Distribution Pattern and Differentiation Thresholds of Gas Content in Large Oil-Filled Equipment

In the existing threshold calculation methods, there are problems such as establishing distribution models based on large group data of different types of oil-filled equipment, or specifying the threshold using normal dissolved gases analysis (DGA) data of some individuals while fault data of others, etc. In order to address these problems, this paper proposes a differentiation gas content threshold calculation method applicable to large oil-filled equipment individuals. The method is based on the gas content normal data and fault data of individuals of oil-filled equipment, establishes parameter model based on the Weibull distribution to fit the data, and uses the inverse cumulative distribution function and the P-R curve to calculate and select the reasonable thresholds as the attention value and warning value. The differentiated thresholds are obtained for five kinds of large oil-filled equipment individuals, and compared with the traditional threshold calculation method, the thresholds obtained by this method are more effective in identifying defects and faults, and can significantly reduce the occurrence of false alarms and omissions.