Optimization of computational electromagnetics (CEM) simulation models can be costly in both time and computing resources. Mesh refinement is a key parameter in determining the number of unknowns to be processed. In turn, this controls the time and memory required for a simulation. Hence, it is important to use only a mesh that is good enough for the objectives of the simulation, whether for direct handling of high-fidelity EM models or, even more importantly, for setting up low-fidelity models in a variable-fidelity optimization. On the other hand, in the early stages of an optimization process, a relatively coarse mesh can show whether the governing parameters of the simulation are being appropriately modeled. As the simulation geometry approaches its target, the mesh definition becomes more refined. This letter presents initial results for an approach to identify-ing the minimum acceptable mesh coarseness based on the projected evolution of FSV's global difference measure when a model is refined from a very crude representation rather than the more usual high-fidelity model. Future work to verify the generality of this letter could provide substantial savings in time and effort for CEM analysis in EMC.
Authors
- Alistair Duffy,
- Gang Zhang,
- prof. dr inż. Sławomir Kozieł link open in new tab ,
- Lixin Wang
Additional information
- DOI
- Digital Object Identifier link open in new tab 10.1109/temc.2015.2417523
- Category
- Publikacja w czasopiśmie
- Type
- artykuł w czasopiśmie wyróżnionym w JCR
- Language
- angielski
- Publication year
- 2015