用改进的灰色多变量模型预测变压器油中溶解气体的浓度  

Prediction of Gas Dissolved in Power Transformer Oil by Improved Multivariable Grey Model

作　　者：肖燕彩[1] 陈秀海[2] 朱衡君[1] 

机构地区：[1]北京交通大学,北京市海淀区100044 [2]北京电力公司,北京市西城区100031

出　　处：《电网技术》2006年第10期86-89,共4页Power System Technology

摘　　要：油中溶解气体分析(dissolved gas analysis,DGA)是评估变压器绝缘状态和分析变压器绝缘故障的重要依据。灰色多变量预测模型可使用从同一信号中提取的多个诊断指标同时进行预测,从系统的角度对各特征参数进行统一描述, 因而能够获得比较精确的预测结果。在现有文献研究成果的基础上,对灰色多变量模型做了进一步改进,将灰色多变量模型中的已知条件做了适当的拓广,提出了新的预测方法, 从而使新模型具有更高的精度。变压器油中七种特征气体的预测实例分析验证了该方法的有效性。Dissolved gas analysis （DGA） is the essence to evaluate the state of transformer insulation and analyze transformer insulation faults. With multi-variable grey model the diagnosis indices extracted from same signal can be simultaneously used to perform fault prediction and in the viewpoint of system the unified description of characteristic parameters can be realized, so a more accurate prediction result can be obtained. Based on references, the multi-variable research results from existing grey model is further improved, the known conditions in multi-variable grey model is expanded properly and a new prediction method is proposed, thus the prediction results by improved model will be more accurate. Actual case of seven featuring gases dissolved in the power transformer oil is employed to demonstrate the validity of the proposed approach.

关 键 词：多变量模型 灰色模型 变压器 油中溶解气体分析 高电压绝缘技术 

分 类 号：TM855[电气工程—高电压与绝缘技术]