Rapid adjoint-based design optimization of compact microwave structures using multi-fidelity simulation models

Design closure of compact microwave structures requires - for the sake of reliability - utilization of full-wave electromagnetic (EM) simulation models. The reason is that due considerable cross-coupling effects in tightly arranged layouts, neither analytical nor equivalent circuit models can provide sufficiently accurate performance evaluation of miniaturized circuits. At the same time, high computational cost of accurate EM analysis makes the dimension adjustment process challenging, especially for traditional methods based on parameter sweeps but also for conventional numerical techniques. In this paper, low-cost optimization of compact structures is demonstrated using gradient search with adjoint sensitivities as well as multi-fidelity EM simulation models. The process is arranged sequentially, with the largest steps taken at the level of coarse-discretization models. Fine tuning of the design is realized with the models of higher fidelity. Switching between the models is controlled by adaptively adjusted termination conditions. This allows for considerable reduction of the design cost compared to single-level optimization. Our approach is illustrated using a compact microstrip rat-race coupler with two cases considered, i.e., (i) bandwidth enhancement, and (ii) minimization of the structure size. In both cases, the optimization cost corresponds to just a few high-fidelity EM simulations of the coupler.