The Feko 2022.0 release features extensions to many different solvers - enhancing Feko’s capabilities in propagation modelling, network planning as well as general and specialist EM analysis application areas. There are various performance improvements, interoperability and API extension, enabling more flexible workflows and automation.

Propagation modelling and network planning

The Longley-Rice propagation model used in WRAP has been enhanced and updated to account for terrain profile.

A new hybrid ray-tracing technique is employed to improve accuracy results when using the SBR solver for propagation modelling.

A ‘Soft Break’ in the WRAP Cost and Coverage Optimizer allows the user to inspect the best solution found so far by the optimizer before deciding whether to continue with the optimization search or to use the best solution already available.

The impact of Doppler Shift due to rotation around any axis and movement in any direction can be evaluated and included in propagation modelling simulations.

General EM solver extensions and performance improvements

The newFASANT MoM solver now supports the usage of the Characteristic Basis Function Method (CBFM) for non-RCS calculations.

Improved creeping wave accuracy , as well as additional optical effects (corner/tip diffraction) and higher-order effects (multiple reflections plus 1 wedge diffraction), are available for faceted UTD simulations.

The faceted UTD solver may now be used in both coupled and uncoupled simulations including parts solved with the MoM. This extends the hybrid simulation options for Feko solvers. Receiving antennas defined using far field, near field or spherical modes may also be used in simulations involving faceted UTD .

Improvements to the application of Shared Memory in the MLFMM solution have reduced the parallel memory footprint considerably, resulting in improved parallel efficiency (both runtime and memory) for highly distributed simulations (many nodes/cores). For large MLFMM simulations where memory resources may be insufficient for certain nearfield computations, a chunking approach has been implemented, reducing the memory requirement. 

When using MLFMM, faster far-field calculation times and better scaling (when the number of parallel processes is increased) are achieved by optimization of the number of low-level MPI communications.

A new workflow allowing for accurate and fast simulation involving more complex realistic multi-ply radomes including frequency selective surface (FSS) structures is now supported in newFASANT. The radome material is first characterized in terms of reflection and is then applied as a material to the relevant surface when configuring the full radome analysis.

Automation, interoperability and results analysis

The WinProp API has been extended to allow automation of the generation of 5G Beam Patterns as well as FM-CW Radar Post Processing.

Maps can now be exported from WRAP to ASCII grid files and the use of Touchstone files when configuring Collocation Interference analysis is now supported.

Nearfields calculated using Feko on a Cartesian boundary may now be imported as a source for simulations using newFASANT.

3D radiation patterns of antennas may be shown along with coverage results to improve visual interpretation of propagation simulation results.

Results generated using the Feko parameter sweep script may now be plotted on a surface graph in POSTFEKO with a sweep parameter as the independent axis.

An application macro script has been added that allows for the iterative application of the CADFEKO automatic mesh refinement for simulations involving FEM. Refinement parameters (such as number of refinement iterations, maximum target error level etc.) can be controlled using variables in the model allowing for the script to be run in a non-interactive fashion. The volume refinements have also been improved.

The plugin which enables farming of RCS calculation steps to a simulation cluster been extended to support user-specification of the number of chunks required for farming. These chunks may be split over angle and/or frequency, catering for multi-frequency and single-frequency RCS simulations as well as non-RCS multi-frequency simulations.

Cable modelling validity checks have been extended to validate factors such as the minimum pitch length of twisted pairs, the distance to ground for Individual signals, validity of cable interconnecting pin definitions and inherent frequency limitations of braided shield models employed in calculating transfer impedances.