Manifold Mapping Algorithm Based on High and Low Order Time Domain Spectral Element Method

Abstract

The invention discloses a manifold mapping algorithm based on a high-and-low-order time domain spectral element method. The algorithm comprises the steps that firstly, an electromagnetic structure geometric model including a plurality of physical variables is established, and variable optimization simulation is quickly conducted through a time domain spectral element method with a first-order vector basis function till a simulated frequency response meets a target frequency response; secondly, according to the optimized variables, simulation is conducted through a time domain spectral elementmethod including a second-order basis function, whether the simulated frequency response meets the target frequency response or not is judged, and if yes, optimization simulation is finished; thirdly,if not, a correction frequency response model including the two simulation frequency responses is constructed, fast variable optimization simulation is carried out again through the time domain spectral element method including the first-order basis function till the correction frequency response meets the target frequency response; the second step is repeated after optimization is finished. Themethod has the advantages of being high in optimization speed, high in calculation precision and the like, and an important value reference can be provided for simulation design of a microwave device.

Description

technical field [0001] The invention belongs to the technical field of electromagnetic simulation, in particular to a manifold mapping algorithm based on high and low order time-domain spectral element methods. Background technique [0002] Computer-aided design (CAD) has been widely used in microwave circuit design, and many robust optimization methods have been proposed. However, optimization for EM simulations is usually very time-consuming. The traditional circuit optimization technology is to directly use the electromagnetic simulation response or its derivatives, and use mathematical methods to optimize the response to approach the requirements. Although this method is accurate, the time cost is huge, and it is almost impossible to implement for problems with large and complex parameters. [0003] Generally speaking, choosing an accurate model for simulation calculation can often obtain satisfactory accuracy, but the time consumed by the calculation is greatly increa...

AI Technical Summary

Problems solved by technology

[0006] At present, the operation of the manifold mapping algorithm generally needs to be alternately performed between the two softwares, and the process is relatively cumbersome.

In addition, this algorithm is mostly used for frequency domain simulation, and the simulation efficiency and simulation accuracy are affected to varying degrees when performing broadband optimization design.

Images