Capacitive voltage transformer (CVT) are widely used measurement equipment in high-voltage power grids. Real-time online assessment of the internal insulation state of CVT will help maintain the safety and stability of power systems. However

the existing online assessment methods of CVT internal insulation state based on group correlation do not consider the influence of primary voltage fluctuations

resulting in insufficient reliability of the assessment results. Aiming at this problem

an on-line evaluation method of CVT internal insulation state considering the influence of primary voltage fluctuation is proposed. This method uses the intrinsic correlation between CVT populations and principal component analysis method to extract features from the output signals of the population to exclude the primary voltage fluctuation

and the characteristic information that characterizes the insulation anomaly in the CVT is obtained. And then

the random forest method is used to analyze the characteristic information to realize the real-time online assessment of the CVT internal insulation state. The experimental analysis shows that this method can effectively eliminate the influence of primary voltage fluctuation on the online evaluation of the CVT internal insulation state

and realize the accurate evaluation of the CVT internal insulation state.