Numerical investigation of acoustic emissions distribution from partial discharge in transformer

A new concept is presented that extrinsic Fabry-Perot interferometer (EFPI) sensors are designed to detect partial discharge (PD) in liquid insulation, and numerical simulation is used to analyze the distribution characteristics of acoustic emissions generated by partial discharge. The sensing head of EFPI mainly consists of an optical pigtail and a thin silica glass diaphragm. Based on wave propagation theory, numerical calculation by finite element method (FEM) is used to calculate the distribution of acoustic emissions generated by partial discharge. A 3-D model of 35kV transformer is built, which is the same scale with real equipment. During the simulation the PD sources are replaced by acoustic sources. The acoustic propagation properties and enhancement of detection sensitivity are studied according to the simulation results. It is presented that there will be an acoustic intensity region built around the solid medium in liquid-solid media. Area sizes of higher intensity region are affected by Young's modulus of solid. The higher acoustic intensity region will become larger with the Young's modulus of solid increasing. The results show that EFPI sensors placed around the solid medium will be more sensitivity to detect PD acoustic signals than other regions in the transformer.