< ,

Modeling of Human Intrusion Scenarios at the Waste Isolation Pilot Plant

M.K. Knowlesl, F.D. Hansenl, K.W. Larsonl, T.W. Thompson2, and M.B. Gross2

1Sandia National Laboratories, Carlsbad, NM, USA 2Carlsbad Technical Assistance Contractor, Carlsbad, NM, USA

ABSTRACT QSTS The Waste Isolation Pilot Plant is a mined, geologic repository designed for permanent disposal of transuranic waste. The facility is owned by the United States Department of Energy, and licensed for operations by the Environmental Protection Agency. Compliance with license requirements dictates that the repository must comply with regulatory stipulations that performance assessment calculations include the effects of resource exploitation on probable releases. Scenarios for these releases incorporate inadvertent penetration of the repository by an exploratory drilling operation. This paper presents the scenarios and models used to predict releases from the repository to the biosphere during. an inadvertent intrusion into the waste disposal regions. A summary of model results and conclusions is also presented.

Key WorrLF: Transuranic Waste /Modeling Scenarios / WIPP

INTRODUCTION

The Waste Isolation Pilot (WIPP) is a mined, geologic repository designed for permanent disposal of transuranic waste. The facility is located in southeastern New Mexico, USA, and was developed by the U.S. Department of Energy. The repository horizon is located approximately 650 m below the surface in a sequence of bedded salt deposits known as the Salado Formation. A sedimentary sequence comprised primarily of sandstones, siltstones, and carbonates overlays the Salado. Site characterization studies of the overlying formations indicate the presence of only limited groundwater. The inherently low porosity and low permeability nature of the Salado Formation isolates the waste from other groundwater sources and essentially eliminates the possibility of transport of contaminants from the repository horizon along groundwater pathways. The WIPP site selection was based on these positive attributes as well as on the ability of salt to creep around the waste, thereby forming a virtually impermeable barrier between the disposal regions and the biosphere.

The WIPP was licensed for operations in the spring of 1998. The licensing process required that the Department of Energy demonstrate that the WIPP complies with U.S. Environmental Protection Agency regulations. Compliance with the applicable regulations is demonstrated by comparison of a summation of probable releases over the 10,000-year regulatory period with an allowable standard. Release pathways include groundwater releases associated with advective and diffusive contaminant migration from the waste disposal areas and direct releases that occur as a consequence of natural resource exploitation (Le., mining or drilling). Regulations state that repository performance assessment calculations must include effects of inadvertent and intermittent intrusion into the disposal areas by drilling for resources. Because the geology of the WIPP site essentially eliminates groundwater releases from the facility, an inadvertent drilling intrusion into the repository area represents the most probable release scenario. In addition to regulatory oversight, the Department of Energy encourages review by stakeholder entities concerned with issues of public health, safety, and economics. Procurement of government and public approval of the first licensed permanent

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, make any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

DISCLAIMER

Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

f

disposal site for radioactive waste presented unique challenges to the project. Many of these challenges were related to perceptions regarding the effects of petroleum industry activities and the associated direct releases on repository performance.

Petroleum exploration and production activities are common in the vicinity of the WIPP; extrapolating the current drilling rate into the future results in hypothetical penetration of the repository several times during the regulatory period. For each probable intrusion, a quantitative value for release of contaminated material is calculated. Evaluation of repository performance depends on the volume of material released and on the radionuclide loading, or activity, of the contaminated material. The activity is estimated from that of the expected waste inventory. Total releases are calculated through incorporation of stochastic uncertainty in a series of performance assessment calculations included in the WIPP compliance certification application(1 1. Uncertainty is captured through variation of conceptual models of the repository and variation of model parameters. A detailed discussion of the probabilistic approach utilized for WIPP performance assessment can be found in the compliance application(l). The performance assessment system includes conceptual models of the direct release processes associated with petroleum exploration; these models are presented in this paper. A description of direct release processes is followed by discussion of models used to evaluate the consequences associated with direct releases. A retrospective review of the conceptual model assumptions and consequences concludes this paper.

Direct Releases at the WIPP Evaluation of the events associated with petroleum exploration led to the formulation of a set

of four mechanisms whereby contaminated material could be transported from a penetrated waste panel to the surface during a drilling operation. These mechanisms are as follows: 1. 2. 3.

