变压器绝缘油中产气趋势突变的检测方法研究

Detection Method Study on Mutation Trend of Gas Production in Transformer Oil

提出一种检验绝缘油中产气监测数据突变趋势的方法,发现变压器的潜伏性故障。该方法主要以非参数统计检验法计算监测数据时间序列的正态分布统计变量,比较正序与负序变量曲线的关系,检验统计变量的发展趋势及时域范围,并给出置信度。实际案例分析表明:该方法可描述监测数据的起始变化、趋势显著、趋势最大及趋势稳定4种渐进变化过程信息,适用于分析低浓度气体及微小变化值,有利于发现变压器劣化迹象。

A method for detecting the mutation trend of gas production data in transformer oil was proposed to find latent defects of transformers. The method used non parametric statistical test method to calculate the normal distribution statistical variable of time series of monitoring data, determined the trend and range of statistical variable, and provided the confidence level. The case analysis shows that the method can accurately describe the origin state, significant, maximal, and stable development trend of monitoring data. It is suitable for the analysis of low concentration gas and small changes, and is benifi- cial to find the early signs of transformer defects.