Mechanical State Assessment Method for Transformer Windings under Multiple Short-Circuit Impacts Based on Gaussian Mixture Model Clustering

Under the cumulative effect of multiple short-circuit impacts, transformer windings are prone to axial clamping force relaxation and structural micro-deformation, posing a severe threat to equipment safety. To achieve accurate assessment of the winding mechanical state, this paper first constructs a mathematical model of winding axial vibration and elucidates the nonlinear relationship between clamping force and equivalent winding stiffness. Subsequently, a short-circuit impact test platform is established, and tests with varying current intensities are conducted on a 110 kV transformer. Transient vibration entropy, main frequency ratio, and vibration skewness are extracted as key state feature vectors. The vibration features are then analyzed using the Gaussian Mixture Model (GMM) clustering algorithm. Experimental results indicate that the evolution of GMM cluster centers effectively characterizes the progressive deterioration process of the winding mechanical state, providing significant engineering value for transformer condition monitoring and fault early warning.