文档介绍:电工技术 理论研究
基于 LSTM 的换流变油温预测方法
廖 毅1 ,陈奕海1 ,李亚锦2
(1.中国南方电网超高压输电公司广州局,广东 广州 510405;2.山东大学电气工程学院,山东 济南 250061)
摘 要:换流变的负载能力与其油温密切相关,对油温的准确预测可为换流变绕组热点温度及换流变的安全运行评
估提供辅助依据。受直流运行功率、环境温度等多源因素的影响,对油温单一的预测方法难以实现较精准的预测。为
解决上述问题,提出基于 LSTM 算法的换流变油温预测方法,考虑油温、直流运行功率、环境温度信息,构建了油
温预测模型,实现未来 1 天内换流变油温预测。通过实例分析验证了该方法可更好地预测换流变油温。
关键词:换流变;LSTM;油温预测
中图分类号:TM406 DOI:10.19768/j.cnki.dgjs.2021.09.008
Prediction Method of Converter Oil Temperature Based on LSTM
LIAO Yi1 ,CHEN Yihai1 ,LI Yajin2
(1.Guangzhou Bureau,China Southern Power Grid EHV Power Transmission Co.,Guangzhou 510405,China;
2.School of Electrical Engineering,Shandong University,Jinan 250061,China)
Abstract:The load capacity of the converter is closely related to the oil temperature,and the accurate prediction of the oil
temperature can provide auxiliary basis for the hot spot temperature of the converter winding and the safe operation evalu-
ation of the converter.However,the current single prediction method for oil temperature is affected by DC operating pow-
er,environmental temperature and other factors,so it is difficult to achieve a more accurate prediction.To