Bayesian methods in probability of detection estimation and model-assisted probability of detection evaluation

Aldrin, John C.; Knopp, Jeremy S.; Sabbagh, Harold A.

doi:10.1063/1.4789250

Cited by 14 publications

(12 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…From the literature, there appear to be several ways in which human effects may be incorporated into POD models for MAPOD applications: 1) represent human effects as influencing parameters in the NDE response model, for instance, using a multi-parameter POD model (Aldrin et al 2012;Pavlović et al 2012), and/or 2) represent human effects explicitly as a random effects term in the POD model (Li et al 2012), and/or 3) account for human effects through the definition of the threshold condition for observation of defects. In these scenarios, the challenge is to define the relationship between NDE response and the variable representing human effects and/or to quantify the variability contributed to the response by human effects.…”

Section: Discussion and Analysis Of Human Factors And Integration Witmentioning

confidence: 99%

“…Explicit representation of POD as a function of multiple influencing parameters has been explored in works by Pavlović et al (2012) and Aldrin et al (2012). Instead of simply representing POD as a function of a single defect size parameter, it is suggested that POD can also be represented as a function of other relevant parameters such as flaw orientation, surface roughness, etc.…”

Section: Pod Model For Continuous Nde Response (â Vs a Models)mentioning

confidence: 99%

“…The results from measurements on artificial flaws are considered the prior information that is updated based on limited real flaw data, through the likelihood function. Aldrin et al (2012) introduce a Bayesian method for estimating the POD model parameters for a high-frequency eddy current application based on â vs. a responses. A Bayesian method is used to estimate the model parameters for different types of â vs. a models including linear single parameter, linear multi-parameter, and physics-based models using data generated by computer simulation.…”

Section: Bayesian Approachesmentioning

confidence: 99%

“…Elsewhere, it is defined through its motivation for (Aldrin et al 2012), "…reducing POD sample number by better managing experimental variation using more accurate statistical models and appropriately designing experiments based on prior experience and data. "…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Review of Literature for Model Assisted Probability of Detection

Meyer

Crawford

Lareau

et al. 2014

View full text Add to dashboard Cite

This report documents a literature survey of so-called "Model Assisted Probability of Detection" (MAPOD) approaches that may be useful in determining the probability of detection (POD), a metric for quantifying the performance of nondestructive evaluation (NDE) methods. The objective of this report is to summarize MAPOD concepts that have been proposed to date in order to assess specific approaches that may be appropriate for application to improve estimates of POD for field NDE of nuclear power plant components. The limitations of laboratory-based studies to replicate actual field conditions are well-recognized and not limited to the nuclear power industry. Probability of detection estimates based on laboratory studies generally provide ideal environments for performing examinations as compared field settings, which typically will result in non-conservative estimates of POD for field applications. As a consequence of this effort, the authors conclude that use of MAPOD concepts to improve estimates of field NDE performance may require access to certain field data, or alternatively, may require significant laboratory studies to assess the influence of human and environmental shaping factors. If the necessary field data cannot be made available, it is proposed that laboratory efforts will have greater chance of success by focusing on a specific examination technique and component application. The authors also find that MAPOD concepts may have a more immediate contribution to the nuclear power industry through their use in extending personnel and procedure qualifications beyond established limits. Finally, another basis was identified that may be used to adjust POD curves generated in previous reliability studies and performance demonstrations. Many of the previous studies calculated POD as the average of performance data which may be inappropriate for many applications, as the calculation may be nonconservative. A more conservative approach to estimating POD would include basing the calculation on a lower statistical quantile of the data set. v

show abstract

Section: Discussion and Analysis Of Human Factors And Integration Witmentioning

confidence: 99%

Section: Pod Model For Continuous Nde Response (â Vs a Models)mentioning

confidence: 99%

Section: Bayesian Approachesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Review of Literature for Model Assisted Probability of Detection

Meyer

Crawford

Lareau

et al. 2014

View full text Add to dashboard Cite

show abstract

“…The POD is typically determined through experiments that are both time-consuming and costly. This motivated the MAPOD methods with the aim for reducing the number of experimental sample points by introducing insights from controlled experiments using information from physics-based simulations [11,12]. However, when it comes to a large amount of simulations, especially when containing statistical uncertainty in random inputs and exploring the joint distributed statistical moments (Fig.…”

Section: -2mentioning

confidence: 99%

Model-assisted probability of detection of flaws in aluminum blocks using polynomial chaos expansions

Leifsson

Grandin²,

Meeker³

et al. 2018

AIP Conference Proceedings

View full text Add to dashboard Cite

Probability of detection (POD) is widely used for measuring reliability of nondestructive testing (NDT) systems. Typically, POD is determined experimentally, while it can be enhanced by utilizing physics-based computational models in combination withmodel-assisted POD (MAPOD) methods. With the development of advanced physics-basedmethods, such as ultrasonic NDTtesting, the empirical information,needed for POD methods, can bereduced. However, performing accurate numerical simulationscan be prohibitivelytimeconsuming, especially as part of stochastic analysis. In this work, stochastic surrogate models for computational physics-based measurement simulations are developed for cost savings of MAPOD methods while simultaneously ensuring sufficient accuracy. The stochastic surrogate is used to propagate the random input variables through thephysics-basedsimulation model to obtain the joint probability distribution of the output. The POD curves are then generated based on those results. Here, the stochastic surrogates are constructed using nonintrusive polynomial chaos (NIPC) expansions. In particular, the NIPC methods used are the quadrature, ordinary leastsquares (OLS), and least-angle regression sparse (LARS) techniques. The proposed approach is demonstrated on the ultrasonic testing simulation of a flat bottom hole flaw inanaluminum block. The results show that the stochastic surrogates have at least two orders of magnitude faster convergence on the statistics than direct Monte Carlo sampling (MCS). Moreover, the evaluation of the stochastic surrogate models is over three orders of magnitude faster than the underlying simulation modelfor this case,which is the UTSim2 model.

show abstract