Parasitic Parameter Prediction for Planar Transformers Based on Neural Network

Abstract Parasitic parameters such as leakage inductance and distributed capacitance of planar transformers have a direct impact on the performance efficiency transformers. Traditional methods for parasitic parameter prediction are commonly based empirical formulas or simulation software, but they problems high computational complexity, time-consuming low accuracy. In this paper, method predicting multilayer perceptron (MLP) under specific winding structure is proposed, which can improve transformer design. The experiments demonstrate that model effectively predict inductance, capacitance, AC loss