《White_Paper_Meshing in TD_gro》.pdf

Understanding Time Domain Meshing in CST MICROWAVE STUDIO® When performing numerical solutions of Maxwell’s Equations in electromagnetic simulations, it is well known that the quality and type of mesh used for discretizing the computational volume plays a critical role in simulation speed and accuracy. In fact, the choice of a particular meshing technique is often determined by the trade- offs between speed, accuracy and memory requirements. In this respect, meshing in the time domain stand outs because it offers a number of advantages over other existing techniques such as those based on Finite Elements (FEM). In particular, it offers more flexibilit for achieving a good compromise between simulation speed and accuracy. It is less memory demanding and it has a linear scaling of computational effort (the simulation time increases in the same proportion as the number of mesh cells), which is the best one can hope for in a numerical technique. For many applications, this translates into the ability to simulate larger problems using fewer computational resources in a shorter time. The purpose of this paper is to provide our CST MICROWAVE STUDIO® user community the required meshing knowledge and techniques so they can achieve the aforementioned benefits. Both automated and “manual” meshing techniques are discussed in detail. 1. Meshing Basics In time domain methods, the type of mesh most commonly used is the so called hexahedral mesh. A hexahedral mesh is that where the computational volume (simulation space containing our model) is discretized by means of variable size rectangular cuboids (in general, ∆x≠∆y≠∆z). Each cuboid is referred to as a mesh cell. See Fig. 1. a)