To comprehensively investigate the intricate relationship between transformer winding deformation categories, deformation degrees and winding electrical quantities, and to extract more winding information from electrical parameters, an explicit function model employing the stochastic response surface method(SRSM) is established. This model integrates three?winding transformer electrical parameters and structural variables, facilitating a deeper understanding of their interplay, and act as objective function for the sensitivity analysis of transformer. A dual?layer global sensitivity analysis model based on the Morris?Sobol method is proposed to analyze the objective function and identify the deformation types most sensitively reflected by each electrical parameter. The Morris method excludes the impact of irrelevant structural variables, increasing the efficiency and accuracy of sensitivity analysis. The Sobol method is subsequently utilized to assess the sensitivity of electrical parameters to different deformation types. It is found that the sensitivity of capacitance parameters to winding radial deformation is much higher than that of inductance parameters. Therefore, the variation in winding capacitance is adopted as the criterion for detecting winding deformation. and the relationship between the change rate of winding test capacitance and equivalent deformation is established. This approach not only enhances understanding but also paves the way for effective detection of winding deformation based on electrical parameters.