4.

cuttings, which include material intersected by the rotary drilling bit; cavings, which include material eroded from the borehole wall during drilling; spallings, which include solid material carried into the borehole during rapid depressurization of the waste- disposal region; and direct brine releases, which include contaminated brine that may flow to the surface during drilling.

Understanding of the WIPP direct release scenarios requires a perspective on the repository and disposed waste. Disposal rooms are mined from the salt in a series of panels, as depicted in Figure 1. Each panel consists of ten rooms, which will be filled with containers of waste and magnesium oxide (MgO) backfill material. The backfill material acts as a chemical buffer in the waste disposal regions. Table 1 provides a summary of the initial waste inventory.

The original waste, consisting of a diversity of materials, will evolve as a consequence of host rock response and chemical reactions within the disposal areas. Creep of the incumbent salt formation into the excavated openings will dominate the system response during the first few decades following waste disposal. Within 100 years, WIPP rooms will close to approximately one half of their original height, thereby compacting the disposed waste and backfill material(2). Although the incumbent salt formation has relatively little water, field evidence suggests that some brine will flow into waste areas following disposal. This moisture could, in turn, cause corrosion of steel drums and iron-based materials within the waste containers. Current conceptual models consider microbial degradation of cellulosics, plastics, and rubbers present in the waste stream to be probable. These degradation processes lead to alteration of the waste form and to possible generation of hydrogen, carbon dioxide, and methane gas in the repository. Calculations predict pressures could rise from atmospheric at the time of waste disposal to approximately lithostatic pressure (1 4.8 MPa) within several hundred years following repository cIosure(1).

W.ICSb

Figure 1. Schematic of WIPP underground layout. Disposal areas are mined from the Salado Formation, approximately 650 m below ground surface.

Table 1. Initial Waste Concentrations

Figure 2 presents a schematic of waste evolution over the regulatory period. This schematic is based on several assumptions regarding future processes. The volume of brine in a disposal room is assumed sufficient to generate steel corrosion. Possible sources of brine to the repository include seepage from the Salado Formation, upward flow from a punctured brine reservoir below the repository (Figure 3), and downward flow of brackish groundwater via a plugged exploratory

I

borehole (Figure 4). It is further assumed that the backfill material (MgO) will function only as a sink for carbon dioxide gas generated as a consequence of microbial degradation; it is otherwise assumed inert. It is assumed that creep closure of salt into disposal rooms will cause container rupture, thus exposing waste directly to fluids present in the repository. These assumptions, as well as those discussed in the next section, result in a bounding model of waste evolution over the regulatory period. Implementation of bounding, or conservative, assumptions are consistent with the global strategy to take a conservative approach in the modeling of WIPP performance.

Figure 2. Artist’s schematic of WIPP waste evolution over time.

Modeling Strategy

The modeling strategy applied to human intrusion scenarios at the WIPP incorporates conceptual models of the waste, the drilling process, and the incumbent host rock. A review of scenario development and implementation demonstrates that calculations for the four direct release mechanisms rely on four assumptions: 1. waste rooms and panels are hydraulically connected, 2. waste properties are analogous to a weakly consolidated, particulate material, 3. transient wellbore processes can be neglected, and 4. drilling of exploratory wells in the vicinity of WIPP will continue at present-day rates and current

technology.

0 x - w

a - .- e U e

El

Boundaryof I Upper Seal System-

Laver Seal Systern-

1 MB139 Access Drifts

I I I (Not to Scale) Pressurized - I Brine

Note: I3orcholc pcnctrab wzstt and prcssurized brine in the underlying Castile Formation. h o w s indicate hypothetical dimtion of goundwt%er f l ~ w and radionuclide transport.

Repository and shafts Groundwater flow and I-)LI radionuclide transport . . . . Anhydrite layers a and b

0 Cutebra Disturbed rock zone Increase in Culebra hydraulic conductivity due to mining

ccc31c-2

Figure 3. Schematic of a human intrusion scenario resulting in penetration of the waste regions and a pressurized brine pocket by an exploratory borehole. Brine from the pocket could migrate upward into the waste disposal regions.

