Modeling Single Event Transient in 28 Nm FDSOI MOSFETs Using a Neural Network Approach

It's hard to accurately consider various operating factors for the traditional single event transient (SET) SPICE modeling. This paper proposes a novel method based on a neural network. The proposed method can unify the intricate data correlations among drain voltage, linear energy transfer (LET), temperature, strike position, time, and drain transient current in a single model with high accuracy. Technology computer aided design (TCAD) simulation is used to get the original SET data for training. The genetic algorithm (GA) optimized back propagation (BP) neural network established herein has a root mean square error (RMSE) of less than 2.0042%. This optimized neural network is converted to the SET current SPICE model through the Verilog‐A language, and its practicality has been verified through circuit simulation of a two‐input NAND gate.