分享到:

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
41
期数:
2020年02期
页码:
70-75
栏目:
出版日期:
2020-05-31

文章信息/Info

Title:
Time Series Prediction Method of Dissolved Gas Content  in Transformer Oilased on ARIMA Model
作者:
贾茹宾高金峰
文献标志码:
A
摘要:
变压器油中溶解气体含量是衡量变压器运行状态的重要指标。运用差分自回归移动平均模型(ARIMA)对变压器油中气体含量进行预测,该方法通过python编程以气体含量值对应的时间为索 引输入预测模型,在建模中首先对时间序列平稳性进行单位根检验,采用差分处理的方法将原始不平 稳时间序列转换为平稳时间序列,而后利用自相关函数和偏自相关函数参数选择原则得出若干组模型,在对若干组模型进行优选的过程中分别使用赤池信息、贝叶斯信息、汉南⁃奎因3种准则得出一组最优模型,最后通过相关检验方法对优选模型进行残差检验,并利用满足残差要求的模型对气体含量预测。实验表明,提出的预测方法有较高的预测精度,可以为合理安排变压器的状态检修提供有价值的参考。
Abstract:
The dissolved gas content in transformer oil was an important index to measure the operating state oftransformer. Prediction method of gas content in transformer oil was explored by using differential autoregressive moving average model. In this method, the time corresponding to the gas content value was used .as the index input model to predict the time series by python programming. In the process of modeling, the unitroot test of the stationary time series was carried out by augmented dickey-fuller test statistic. The original non-stationary time series was transformed into stationary time series by differential processing. Then, by using the principle of parameter selection of auto correlation function and partial auto correlation function, several groups of models were obtained, and the akaike information criterion, bayesian information criterion, hannan-quinn criterion, were used in the process of selecting several groups of models. Finally, the residual error of the optimal selection model was tested by the correlation test method, and the gas content was predicted by the model which meets the residual requirements. The experimental results showed that the proposed prediction method has high prediction accuracy and could provide valuable reference for the reasonable arrangement of transformer condition maintenance.
更新日期/Last Update: 2020-05-30