The logical model of the repository (Figure 5, not drawn to scale) captures the first assumption. This model represents an abstraction of the physical domain into a two-dimensional region. Although the model does not retain details of stratigraphy and geometry, the volume and properties of each region are captured. Note that the single panel (Component 23) intersected by an intrusion borehole is modeled as a homogeneous segment; the remaining panels are lumped together

(Component 24), with properties uniformly distributed. Although a more detailed mesh was created for direct brine release(31, the assumption of hydraulic connectivity of waste areas was maintained

E2

I

l a, I (Not to Smle)

d - .-

Pressurized e- Brine

- u)

Note: Borchole pznetrates waste and Qes not penztrate pressurized brine in the underlying Castile Formation. Arrows indiate hiypckhetical direction of groundwater flow and radionuclide tramit.

Anhydrite layers a and b Groundwater and Repository and shafis radionuclide transport 0 Cuiebra Disturbed rock zone Increase in Culebra

hydraulic conductivity due to mining

ccA-011-i

within the model depicted in Figure 5 .

Figure 4. Schematic of a human intrusion scenario in which an exploratory borehole intersects a waste disposal panel. Groundwater from the Culebra could migrate downward to the waste disposal regions following plugging of the borehole.

Material properties assigned to the waste, as shown in Table 2, imply complete degradation of waste containers and contents. As discussed in Hansen et a1.(4), these products include.ferrous and

ferric oxides mixed with salt fragments, reacted and unreacted magnesium oxide, and undegraded materials. Porosity of waste regions is calculated using a synthesized model of room closure resulting from salt creep and resistance to closure caused by pressure created by gas generation. Rooms that are originally about 4 m in height will reduce to between 1 and 2 m in height at the time of a hypothetical intrusion.

Transient wellbore processes include actions taken by the driller to control a blowout. Current drilling practices within the WIPP vicinity utilize common well control technology, such as blow-out preventers. However, since the repository horizon lies above those formations expected to produce high-pressure gas pockets, the conservative assumption of no well control was applied. Other transient processes include ejection of drilling mud, transport of cuttings, and dynamic processes associated with a spallings scenario. Spallings will be discussed in more detail later in the paper.

Table 2. Material Properties Assigned to WIPP Waste

Property WIPP Value Silty Sand Value Sandstone Value Hydraulic Conductivitv (cm/s) 10-3 10-1 - 10-3 10-3 - 10-5

Specific gravity 2.65 2.5 - 2.6* 2.5 - 2.6 Unconfined Compressive Strength &Pa) 6.85 Not Applicable 6.8 - 137,000 Porosity 0.2 - 0.85 0.25 - 0.75 0.05 - 0.3

* Iron oxide specific gravity is 5.4; salt specific gravity is 2.1.

Present-day drilling rates in the WIPP vicinity result in approximately four wells drilled per acre within the Delaware Basin, which includes the WIPP region. Activity in the Delaware Basin has increased substantially over the past ten years. Regulations dictate a drilling rate that results in approximately sixteen wells drilled per acre over the 10,000 years simulated. Wells are drilled using saturated brine as a drilling mud, with penetrations rates varying from about 15 to 60 m per hour. Although these assumptions can be considered conservative for the WIPP, they also preclude speculation regarding future drilling objectives and technology. The geometry (Figure 6 ) of a hypothetical drilling intrusion is based on current practices.

These assumptions lead to the initial and boundary conditions for the direct release models. Cutting releases are calculated from the cylinder defined by wellbore diameter (Figure 6 ) and room height at the time of the intrusion. Cavings releases require calculations of the erosional shear stresses of the flowing drill mud, as defined by the fluid viscosity, density, and helical velocity. Erosion of waste material occurs if the fluid shear stress exceeds a threshold value known as the critical shear stress. The absence of data on critical shear stress for disposed waste necessitated use of literature values for this parameter. A minimum value of 0.05 Pa can be assigned based on field and laboratory experiments on fine-grained, cohesive sediments. The maximum value of 80 Pa was assumed correlated to a waste particle size diameter using the Shield’s diagram(5). This diagram, derived from experimental data, presents critical shear stress for noncohesive particles as a function of grain size.

