With the development of intelligent sensing technology, a large number of partial discharge (PD) time-domain waveform image data are generated in the field of gas insulated combined appliance (GIS). Traditional pattern recognition methods cannot identify the defect types of such data. At the same time, the deep learning method for GIS PD pattern recognition is generally faced with the problem of small samples. In order to solve the above problems, this paper proposes a pd pattern recognition method based on transfer learning and DenseNet model. Compared with other network structures, DenseNet model has higher accuracy and training time end. This study can be used to diagnose pd defects in GIS equipment.

GIS,pattern recognition,transfer learning,deep learning