Novel Strategies for Efficient Computational Electromagnetic (CEM) Simulation of Microstrip Circuits, Antennas, Arrays, and Metamaterials Part-II: Characteristic Basis Function Method, Perfectly Matched Layer, GPU Acceleration

As mentioned in Part-I [1], rapid prototyping plays a critical role in the design of antennas and related planar circuits for wireless communications, especially as we embrace the 5G/6G protocols going forward into the future. Existing commercial software modules are often inadequate for this task in the millimeter-wave range since the memory requirements and runtimes are often too high for them to be acceptable as design tools. Using approximate equivalent circuit models for various components comprising the antenna and the feed system is not the answer either, because these models are not sufficiently accurate. Consequently, it becomes necessary to resort to the use of more sophisticated simulation techniques based on full-wave solvers that are numerically rigorous, albeit computer-intensive. Furthermore, optimizing the dimensions of antennas and circuits to enhance the performance of the system is frequently desired, and this often exacerbates the problem since the simulation must be run a large number of times to achieve the performance goal, namely an optimized design. Consequently, as pointed out earlier, it is highly desirable to develop accurate yet efficient techniques, both in terms of memory requirements and runtimes, to expedite the design process as much aspossible. In the first part of this paper [1], we presented three strategies to address these issues, mostly related to Green’s Functions of layered media. We have shown that the proposed techniques are not only useful for antennas and printed circuits on layered media but also for antennas embellished with metamaterials for the purpose of their performance enhancement. In this sequel to Part-I, we present several other Efficient Computational Electromagnetic (CEM) simulation strategies for expediting the runtime and improving the capability of handling large problems that are highly memory-intensive. These include a domain decomposition technique, which utilizes the Characteristic Basis Function Method (CBFM); the T-matrix approach which is also useful for hybridizing Finite Methods (FEM or FDTD) with the Method of Moments (MoM); Mesh truncation in Finite Method by using a conformal Perfectly Matched Layer (PML); and Graphics Processing Unit (GPU) acceleration of MoM and FDTD codes.