Modelling wire antennas in wider simulations

The thin wire models currently available in TLM and FDTD have been found to give comparable results to MoM techniques, except in close proximity to the TLM thin wire at low frequency. Thus, it is concluded that simple wire antennas can be efficiently represented in TLM and FDTD models. The performance of the TLM models was seen to be influenced by boundary proximity, but these effects have not yet been investigated in FDTD. Based on the evidence obtained so far the FDTD implementation of the thin wire model would appear to be more reliable than the TLM version, but this may be due to more effective absorbing boundaries in FDTD. The TLM tool is currently much more flexible than the FDTD code since it permits cell grading in three dimensions. This eases the problem of matching cells to rectangular geometries whilst minimising the demand on computational resources. A related feature of the TLM code is a multi-grid facility, which allows cells to be lumped together in regions where field variation is not expected to be significant. Nonetheless, care is needed when exploiting this facility, since the largest dimensions of such combined cells determine the high frequency limit for the model. This capability to make limited adjustments to the mesh makes the modelling of more complex structures, such as the log-periodic dipole array, much more practicable using TLM. At present the only alternative with the FDTD code is to reduce the cell size, which is associated with a significant increase in the demand for computational resources. The `staircase´ approximation cannot yet be avoided in TLM and FDTD for boundaries which are curved or inclined with respect to the mesh axes. However, enhancements have now been added to the TLM code to correct for errors associated with these discretisation artifacts for inclined wires, and to eliminate the need to add a metal block at wire bends