A Recurrent Neural Networks Based Modeling Approach for Internal Circuits of Electronic Devices

In this paper, a modeling approach is developed for internal circuits of electronic devices. Two types of recurrent neural networks (RNN), both with and without time sequence, are trained to learn the dynamic responses of interferences in frequency and time domain respectively. After training, the RNN model provides fast evaluation of interference responses of the original internal circuits, which is useful for electromagnetic susceptibility (EMS) analysis and optimization of electronic devices. Two examples are provided to demonstrate the validity of the proposed modeling approach.