AI optimization of high-power microwave filters.

Accurate modeling and optimization of high-power microwave filters necessitate the use of Multiphysics-based approaches, including electromagnetic, thermal and mechanical stress coupled analysis. However, such modeling and optimization approaches are computationally prohibitive. This talk presents AI/machine learning oriented space mapping approaches to exploring Multiphysics based data (as fine data) augmented with electromagnetics-based data (as coarse data) for training Multiphysics oriented models. Parallel and space mapping-based approaches for Multiphysics oriented optimization of high-power microwave filters are also formulated and examples presented.