As the development of power equipment operation detection and monitoring technology, the transformer as a key master device , the detection of insulation defects inside the equipment has become the focus and direction of the research. However, due to the complexity of the internal structure of the transformer, the difficulty of reproduction of different insulation defects, the difficulty of identifying the local discharge signal interference signal and other issues, the level of the transformer internal insulation defects detection, analysis and identification is still not high at present. In order to explore the effective detection method, this paper has carried out targeted research work. Based on the transformer failure simulation mode, research on Simulation Technology of Transformer Internal Insulation Defect has been carried out, two kinds of transformer fault hybrid models are designed to realize the accurate reproduction in the real environment of transformer internal defects. Through the composite signal acquisition system based on HFCT, UHF, AE, optical four kinds of sensors, research on Signal Characteristics Based on Different Signal Principle in Transformer Partial Discharge has been carried out. At the same time, the correlation analysis of different discharge signal characteristics based on the acquisition signal is carried out to realize the comprehensive analysis of multi-signal characteristics and further improved the accuracy of PD signal identification. The research shows that the PD signal acquisition system can accurately measure the partial discharge signal in the transformer and reflect the signal characteristics of all kinds of signals. By comparing the signal characteristics and the energy distribution characteristics of the local discharge under the single PD model and the hybrid PD model, it provides a basis for the future study of the hybrid partial discharge pattern recognition of the transformer.