摘要
目前变压器油中气体浓度预测普遍采用灰色模型,但灰色模型的使用存在一定局限性。为提高预测的精度和可靠性,应用最小二乘支持向量机(least squares support vector machine,LS-SVM)理论建立了同时预测变压器油中7种主要特征气体(氢气、甲烷、乙烷、乙烯、乙炔、一氧化碳和二氧化碳)的预测模型。该模型既综合考虑了气体之间的相互影响,又充分发挥了LS-SVM解决有限样本问题的优势, 具有较高的预测精度和泛化能力。实例分析验证了该模型的有效性。
At present available prediction models of gases dissolved in transformer oil are grey models, but there are certain limitations for them. To improve accuracy and reliability of the forecasting, a simultaneously forecasting model for seven gases dissolved in transformer oil, namely acetylene, hydrogen, ethane, methane, ethylene, carbon monoxide and carbon dioxide, is built by use of least square support vector machine （LS-SVM）. In the built model the interaction among these gases is comprehensively considered, and the superiority of LS-SVM in processing finite samples is fully brought into play. This model possesses high forecasting accuracy and generalization ability, its effectiveness is verified by case analysis.
出处
《电网技术》
EI
CSCD
北大核心
2006年第11期91-94,共4页
Power System Technology
关键词
最小二乘支持向量机
溶解气体分析
变压器
高电压与绝缘技术
least square support vector machine （LS-SVM）
dissolved gas analysis
transformer
high voltage and insulation technology
作者简介
肖燕彩（1972-）,女,博士研究生,讲师,研究方向为检测技术与故障诊断;
陈秀海（1969-）,男,硕士,高级工程师,研究方向为电力系统自动化;
朱衡君（1950-）,男,教授,博士生导师,研究方向为检测技术与故障诊断.
