Machine learning based prediction model for single event burnout hardening design of power MOSFETs.

Aiming at the interaction between different hardening techniques and the search for the optimal hardening scheme, we propose a machine learning based prediction model for single event burnout (SEB) hardening design of power MOSFETs in this paper. The feedforward neural networks are used to build the prediction model, which covers at least three commonly used SEB hardening techniques considering their interaction and predicts the variations in the radiation tolerance and the critical electrical parameters under the complex hardening conditions. Then, based on the prediction model, we can search for the optimal value of each hardening parameter to obtain the optimal hardening scheme, for which the radiation hardness can be greatly improved while keeping the critical electrical parameters at an acceptable level using the search algorithm. The prediction model proposed in this paper provides a new research method for the radiation hardening design of aerospace electronic devices, and it effectiveness has been successfully verified by the TCAD simulations. [ABSTRACT FROM AUTHOR]