1. BOrehole (first 2DEyearsJ 8. &per Sllado corrpatted day co&lmn 9. bwer Sdado corrpacted day column IA. Borehole concrete plug

1B. Upper unrestricted borehole 1 E. bwr cby component IC. Lower unrestricted borehole 11. Concrete monolith

2. Shaff 12 Unlts abme the Dewy W e 3. Earth fill 13. Dew bls 4. Rustler compacted clay cbturfn 14. Forty-nhw 5. Asphalt 15. Magenta 6. CDntrete 16. TvnarisK 7. Crushed salt (cowacted salt cdurm) 17. Culebn

1 B. Unnamd bwer Member 19. lmprrre halbe Boundary of accessible environment

---I

E. MI3138 2l.An~driteiaywaand b 22. Distuhed rockzone 23. Waste panel 24. Red d repstoy 25. Panel closures 26. Operation region 27. brperimerbl area 2B. MB139 29. Castile 3Ij. Brine reswvoi

cc1-01a-2

Figure 5. Model grid and material designations for numerical model used to simulate fluid flow performance assessment calculations; grid includes schematic of an intrusion borehole.

Spalling is assumed to occur if repository pressure exceeds the hydrostatic pressure of a column of drilling mud, or about 8 MPa(4). Because the waste panels are assumed hydraulically connected, the entire gas storage volume of the disposal area is accessible to a drillhole penetrating the repository horizon. Production of solids from a formation in the petroleum industry is generally

divided into two categories: dynamic response resulting from a gas kick during exploratory drilling(6), and steady-state production of solids from a producing formation(7). The modeling strategy for spallings releases at the WIPP incorporates both of these processes.

Three conceptual models for spallings releases currently exist for the WJPP. These models include the first conceptual model for spallings, as implemented in the October 1996 submittal of the licensing application; a mechanistically based model developed to address peer review concerns; and a model that will be implemented in recertification calculations. Federal regulations governing the WIPP specify that conceptual models used to predict repository performance must undergo peer review. A Conceptual Model Peer Review Panel, convened to ensure that the conceptual models used in the WIPP performance assessment reasonably represent possible future states of the disposal system, found the first conceptual model for spallings inadequate for performance assessment calculations. The mechanistically based model was developed to demonstrate that, although the initial model is missing some of the physics considered necessary to fully capture the process, predictions of the resultant releases are bounded by the original model. The third model is under development and will be used in future performance assessment calculations of spallings releases at WIPP. This model adopts features of the two earlier models. Each of these models will be briefly reviewed.

The original conceptual model for spallings assumes that the dynamic response of the system can be included in the results of a steady-state model. Laboratory experiments of a simulated blowout were conducted using sand as a waste surrogate@). Results indicated that production of solids would cease under steady-state conditions. In order to develop a mathematical model of the observed spalling process, it was assumed that experimental results could be correlated to the geometry of the apparatus, particle size of the surrogates, and cohesive forces between particles. Capillary and other cohesive forces are assumed to account for all adhesion between sand grains. An abstraction of these forces was devised, with fitting parameters identified to capture effects of geometry and scale. Details regarding the mathematical formulation of the model are found in the the certification application(1). Experimental results and interpretations are found in Lenke et a].(*). The original model calculates removal of solid material from the repository to the surface through assumption that erosional forces dominated the process. Solids transport within the waste regions and drillhole are assumed to be inherently incorporated in the mathematical formulation of the model.

Principal concerns of the peer review related to an apparent mismatch between the mathematical model of spallings and the experimental results(9). Experiments and analyses conducted subsequent to publication of peer review concerns resulted in development of a mechanistically based conceptual model for spallings(4). Rigorous evaluation of waste character was the first objective of the new model. A review of performance assessment predictions of iron corrosion, brine inundation, and pressure history supports a model of the waste as a heterogeneous mixture of variably sized materials. Chemical reactions within the waste lead to cementation of the original waste as well as cementation of degradation byproducts. Technical studies, including laboratory experiments, theoretical evaluations, and analog comparisons, provide the basis for a mechanistically based conceptual model for spall. A systems approach was applied to model development, as illustrated in Figure 7. The mechanistically based model is logically divided into three elements: motion of the mud column, gas flow within and fiom the repository, and variation of stresses in the waste. As discussed in Veeken et aI.(7), seepage stresses and erosional processes contribute to production of solids. However, peer review concerns focused on calculation of stress and potential material failure in the waste. Model calculations therefore focus on the volume of waste subject to tensile stresses likely to result in material failure. Erosional processes and solids transport are not included in this model. Instead, calculations of the volume of material subject to tensile failure are compared to solids releases predicted using the erosional model. Embodied in this approach is the assumption that the waste will consist of a competent, nongranular material unless stresses result in material failure. Such failure may

disaggregate waste matter and degradation byproducts, thereby making this material available for transport. The volume of material (Figure 8) likely to experience tensile failure is, in all cases, smaller than solids releases predicted using the erosional model. Peer review concerns were thus addressed, and the mechanistically based model was then retired. Development of a model that captures all the physical processes leading to spallings releases is under way.

