Using Solver Parameters the user can specify settings that are used by the solver being used to execute the calculations. When the Parameters menu item is selected from the Solver menu, the following panel appears:
The Solver method is used to specify which solution method is used to execute the calculations. Each method has a different operation mode.
Physical Objects (PO) performs a high frequency approximation based in the combination of optics and full wave methods. It computes the RCS considering only the illuminated subdomains by the incident wave and with the PO current.
Method of Moments (MoM) is an accurate full wave method that can be used to solve many kinds of simulations. The Method of Moments approach is used in subdomains.
A multiprocessing architecture strategy can be selected using the appropiate options as well. MPI is an multiprocessing architecture that works on many computer architectures, but it requires more shared memory than OpenMP, which is a memory architecture that can be selected when the constraints for memory requirements are important.
Electromagnetic equation and Solver Functions can only be set when the solver method have been set to MoM. These settings are covered later on its own section.
The user can set the relative error that is used by the simulation process. This is the maximum delta error accepted by the simulation. The maximum number of iterations can be set as well. The simulation process makes use of an iterative process. This setting can limit the number of iterations performed by the calculation in order to prevent simulations that don't converge simulating infinititely.
When the solver method has been set to MoM, there are extra settings that can be selected in order to configure how the analysis is made during the simulation process.
It is possible to set the electromagnetic equation as well. This option is only available when the solver method has been set to MOM. This electromagnetic equation defines the integral equation to solve during the simulation process. There are three possible equations:
- EFIE makes use of the Electric Field Integral Equation, which solves most of the problems provided a that a good convergence is achieved.
- MFIE makes use of the Magnetic Field Integral Equation. Note that this solver requires geometries to be closed, having their normal vectors pointing outside the objects, in order to obtain valid results.
- CFIE combines EFIE with MFIE -therefore, requirements for MFIE also apply on CFIE. CFIE uses a weighted combination of EFIE and MFIE. The CFIE Parameter sets this weight. CFIE uses the following equation:
CFIE = EFIE · α + MFIE · (1 - α)
Tip: EFIE is best recommended on projects with open surfaces, although it works on closed surfaces as well. CFIE is best suited for closed metallic surfaces. You can use a combined approach using EFIE for some surfaces and CFIE for closed surfaces.
Warning: if the solver uses the CFIE approach, it is mandatory for the normal vectors to be facing outside. Unexpected results and likely errors will happen if normal vectors point inside of volumes.
The selected solver function is ued to set the electromagnetic technique being used through the simulation process:
- If the subdomains option is chosen, the MLFMA-MoM (Multi-Level Fast Multipole Algorithm) will be used. This is the most conventional technique.
- If the macro basis functions (CBFs) option is selected, then the CBFM-MLFMA (Characteristic Basis Function Method - Multi-Level Fast Multipole Algorithm) is used instead. This option improoves the convergence of the MLFMA algorithm, reducing CPU consumption and time since the number of unknowns is reduced.
Tip: CBFs approach is more efficient than the subdomains approach. However, the CBF method has accuracy issues in cavities.
As covered before, it is possible to set the Solver Functions when the solver method has been set to MoM. Solver functions have custom settings that can be changed by selecting the Advanced Options button in the Solver Functions. Pressing this button shows a dialog with two visible tabs:
Main Properties: this tab has settings for changing how the solver works, including which kind of algorithm will the solver use.
Preconditioner: this tab contains available preconditioners for the selected solver.
CBFs Properties: this tab is only enabled when the solver functions are set to CBFs.
The Main Properties tab looks like this:
The following settings are available:
Solver: this sets the algorithm used by the iterative method. The user can set the algorithm either to BICGSTAB (Biconjugate Gradient Stabilized method) or to GMRES (Generalized Minimal Resudial method). Other option is to enable the Direct Solver to improve the efficiency in time when the problem is less than the number of unknows introduced. Direct Solver is enabled only if OpenMP Architecture is selected on the mainSolver window. Direct Solver is not compatible with Speed Up, preconditioners and CBF's blades.
More Options: this group lets the user set other special parameters used by the solver, such as the conductor losses, region size or maximum multipole level. It is possible to enable Speed Up as well, which will reduce the analysis time for problems that have several incident angles at different frequencies. Rigorous Radiation improves efficiency on the solve solution, but is not compatible with the 3D Radiation computation. The 3D Radiation computation is not compatible with Monostatic RCS.
The Preconditioner tab contains the following options:
This tab mainly covers the usage of the Preconditioner, which speeds up the resolution of the problem by selecting one of the possible preconditioners provided.
Diagonal Preconditioner: Use this preconditioner only when the mesh density is 10 div/lambda or highter.
Sparse Approximate Inverse Preconditioner (SAI):
- Sparsity Distance: Set this parameter between 0.2 and 2. Usually hight values increase convergence speed and cpu / memory needs.
- Filtering Threshold: These parameters should contain a value between 0.0 and 1.0. The default values should be adequate in most cases.
- Pre-Processing: This parameter controls the amount of data considered to generate the preconditioner. Lower values entail a more accurate generation, while higher values entail a faster computation.
- Post-Processing: This parameter controls the amount of data to be stored after the generation of the preconditioner. Lower values entail better convergence, while higher values entail less RAM required to store the preconditioner.
- MPI Data Exchange Frequency: This parameter sets up how often the MPI nodes request more coupling terms to generate the preconditioner. Larger values require less interactions speeding up the simulation, although more memory will be needed to store these terms. A negative or 0 value indicates that the coupling terms are only exchanged once.
When CBFs are enabled, the following options are available as well:
These settings let the user enable the preconditioner for speeding up the process, and enabling blades when the simulation is done using a monostatic RCS.
CBF Excitation Drop Threshold: Select a value between 0.0 and 0.1. Lower values give more accuracy.