Summary for: SimulationParameters

Class summary

SimulationParameters Parameters for MagneticsProblem solutions.

This class is intended for supplying parameters to the MagneticsProblem class.

** NOTE Unlike most EMDtool classes, this is NOT a handle class, meaning that changes to any of the class properties, made after giving an instance of the class as an argument to a MagneticsProblem solve method will not be propagated to the problem object.

Properties

.N_periods Number of electrical periods to simulate.

.N_stepsPerPeriod Number of time-steps per electrical period

Number of time-steps per electrical period 1/this.f.

.alpha2 Time-stepping scheme parameter.

A parameter for specifying the time-stepping scheme, with the following options:

  • 2 : Backward (or implicit) Euler. The most robust options, but can sometimes require a short step length for reaching a given accuracy.
  • 1 : Pure trapezoidal rule (or Crank-Nicolson according to some authors. Can result in non-decaying spurious oscillations.
  • 1 < alpha2 < 2 : A hybrid approach. Often the best of both words.

See also SimulationParameters.use_hybrid_stepping

.convergence_tolerance Newton convergence tolerance.

Newton iterations are performed until the relative residual, i.e. the norm of the Newton residual divided by max(norm(load_vector), 1) is strictly smaller than the convergence tolerance.

.f Supply frequency.

Supply electrical frequency. For non-typical problems with no clearly defined supply frequency, the frequency can be understood purely via the time-step length, through this.N_stepsPerPeriod

.maxIter - maximum number of Newton iterations

.maximum_damping_iterations Maximum number of damping iterations.

Number of damping iterations to perform for each Newton step (after the aforedescribed number of regular steps). On each damping step, the step size is halved, until the residual norm strictly decreases or the maximum number is reached.

.maximum_regular_iterations Number of maximum regular iterations.

The maximum number of pure-Newton steps to perform, with the full step size. The count restarts for each time-step, obviously.

.rel - relaxation factor for iterations, default = 1 (no relaxation)

.rotorAngle Rotor angle.

Used directly in static analysis, in which an array of angles can be specified. In harmonic or stepping analysis, a scalar value must be given, corresponding to constant bias in the rotor angle.

.silent - display progress in command line, true/false

.slip Rotor slip.

Used in harmonic and time-stepping analysis.

.ts - time-step time-stamps

.use_cached_Jacobian Use saved Jacobian constructor.

If true, repeated calls to the solve methods of MagneticsProblem will use the same MagneticsJacobian object. Defaults to true.

.use_hybrid_stepping Use a hybrid time-stepping scheme.

Use a hybrid time-stepping scheme in which the implicit Euler method is used for those nodal potentials (pure FE-part of the model, nodal values of magnetic vector potentials) that are fully non-damped aka do not appear in the FE part of the problem mass/damping matrix. For the remaining nodes, and all circuit variables, the scheme defined by pars.alpha is used.

This benefit can eliminate or limit spurious oscillations sometimes (rarely) occurring with hybrid trapezoidal-IE stepping schemes (with pars.alpha2 strictly less than 2).

Defaults to true.

Methods

Class methods are listed below. Inherited methods are not included.

.default parameters

Documentation for SimulationParameters/SimulationParameters doc SimulationParameters

.SimulationParameters/dispf is a function.

dispf(this, varargin)

.SimulationParameters/miscpar is a function.

p = miscpar(this, name)