655.32 m

Psurface = 10 1.32 kPa Tsurface = 27 C

1

182.88 m

1

1

289.56 rn

182.88 m

4

t

I-- Kelly Joint

j Surface

Ij Surface Casing 323 rnm ID

Drill Pipe 88.9 rnrn OD

Drill Collars 203.2 mm OD

Borehole

Figure 6. Schematic of typical drilling configuration for the Delaware Basin. WIPP is within the Delaware Basin

Underground Setting for SpaIl Event

Bottom Hole Pressure

C O M P U T A T I O N A L A P P R O A C H E S I I

SEMI-ANALYTICAL APPROACHES ~~ - ~~

.Quasistatic Solution Cavity Growth Transient

Solution

+ Verify/Confirm Verify/Confirm NUMERICAL APPROACHES

Conceptual Model Approaches *Two-Phase Finite Difference Pressure

Semi-Analytical OnE-mY Coupled

Decay (TOUGH28W)

I *PoroelaStic Finite Element Stress 1 Analog Comparisons Analyses (SPECIROM-32)

t ~ - j G D e i ~ i a T i o n s ’ I

I + Compare Volnma to CCA Model .

Figure 7. Diagram illustrating systems approach used to develop the mechanistically based conceptual

-0.W

0.350

0.300

0200

0.150

Q.100

5.000

Tensik Shngth, psi 15.Ooo

model for spallings releases.

0.4500.500 mo.400-0.450

m 0 . w o . w ~ mo.250-0.300 mo.2wo.m

00.350-0.400

00.150-0.200 00.1000.150

. 0 . ~ . 0 5 0 mo.omo.ioo

Figure 8. Volume of waste material predicted to fail in tension as a funtion of initial fluid pressure and tensile strength; direct releases predicted using the original model ranged from 1 to 4 m3.

The current model for direct releases resulting from spalling integrates the erosional processes of the first model with the coupled response of the mechanistically based model, and adds a solids transport feature to the system. In this model, releases to the surface are limited by the carrying capacity of gases flowing from the repository; a chemical engineering method based on fluidized beds transport forms the basis for this component of the spallings model. Transient drilling effects are directly coupled with gas flow and stress calculations in the waste. Uncertainty in the coupled physical processes and parameters is included via a probabilistic assessment based on engineering judgment. Waste evolution studies also support the model; reevaluation of the conceptual model of the entire waste disposal area is included in the new conceptual model for spallings.

The direct brine release mechanism is active when repository fluid pressures exceed the hydrostatic pressure (8MPa) of the drilling mud. Other parameters include the flowing bottom-hole pressure and brine saturation of the waste regions. To determine the flowing bottom-hole pressure, an iterative procedure is used based on a petroleum industry multiphase flow correlation developed by Poettman and Carpentedlo). Flow up the intrusion borehole during drilling is govemed by complex physics dependent on frictional effects and two-phase fluid properties. This behavior is much studied in petroleum engineering, and many correlations have been developed to predict flow rates and pressures in vertical two-phase pipe flow. The Poettman-Carpenter approach was chosen to calculate the necessary bottom-hole pressure because it accounts for multiphase frictional effects based on empirical (field) data from flowing wells, is one of the few correlations that includes flow between the drill pipe and open hole (annulus) in its development, and is easy to implement.

The wellbore is discretized into finite segments to capture the geometry of the open hole, drill pipe, drill collars, and casing(s). Similar to the spallings model, brine is assumed to flow to the surface through the annular area only. This is consistent with gas well blowout behavior since the inside of the drill pipe is filled with drilling fluid. The fluid saturations and panel pressures in the waste are calculated in a probablistic manned11 for each intrusion time; it is necessary to match these parameters to values predicted using the Poettman-Carpenter approach. As discussed in O’Brien et al.(3), an iterative process is applied to achieve this.

