An expanding-grid algorithm for the finite-difference time-domain method (EM deposition in human body)

Comparison of numerical and experimental methods for determination of SAR and radiation patterns of handheld wireless telephones

“…To this objective, we used the previously described expanding grid algorithm for the FDTD code [Gao and Gandhi, 1992], which has now been adapted for the present problem (Tinniswood et al, submitted). The CAD-derived mobile telephone of Figure 1B was sampled with a resolution of about 1 mm (actually, 0.987 ϫ 0.987 ϫ 1.0 mm) to match the resolution of human body Model A, which was resampled by subdividing the cells by factors of 2 ϫ 2 ϫ 3 in x-, y-, and z-directions, respectively, to obtain Model B.…”

“…Another development for numerical modeling was to use the expanding grid formulation of the finite-difference time-domain (FDTD) method [Gao and Gandhi, 1992], where finer resolution was used for the antenna and the highly coupled region of the ear and the head, while a gradually expanding grid was used for coarser modeling of the more-distant, weakly exposed parts of the head. Together with the recently described truncation algorithm for the model of the head [Lazzi and Gandhi, 1997], this procedure led to a savings of computer memory for numerical calculations of SAR distributions and radiation patterns by a factor of over 20, making it possible to use a workstation instead of a multinode parallel computer (Tinniswood et al, submitted).…”

Empirical validation of SAR values predicted by FDTD modeling*†

“…For far ®eld sources, simulation results have been compared with analytical results for a square [Umashakar and Ta¯ove, 1982], cylinders [Borup et al, 1987;Furse et al, 1990], spheres [Gandhi and Chen, 1992;Gao and Gandhi, 1992;Gandhi et al, 1999], and plates [Ta¯ove et al, 1995]. Calculations of currents induced in a standing human have compared well with empirical data [Chen and Gandhi, 1989;Furse et al, 1997].…”

Exposure Assessment Using the Dual-Grid Finite-Difference Time-Domain Method

“…Expanding-grid (EG), also called graded mesh, has also been proposed [7], [8]. In that case, the spatial resolution is nonuniform all through the volume, which reduces the computation time.…”

“…In that case, the spatial resolution is nonuniform all through the volume, which reduces the computation time. However, the mesh generation must fulfil several requirements to ensure a good accuracy since the nonuniform grid leads to first-order error locally and nonuniform numerical dispersion [7].…”