SPARK::DefaultGlobalSettings Namespace Reference

Defines the default values for the global settings. More...

Tolerance settings
const double	Tolerance = 1.0e-6
	Relative tolerance used throughout the simulation.
const double	MaxTolerance = 1.0e-3
	Looser relative tolerance used to recover if convergence cannot be achieved against Tolerance.
Safety factors
Safety factors are used to modify the convergence check depending on the context (i.e., check after prediction step or after an iteration) and on the type of the unknwon variable (i.e., a normal unknwon or a break unknown). A safety factor smaller than 1 makes the convergence check tougher to satisfy, with 0 making it impossible. A safety factory larger than 1 essentially corresponds to a relaxation of the convergence check. E.g, if IterationSafetyFactor=10 then the convergence check at each iteration after the prediction will be made 10 times easier to satsify than the specified tolerence, because the convergence error must be samller than < SafetyFactor*Tolerance. The default safety factors are equal to 1. therefore, they treat break and normal unknowns with the same weight in the convergence check. For certain applications, it might be useful to only perform a convergence check on the break variables. This is achieved by setting the NormalUnknownSafetyFactor entry to a number larger than 1.
const double	PredictionSafetyFactor = 0.01
	Safety factor used to check convergence after prediction [feature added in SPARK 2.0].
const double	IterationSafetyFactor = 0.9
	Safety factor used to check convergence after each iteration [feature added in SPARK 2.0].
const double	BreakUnknownSafetyFactor = 1.0
	Safety factor used to check convergence for a break unknown [feature added in SPARK 2.0].
const double	NormalUnknownSafetyFactor = 1.0
	Safety factor used to check convergence for a normal unknown (i.e., not a break variable) [feature added in SPARK 2.0].

---

Detailed Description

Defines the default values for the global settings.

---

Variable Documentation

const double SPARK::DefaultGlobalSettings::Tolerance = 1.0e-6

Relative tolerance used throughout the simulation.

const double SPARK::DefaultGlobalSettings::MaxTolerance = 1.0e-3

Looser relative tolerance used to recover if convergence cannot be achieved against Tolerance.

const double SPARK::DefaultGlobalSettings::PredictionSafetyFactor = 0.01

Safety factor used to check convergence after prediction [feature added in SPARK 2.0].

const double SPARK::DefaultGlobalSettings::IterationSafetyFactor = 0.9

Safety factor used to check convergence after each iteration [feature added in SPARK 2.0].

const double SPARK::DefaultGlobalSettings::BreakUnknownSafetyFactor = 1.0

Safety factor used to check convergence for a break unknown [feature added in SPARK 2.0].

const double SPARK::DefaultGlobalSettings::NormalUnknownSafetyFactor = 1.0

Safety factor used to check convergence for a normal unknown (i.e., not a break variable) [feature added in SPARK 2.0].

---