In order to accurately analyze the test results and decrease the experience requirements for analysts when using the frequency response method to diagnose transformer winding condition

a method for classification of transformer winding deformation based on complex values of frequency response and digital image processing（DIP） techniques was proposed. The frequency response characteristics of normal and deformed windings were represented in polar coordinates on the basis of amplitude and phase angle of frequency response measurement data. Moreover

the Fisher discriminant analysis（FDA） and the gray projection method which is one of digital image processing techniques were used in the proposed classification method to realize the classification of different winding deformation faults. The data of winding deformation were obtained by modeling. Then the proposed method was used to classify 40 groups of test samples

and the results show that the distribution of different types of samples in projection space has obvious distinction andno misclassification occurs. The test results show that the method proposed in this paper has a good classification effect on transformer winding fault

which is beneficial to realize the intelligent diagnosis of transformer winding condition.