Breakdown Voltage Prediction of SOI Lateral Power Device Using Deep Neural Network

Breakdown voltage is an essential indicator for evaluating the performance of power devices. The early prediction of the breakdown voltage is beneficial for subsequent device design closure. In our paper, we propose a method to predict the breakdown voltage for SOI lateral power device using machine learning. Our prediction scheme comprises a deep neural network (DNN) model with four hidden layers. The experimental results show that the proposed approach achieves 8.6% average prediction error for the breakdown voltage compared with the result from Medici simulation, while achieving a 5.4E6x speedup.