Development and calibration of a network-based streaming electrification model for core-form transformer structures

Streaming electrification occurs in large power transformers when oil circulated for cooling purposes acts to separate charges from the Debye layer on the cellulose insulation surfaces. This phenomenon is believed to have caused several catastrophic transformer failures and remains to be satisfactorily understood in realistic geometries. A network-based streaming electrification model is presented which utilizes single winding ducts as the basic element to investigate a core-form structure. The network is first analyzed using hydraulic duct parameters to calculate the oil flow distributions within the structure. This data is then provided to the electrification solver which predicts the charge densities at each duct outlet. The charge density results are ultimately used to determine the streaming potentials at each node based on an a suitable matrix of leakage resistances. Development of the network model is presented along with empirically-based calibration data for a realistic section of core-form transformer geometry.