Simulation parameters are located in python files, themselves located in the "Puma-EM/parameters" directory. There are three parameters files: parameters.py, RCS_parameters.py and TDS_parameters.py. The parameters contained therein are explained hereafter.

parameters.py is a default file. The user should not change it. Simulation parameters are best changed in RCS_parameters.py, the front-end file for parameters.

Default and/or typical parameters values are given next to the parameter name.

params_simu.pathToTarget = "./geo"

the absolute or relative path to the target geo file location

params_simu.targetName = "cubi"

the name of the target, for example "cubi", "sphere", "boat_12", etc. The geo file has to exist of course!

params_simu.targetDimensions_scaling_factor = 1.0

the target scaling factor, needed when sometimes targets are designed in cm or mm units, instead of meters

params_simu.z_offset = 0.0

the vertical offset of the target

params_simu.lx = 0.2

parametric x dimension of the target. It will have an effect where applicable, i.e. for geo files that have lx as a parameter, otherwise it has no effect

params_simu.ly = 0.2

parametric y dimension of the target

params_simu.lz = 0.2

parametric z dimension of the target

params_simu.MONOSTATIC_RCS = 0

1 if we want to compute the Monostatic RCS, 0 otherwise

params_simu.FAR_FIELD = 1

1 if we want to compute the Far Fields, 0 otherwise

params_simu.f = 2.12e9

the frequency at which we want to solve the problem

params_simu.eps_r = 1. + 0.j

relative dieletric permittivity of the host medium

params_simu.mu_r = 1. + 0.j

relative magnetic permeability of the host medium

params_simu.TDS_APPROX = 0

do we use the Thin Dielectric Sheet approximation (for imperfect conductors). Use 1 if you intend to use this approximation

params_simu.Z_s = 0.0 + 0.j

the surface impedance used in the TDS approximation

params_simu.EXCITATION = 'dipole'

other choice is 'plane' for plane wave excitation

params_simu.J_src_x = 1.0 + 0.j

if 'dipole' excitation, gives the amplitude and phase of the J dipole in the x-direction. If 'plane' excitation, gives the amplitude and phase of the E-field of the plane wave in the x-direction at a given reference point (usually the origin)

params_simu.J_src_y = 0.0 + 0.j

same as for params_simu.J_src_x but in the y direction

params_simu.J_src_z = 0.0 + 0.j

same as for params_simu.J_src_x but in the z direction

params_simu.r_src_x = 0.1

gives the x-origin of the J dipole or of the plane wave

params_simu.r_src_y = 0.1

gives the y-origin of the J dipole or of the plane wave

params_simu.r_src_z = 20.0

gives the z-origin of the J dipole or of the plane wave

These parameters are for advanced users to change. Most users will be happy with the default values.

params_simu.COMPUTE_Z_NEAR = 1

1 if we need to compute the near field and SAI preconditioner sparse matrices, 0 otherwise

params_simu.MOM_FULL_PRECISION = 1

1 for full precision in the MoM (more integration points on the triangles), 0 otherwise

params_simu.lc_factor = 1.0/10.0

the characteristic length factor. It will multiply lambda (the wavelength) to obtain the average edge length (lc) of the mesh. In Method of Moments, usually lc ~= lambda/10

params_simu.a_factor = 0.25

the a (finest cubes sidelength) factor -- it will multiply lambda (the wavelength) to obtain the side length of the leaf (finest) level. Usually a = lambda/4

params_simu.nu = 0.2

Combined Field Integral Equation factor. Changing this factor may result in fewer, or more likely, more iterations in solving the linear system. Value has to be between 0 and 1

params_simu.BE_BH_N_Gauss_points = 1

the number of points for the leaf cubes radiation functions calculation through Gaussian integration. Values can be 1, 3, 6, 9, 12, 13

params_simu.DIRECTIONS_PARALLELIZATION = 1

this option should be enabled if there is a big number of processors wrt the number of cubes at the coarsest level. To disable, enter 0

params_simu.ALLOW_CEILING_LEVEL = 0

do we allow a ceiling level that could be other than the third coarsest level? If yes: 1, if no, 0. the The ceiling level having the smallest memory footprint by its cubes is chosen

params_simu.NB_DIGITS = 3

number of significant digits carried out during the whole computation

params_simu.TOL = 1.e-3

iterative solver tolerance

params_simu.SOLVER = SOLVERS[0]

here we choose the solver. Since SOLVERS = ["BICGSTAB", "GMRES", "RGMRES", "FGMRES"], 0 yields "BICGSTAB", 1 yields "GMRES", 2 yields "RGMRES" and 3 yields "FGMRES"

params_simu.MAXITER = 150

the maximum number of iterations of the main solver. Not to be confused with the maximum number of matrix-vector multiplications!

params_simu.RESTART = 30

the restart value for GMRES-type solvers

params_simu.INNER_SOLVER = SOLVERS[0]

the choice of the inner solver, in an inner-outer solution scheme. Since SOLVERS = ["BICGSTAB", "GMRES", "RGMRES", "FGMRES"], 0 yields "BICGSTAB", 1 yields "GMRES", 2 yields "RGMRES" and 3 yields "FGMRES"

params_simu.INNER_TOL = 0.25

inner solver tolerance, which does not need to be as strong as the outer solver tolerance

params_simu.INNER_MAXITER = 15

inner solver maximum number of iterations

params_simu.INNER_RESTART = 30

restart parameter for inner solver

params_simu.PRECOND = "FROB"

the preconditioner type, defaulted to a Sparse Approximate Inverse using a Frobenius norm criterion. Not much of a choice here!

params_simu.USE_PREVIOUS_SOLUTION = 1

1 if we want to use result from previous angle in monostatic RCS computation, 0 otherwise

params_simu.MONOSTATIC_BY_BISTATIC_APPROX = 0

1 if we want to use the bistatic approximation for computing monostatic RCS, 0 otherwise

params_simu.MAXIMUM_DELTA_PHASE = 0.0

in case we want to use the bistatic approximation for computing monostatic RCS, this parameter allows for choosing the angle between samples

params_simu.MAX_BLOCK_SIZE = 10.

max block size for the near field and preconditioner matrices (in MBytes): the near field and preconditioner matrices are sliced in blocks and dumped to the disk in order to minimize RAM memory occupation

params_simu.int_method_theta = "GAUSSL"

Gauss-Legendre integration following theta for the MLFMA radiation function

params_simu.int_method_phi = "PONCELET"

Poncelet (mid-point method) integration following phi for the MLFMA radiation function

params_simu.INCLUDE_BOUNDARIES = 0

1 if we include the boundaries of the radiation functions, 0 otherwise

params_simu.PERIODIC_Theta = 1

PERIODIC_Theta = 1 if we use Gaussian interpolator, 0 if we want to use the Lagrangian interpolator

params_simu.PERIODIC_Phi = 1

PERIODIC_Phi = 1 if we use Gaussian interpolator, 0 if we want to use the Lagrangian interpolator

params_simu.CYCLIC_Theta = 1

CYCLIC_Theta = 1 if we use cyclic interpolation for the radiation functions, 0 otherwise

params_simu.CYCLIC_Phi = 1

CYCLIC_Phi = 1 if we use cyclic interpolation for the radiation functions, 0 otherwise

params_simu.CURRENTS_VISUALIZATION = 0

CURRENTS_VISUALIZATION tells if we want to create a view of the currents in GMSH

params_simu.SAVE = True

save the output in a file

params_simu.fileToSaveTo = 'result.txt'

the name of the file which will store the results

params_simu.VERBOSE = 1

gives little to no informational output about program state if 0. Set to 1 for debugging infos