Computationally Efficient Estimation of the Electric-Field Maximums for the MFT Insulation Coordination

This paper proposes a methodology for computationally-efficient estimation of the local E-field maximums within the transformer winding insulation material for design optimization purposes, where besides the estimation accuracy, the computational cost represents an equally important figure of merit. Except for a limited class of very simple symmetric geometries, it is not possible to analytically solve or approximate these phenomena with acceptable accuracy. While it is possible to very accurately model the electric field distribution via some computationally intensive numerical method, such as finite elements method (FEM), the execution time and numerical stability are often limiting factors when it comes to a multi-variable optimization. To that end, this paper proposes a suficiently-accurate (error < 5% referred to FEM) modeling methodology for parasitic capacitance and local E-field maximum estimation, specially designed for very fast execution - more than four orders of magnitude faster compared to FEM.