Sample application of sensitivity/uncertainty analysis techniques to a groundwater transport problem. National Low-Level Waste Management Program

Abstract

The primary objective of this document is to provide sample applications o f selected sensitivity and uncertainty analysis techniques within the context of the radiological performance assessment process. These appl ications were drawn from the companion document Guide7 ines f o r S e n s i t i v i t y and Uncertainty AnaTIyses o f Low-Leve7 Radioactive Waste Performance Assessment Computer Codes (S. Maheras and M, Kotecki , DOE/LLW-100, 1990). Three techniques are illustrated in this document: one-factor-at-a-time (OFAT) analysis, fractional factorial design, and Latin hypercube sampling. sensitivity and uncertainty analysis at the early and latter stages o f the performance assessment process, and potentia'! pitfalls that can be encountered when applying the techniques, as opposed to the actual results, since the results are hypothetical and are not based on site-specific conditions. The report also illustrates the differences in The emphasis is on application of the techniques i i i EXECUTIVE SUMMARY Sensitivity and uncertainty analyses are a valuable, but sometimes neglected, part of the performance assessment (PA) process. In some cases, P A 4 s seen only as a method for demonstrating compliance with long-term dose or risk objectives after a facility has been designed. Furthermore, some view PA as a requirement that must be done, without consideration of the other benefits that can be obtained. used as a management tool during all stages of the licensing process. By indicating the relative impact o f individual design and site Characterization parameters on disposal system performance, sensitivity/uncertainty analysis results can provide justificatior for specific activities based on quantitative, performance-based measures. Of course, technical judgement and qualitative factors (i.e., reality checks) must also be considered because o f the uncertainties involved and complexity of the processes. understanding of the behavior of a system makes it easier for management to defend decisions related to continuing or discontinuing funding for specific activities. PA (through the use of sensitivity/uncertainty analysis) can be In any case, improved The use o f sensitivityluncertainty analyses in the PA process has two primary benefits: e e Cost savings obtained by focusing site characterization, design, and modeling efforts on the site data, design features, and transport pathways that have the greatest impacts on facility performance Quantification o f the uncertainty associated with PA predictions used to demonstrate compl iance, Furthermore, the U.S. Nuclear Regulatory Commission (NRC) has emphasized the need for sensitivity/uncertainty analyses by including a requirement in the Standard Format and Content f o r a License A p p l i c a t i o n for a Low-Levell R a d i o a c t i v e Waste Disposall F a c i l i t y (NUREG-1199, Rev. 1; NRC 1988, pp. 6-7 and 6-10). V The primary objective of this document is to provide sample applications of selected sensitivity and uncertainty analysis techniques within the contex...

GWSCREEN: A semi-analytical model for assessment of the groundwater pathway from surface or buried contamination: Theory and user`s manual

“…The PORFLO peak flux exceeds the GWSCREEN peak flux. A similar observation of contaminantconcentrationwas made by Seitz et al (1991) lt was shown that for Peclet numbers less than about 50, the peak concentration,and consequently,the peak mass flux, increased when the ratio of contaminant half-life to contaminant transit time was decreased to less than 0.1 (Figure 12). The Peclet number for this simulationwas Pe = 9.Im+ 0.5 m = 18 One method to assure conservatismis to make sure the half-life/transittime ratio be limited to a minimum value of 0.1.…”

“…The cumulative flux for both models is the same as time approaches infinity. For contaminantswith short half-lives 2, under conditionswhere the dispersivity is large, the peak flux may actually be larger using a dispersion model as compared to a plug flow model (Seitz et al, 1991) because the contaminant arrives at the receptor sooner resulting in reduced decay time. Dispersivity in the unsaturated zone would not be readily available during the initial groundwater assessmentprocess and since other conservativeassumptions, such as increasingthe percolation rate and reducing the unsaturated zone thickness, may be assumed to compensatefor the effect of dispersion on short half-lived contaminants, dispersionwas not consideredand the simpler plug-flow model IThe contaminant half-life is greater than 1/10 the contaminant transit time in the unsaturated zone was used.…”

“…PORFLO-3and GWSCREENmass flux at the unsaturated zone-aquifer interface as a functionof time for a contaminant with an infinitehalf-3and GWSCREENmass flux at the unsaturated zone aquifer interface as a functionof time for a contaminant with a four year half-life, Normalized peak concentrationvs Peclet number for contaminants with different half-life/contaminant transit time (Ht) ratios. (Taken fromSeitz et al, 1991) Six sample problems are illustrated in this Appendix. Each sample problem was• intended to demonstrate a different aspect of the GWSCREEN code.…”

GWSCREEN: A semi-analytical model for assessment of the groundwater pathway from surface or buried contamination. Theory and user`s manual, Version 2.0: Revision 2

“…For contaminants with short half-lives relative to their residence time in the unsaturated zone, the plug flow assumption may not be conservative since the contaminant will arrive sooner (lower residence time in the unsaturated zone) when dispersion is considered, therefore allowing less time to decay. This observation was noted by Seitz et al (1991) and may become a significant factor when the unsaturated transit time [see Equation 21] is greater than 10 times the half-life of the contaminant. Dispersion coefficients in the unsaturated zone would not be readily available during the initial groundwater assessment process.…”

Modeling approaches for concrete barriers used in low-level waste disposal

“…Sensitlvlty/uncer-prepared as part of the Concrete Barriers Project. tainty analysis has been recognized as a critical New information about approaches not addressed part of a PA (Campbell et al 1993; in the previous documents and papers is also 1993; Seitz et al 1991;Seitz et al 1992a).…”