Abstract: As the most expensive and important equipment in substation, transformer has always been the key object of research. At present, there are few researches on the transformer noise source identification. Therefore, this research combines the beamforming technology to carry out the related research on the transformer noise source identification. Firstly, according to the investigated transformer noise characteristics, the concerned noise frequency band and the sound source identification method are determined, the direction of arrival estimation of wideband signal and the related theory of beamforming are expound, and then a wideband transformer signal location idea is sort out based on frequency domain data. The simulation results confirm the rationality and accuracy of this research. Finally, a set of transformer noise source identification system is built based on LabVIEW platform, and the positioning effect of the system is tested through the field tests of the actual operating transformers under different conditions. The results prove that the noise source identification idea and the system platform built have good reliability and availability.