crackR (version 0.3-9)

lognormalPOD: Define Probability Of Detection (POD) curve for Non-Destructive Inspection (NDI) based on a Log-Normal CDF formulation

Description

The POD curve in probabilistic damage tolerance analysis is often defined using the formulation of a Log-Normal CDF, with several optional modifying parameters. This function can be used to generate the appropriate POD function for use in the crackRparameters component "pod.func".

Usage

lognormalPOD(a,median,slope,a.min.detectable=0,poi=1,far=0)

Arguments

a

Crack length; the sole input to the POD function generated by this function.

median

Median detectable crack length.

slope

Slope parameter (aka sdlog).

a.min.detectable

Crack size below which detection is assumed to be impossible. Below this size the generated POD curve returns a zero probability. Note the far parameter will override this if far > 0.

poi

Probability Of Inspection. In practice an analyst may assume that some percentage of scheduled inspections will not occur as planned or will be conducted incorrectly. This can be represented by setting poi less than 1, in which case all detection probabilities are factored down.

far

Often in Non-Destructive Inspection (NDI) it is assumed there is no false alarm rate. This is most likely untrue. Setting a false alarm rate in this function forces a minimum value of POD that will be returned from the generated function.

References

MIL-HDBK-1823A (Department of Defense, USA) "Nondestructive Evaluation System Reliability Assessment", Apr 2009 Halbert, K. "Estimation of Probability of Failure for Damage-Tolerant Aerospace Structures" PhD Thesis, Temple University Department of Statistics, Philadelphia, PA, Apr 2014

See Also

crackRinit inspection

Examples

myPODcurve <- function(a) lognormalPOD(a, median=0.01, slope=0.5, a.min.detectable =
0, poi = 0.95, far = 0.001)
myPODcurve(c(0, 0.005, 0.01, 0.05, 1))