DISCUSSION AND CONCLUSIONS

The relative importance of direct releases to WIPP performance can be deduced by examination of the distribution function depicted in Figure 9. The curves presented on this plot represent the summation of total releases over the regulatory period as a function of probability. Regulatory compliance is demonstrated if the curve of total releases falls to the left of the limit. In addition to the curve showing total releases, curves are also presented for cuttings and cavings (as a single summation), direct brine releases, and spallings. Groundwater releases are not shown on this plot; these releases are several orders of magnitude smaller than releases resulting from direct releases. The cuttings and cavings mechanisms are the most significant contributors to the summation of releases. Because cuttings and cavings releases occur during every hypothetical intrusion into a waste panel, this contribution is largely a consequence of the assumption that the present-day drilling rate continues throughout the regulatory period. Although drilling within the WIPP area is currently prohibited, it is assumed that future generations will be unaware of this restriction. Direct brine releases are comparatively quite small, at least one order of magnitude less than those predicted for the other mechanisms. Contributions due to spallings releases become important at lower probabilities. The curve presented here represents spallings releases calculated using the original conceptual model. It is expected that the current model, discussed previously, will result in a smaller contribution to total releases for this process.

HEPA Containment I [§I 91 .I?@)] 100

io-’

10-2

10-3

104 10-6 10-5 104 10-3 10-2 10-1 100 I O ’ 10 2

Summed Normalized Release, R

- Direct Brine Release

I 8 .1111 I , 4 , 1 1 1 1 I ! 2 I I I t ? , , 1 , . , , , , , , 1 1 1 1 1 I l , , , , , I , 1 , , , , , , I 1 , , , ,

Figure 9. Summary of the total normalized releases (top curve), and normalized releases for spallings, cuttings and cavings, and direct brine release from WIPP. Normalized releases resulting from groundwater transport are several orders of magnitude less than direct releases and do not plot at the scale of this figure.

The assumptions regarding waste character and properties presented early in this paper significantly influence the magnitude of direct releases(]). These waste properties derived from the assumption that prediction of the physical evolution of the waste represented an insurmountable challenge. Recent studies show that waste character and properties can be predicted on the basis of known and expected processes within the disposal areas. This recognition has led to the support of an experimental program for waste surrogate development. In addition to providing material properties for the spallings release model, an additional benefit is reduction in uncertainty in the critical shear stress parameter implemented for cavings releases. The lower bound for critical shear stress is based on a literature value for extremely soft sediments; these sediments represent a conservative analog for the materials presented in Table 1. It is reasonable to expect that direct releases resulting from cavings will also be reduced as a more realistic evaluation of waste evolution is made.

A retrospective review of the conceptual models for the waste disposal areas and direct releases shows that the conservative approach implemented in model development has mixed results. A perceived benefit of implementing conservative models was that these models could be easily defended to regulatory and stakeholder entities. Defensibility of the models as conservative can only be achieved if the audience is cognizant of the expected system response, as compared to the predicted response, and if it is clear that the essential physical processes are included in model abstractions. For example, in the case of spallings this clarity was lacking, and a significant effort was required to demonstrate model conservatism. Strong coupling between repository and release models renders it difficult to convince stakeholders that the predicted response represents a conservative approximation

of all direct releases. The assumption of disposal panel connectivity is an example of this difficulty. Predicted inflows from a punctured brine reservoir (Figure 3) provide sufficient brine to the waste such that 100% of the iron could theoretically corrode in an intruded panel. This scenario produces a degradation pathway that could produce a waste form consistent with the model of a weakly cemented, granular medium. Presenting a concise argument to public stakeholders of the exceedingly small probability that this same panel will have high gas pressure sufficient to generate a gas blowout or spallings evenf or that the erodability of the degraded material is representative of all the remaining waste, remains a challenging exercise. At the same time, scientific stakeholders question the use of conservative models; instead, these entities favor implementation of engineering models and methods to demonstrate system performance.

