Probabilities.FatigueLifePDF Method

Overload List

#SignatureDescription
1procedure FatigueLifePDF(const X: TDenseMtxVec; Mu: Double; gamma: Double; beta: Double; const Res: TDenseMtxVec);Fatigue Life PDF (vectorized).
2function FatigueLifePDF(x: Double; Mu: Double; gamma: Double; beta: Double): Double;Fatigue life (Birnbaum-Saunders) probability density function (PDF).

Overload 1: procedure FatigueLifePDF(const X: TDenseMtxVec; Mu: Double; gamma: Double; beta: Double; const Res: TDenseMtxVec);

Fatigue Life PDF (vectorized).

#NameDescription
1XDefines distribution domain, real vector or matrix with real values on open interval (Mu,INF).
2MuDefines distribution location parameter.
3gammaDefines distribution shape parameter. gamma must be a positive scalar.
4betaDefines distribution scale parameter. beta must be a positive scalar.
5ResAfter calculation stores the PDF calculated from X, Mu, gamma and beta. Length and Complex properties of Res are adjusted automatically to match Length and Complex properties of X.

Result: stored in self (calling object)

Overload 2: function FatigueLifePDF(x: Double; Mu: Double; gamma: Double; beta: Double): Double;

Fatigue life (Birnbaum-Saunders) probability density function (PDF).

#NameDescription
1xDistribution domain, real value on the open interval (Mu, +Inf).
2MuDistribution location parameter, real value.
3gammaDistribution shape parameter. gamma must be a positive scalar (gamma>0).
4betaDistribution scale parameter. beta must be a positive scalar (beta>0).

Returns: Double - the Fatigue life probability density function (PDF) evaluated at x using the location parameter Mu and the positive shape (gamma) and scale (beta) parameters. Returns NaN if gamma<=0, beta<=0, or x<=Mu.

Remarks:

The fatigue life distribution is also known as the Birnbaum-Saunders distribution. Several alternative formulations appear in the literature. With z=(x-mu)/beta and phi the standard normal PDF, the textbook Birnbaum-Saunders density is

PDF(x | mu,gamma,beta) = ((sqrt(z)+sqrt(1/z))/(2gamma(x-mu))) phi((sqrt(z)-sqrt(1/z))/gamma) , z=(x-mu)/beta, x>mu .

NOTE (implementation): the current source evaluates the standard-normal-PDF argument as bigl(z^2-sqrt(1/z)bigr)/gamma instead of bigl(sqrt(z)-sqrt(1/z)bigr)/gamma, so the returned PDF is not the exact derivative of FatigueLifeCDF away from x=mu+beta. Returns NaN for x<=Mu, gamma<=0 or beta<=0.

See Also: Probabilities.FatigueLifeCDF, Probabilities.FatigueLifeCDFInv