| Abstract |
This paper proposes a modeling and parameter design method for DC-DC converters based on artificial intelligence (AI). Initially, a database of switching losses is constructed using Spice simulation data with a single fixed semiconductor switch. Next, an artificial neural network (ANN) is trained by the database. Then Transfer Learning (TL) is implemented to train other ANNs for other switches with much less training data needed. Finally, under the restrictions of current and voltage ripples, a heuristic optimization algorithm is used to obtain the most efficient and optimal design. The results show that the ANN models give precise estimates of the converter properties. |