The WIPP experience generally supports the typical position of the scientific community that adherence to realistic and expected processes in conceptual model development represents a preferred option. Predictions made using this approach are more likely to withstand public, regulatory, and scientific scrutiny. Waste evolution and form comprise a clear example of the success of the approach. Rational arguments made regarding expected processes and products provided a basis for waste surrogate development. Surrogates were developed for the most degraded scenarios such that material properties could be deduced and implemented in the mechanistically based model. Both the rationale for surrogate development and material properties estimates were presented for peer review and were deemed acceptable. Concerns of stakeholder entities were similarly addressed through logical presentation of the expected physical processes, rather than relying on abstract mathematical models. Although a need remains for some abstractions so that quantitative results can be obtained, it has become evident that the benefits derived by direct consideration of complex problems outweigh the computational difficulty of addressing these complexities.

Experience has shown that the use of simple and abstract models of the waste and direct releases is computationally straightfonvard and can be conservative. However, the difficulties that can arise defending simple abstractions should be considered when choosing models to simulate physical phenomena in performance assessments. Experience has also shown that it is possible to incorporate more realistic models and that use of these models yields improvements to predicted performance.

REFERENCES

(1) U.S. Department of Energy (USDOE). Title 40 CFR Part 191 Compliance CertiJcation Application for the Waste Isolation Pilot Plant. DOWCAO-I 996-21 84, Carlsbad, NM: United States Department of Energy, Waste Isolation Pilot Plant, Carlsbad Area Office (1 996).

(2) G.D. Callahan and K.L. DeVries. Analyses of Baclfzlled Transuranic Waste Disposal Rooms. SAND91-7052. Albuquerque, NM: Sandia National Laboratories (1 991).

(3) D. G. O’Brien, D.M. Stoelzel, P. Vaughn, and T. Hadgu. “Wellbore Flow Treatment used to Precit Radioactive Brine Release to the Surface from Future Drilling Penetrations into the Waste Isolation Pilot Plant”. HA WA ’98, International Conference on Hazardous Waste, Cairo, Egypt, December 12-16,1998. SAND98-0782C. Sandia National Laboratories, Albuqerque, NM (1998).

(4) Hansen, F.D, M.K. Knowles, T.W. Thompson, M. B. Gross, J. McLennan, and J.F. Schatz. Description and Evaluation of a Mechanistically Based Conceptual Model for Spall. SAND97- 1369. Sandia National Laboratories, Albuquerque, NM (1997).

( 5 ) P.D. Komar. “The Transport of Cohesionless Sediments on Continental Shelves.” In Marine Sediment Transport and Environmental Management. Pp. 107-125. Eds. D.J. Stanley and D.J.P. Swift. John Wiley and Sons. New York, NY (1 976).

*

(6) M.F. Khodaverdian, J. D. McLennan, I. D. Palmer, H. H. Vaziri and X. Wang. Cavity Completions for Enhanced Coalbed Methane Recovery. Final Report GRI-95/0432, Contract No, 5091-214-2220. Chicago, IL Gas Research Institute (1996).

(7) C.A.M. Veeken, D.R. Davies, C.J. Kenter, and A.P. Kooijman. “Sand Production Prediction Review: Developing an Integrated Approach,” Production Operations and Engineering, I991 SPE Annual Technical Conference and Exhibition, Dallas, TX, October 6-9, 1991. SPE 22792. Richardson, TX: Society of Petroleum Engineers. 335-346 (1991).

(8) J.R. Lenke, 3. Berglund and R.A. Cole. Blowout Experiments Using Fine Grained Silica Sands in an hisymmetric Geometry. Contractor Report 1996/7/32250. Sandia WIPP Central Files WP0#43 145. New Mexico Engineering Research Institute, University of New Mexico. Albuquerque, NM (1996).

(9) C. Wilson, D. Porter, J. Gibbons, E. Oswald, G. Sjoblom and F. Caporuscio. “Waste Isolation Pilot Plant Conceptual Models Supplementary Peer Review Report”. Sandia WIPP Central Files. WP0#43 153. Sandia National Laboratories, Albuquerque, NM (1 996).

Flow Strings with Application to the Design of Gas-Lift Installations,” Drilling and Production Practice, American Petroleum Institute. 257-3 17 (1952).

(10) F.H. Poettman and P.G. Carpenter; “Multiphase Flow of Gas, Oil, and Water Through Vertical

Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for thi United States Department of Energy under contract DE-AC04-94AL8500.