SiC VDMOSFET Performance Prediction Method Based on Neural Network

Silicon carbide (SiC) vertical double-diffusion MOSFETs (VDMOSFETs) exhibit excellent performance in power electronics. This study employs TCAD simulations to model SiC VDMOSFETs, generating a dataset of 2000 device structures and their simulation results. A neural network model is trained on this dataset to predict key device parameters. Prediction accuracy is evaluated using mean squared error (MSE), mean absolute error (MAE), and R-squared (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>). Furthermore, the trained model facilitates parameter sensitivity analysis. Results demonstrate that appropriate model selection (e.g., CNN) enables accurate prediction of electrical characteristics and effective sensitivity analysis.