AI Optimization of High-Power Microwave Filters

Accurate modeling and optimization of high-power microwave filters necessitates 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.