A summary of the models used for the mechanical response of disposal rooms in the Waste Isolation Pilot Plant with regard to compliance with 40 CFR 191, Subpart B
About: The article was published on 1993-08-01 and is currently open access. It has received 7 citation(s) till now. The article focuses on the topic(s): Waste Isolation Pilot Plant.
Summary (9 min read)
- Results using the mathematical models for disposal room response are described, beginning with closure of empty rooms and becoming progressively more complex.
- Developments currently in progress to improve the evaluation of the disposal room performance are addressing the coupling between brine flow and closure and the two-dimensional capability for analyzing a complete panel of rooms.
- The coupling is accomplished using closure surfaces that describe the relationship among porosity, total amount of gas in the repository, and time.
- After the containers are placed in the room, crushed salt or some other type of backfill will be placed over, around, and between the containers to fill much of the remaining room void space.
- Specifically, the majority of work reported in this document refers to post closure 40 CFR 191B performance assessment.
- Exceptions are those areas regarding RCRA, which are clearly identified.
- The contents represent progress in development over a period of approximately four years, ending March 1992.
- Before discussing these objectives further, a brief discussion of the waste environment after a room is filled is necessary.
1.1 The Disposal Room Closure Process
- Void volume increase will continue until gas generation ceases and pressure equilibrium is eventually established with the suri-I.
- The disposal areas are composed of haulage ways, access drifts, and eight panels, each containing seven disposal rooms.
- Or the purpose_ of post.-._closure R('RA, gas leakage mechanisnls must be defined, also known as Rounding formation, t.
- Io demonslrnle lh;il VOC gase_ do not escape to Ihe accessible en_ir_)nllleni in ex_'ess of the regulated amounts.
- While the steps in normal closure of the repository are hundreds to thousands of years in duration, the,,' can be interrupted at any time by events that disrupt the state of the repository.
- ('h_lplet _ is a briet" desctiplit)n of some t" the development w¢)rk in progress to improve the n!_)dei I_)r ex_lmple, the consequences ot' an inadvertent human intrusion have yet to be _ddre_ed.
- Allhough the methods for compleling this step are already in place.
- _ brine t'1o_ on gas generation and closure by coupled-flow analysis is also being _lddte_ed.
- Is well a_ the t'easibilily of various panel-scale modeling schemes, ,_ major (,bjective ot" this report, a proposal of the way the disposal room model can be ¢_upled inlo assessment of repository performance, is in Chapter 7, also known as.
- The mechanism ofl nt'_rm_ltion transfer is currently specification ot" surfaces des_._ibing the relationship of" lhree _ariables, pc,rosily, total amount ot" gas in the repository, and time, For various assumed sequellcesoi _e_,enls, i:inall_, Chapter 8 is a summary of the contents of the report and recommendations for further de,, elopment. !-5.
2.1 Room and Panel Dimensions
- The configuration for closure analysis is generally a single disposal room.
- Because of computational limitations, most past numerical closure calculations have been limited to 2-D plane strain analyses that examine closure of the 4-m-by-10-m cross section of an infinitely long room.
- These rooms, in turn, either are assumed symmetric, with vertical symmetry planes at the center of each pillar and each room, as would be valid for an infinite series of rooms, or modeled as a single, isolated room.
- Extensive calculations of 3-D effects at intersections of the rooms with the panel access drifts have not been completed largely because of restraints imposed by the numerical methods and computation time.
- Some results are available from calculations by Argtiello et al. (1989) and Argtiello (1990) .
2.4 Backfill Emplacement
- A reference stratigraphy for the region surrounding the disposal rooms was also recommended by Krieg (1984) , as were reference mechanical properties for dominant nonhalite features such as anhydrtte and polyhalite marker beds and clay seams.
- Specification of the parameter values used in the SANCHO model will be deferred until the next section on backfill consolidation because they are also part of the backfill model.
3,2.1 Model Description
- A mathematical model that defines consolidation of two backfill materials, (I) pure crushed salt and (2) 70% by weight salt and 30% by weight bentonite, is described in this section.
- Based on a comparative study of the performance of both backfills, salt/bentonite has been proposed as the most desirable material for use in WIPP disposal rooms (Butcher et al., )991a) .
- Information relative to the usefulness of other, alternate backfills considered by the Engineered Alternatives Task Force (EATF) (US DOE, 1991) is not included because the exact nature of these candidates has yet to be established.
- The model for only one backfill model is described, because with suitable redefinition of material constants it can be applied to both salt-based backfill materials.
- Like pure solid salt, crushed salt continues to deform under stress with time.
- Such as are illustrated in Figure 3 -1 are made using the original model constants, even if their representation of the data is imperfect.
- The reason for retention of the original constants is that frequent alteration of the model constants would make comparison of the results of ongoing calculations exceedingly difficult.
- Updating the material constants is planned only if new data differ significantly from the model, which has not been the case so far, or if a more accurate description is required for performance assessment, which currently does not appear to be necessary.
3.2.4 Backfill Constitutive Equations
- Material constants for the respective constitutive equations are summarized in Table 3-1.
- Studies have shown that the exact method of specification of the elastic response of the various types of backfills and the values used have little influence on compaction predictions .
- Therefore, the reader is referred to the documentation of the various models (Callahan and DeVries, 1991; Sjaardema and Krieg, 1987; Weatherby et ai., 1991a) for more detail about how elastic constant values were estimated.
3.3.1 Model Requirements
- In turn, these factors help determine rates of brine transport throughout the disposal room.
- The porosity of the waste at a given time also determines the amount of soluble radionuclides contained in brine-filled voids within the waste.
- The approach used in compaction model development was based on several considerations.
- First, simulated waste was used because of the difficulties inherent in working with real waste.
- This approach was justified because the mechanical response of the waste depends entirely on its nonradioactive constituents, such as plastics, cloth, sorbents, etc.
- The presence of trace radioactive elements has no effect on compaction.
3.3.4 Compaotion Models
- Although the deviatoric responseof the waste has not been characterized.
- And OeVries, 1991) are shown in The curve labeled "series," representing elements of the various waste components in series (the same consolid,tion load acting on each component), appeared intuitively to be more representative of actual compaction conditions and was selected for use.
- The parallel curve represents elements in parallel (with varying loads applied to each waste component according to its portion at" the inventory).
3.5 A Model for Gas Generation
- Gas produced by decomposition of cellulosic waste, corrosion of metals, or radiolysis of TRU waste has always been a subject of interest to the WIPP Project (Lappin et al., 1989, Sec. 4.2).
- The presence of gas is beneficial in the sense that gas occupies void volume that would otherwise be eventually filled with brine.
- In the opposite sense, gas pressurization could force flow of radioactive brine out of the repository, it could inhibit closure, it could open preexisting fractures and provide paths of easy transport away from the repository, and VOCs could also become entrained in gas and escape from the repository.
- From anoxic metallic ¢orrosl(n of drums, metal boxes, and metallic _(nsliiuenl_ of the waste.
- These reactions require w.aler and produce large amounts of hydroRen gas.
- Water availability in the form of brine (brine availability) deieimines _.helhet these reactions can occur and their rates.
- For e_ample, thai the rate of corrosion of metal waste immersed in brine i_ orders of malnilude fa_ler than the rate of corrosion of metals exposed to water vap_)r 1991a. Sec_ 3,3), A gaseous environment may prevent contact of the waste with brine and thus eliminate the principal mechanism for radionuclide migration: transport of soluble radionuclides in brine.
3.6.1 Gas Pressurization
- For the case of maximum gas potential (e.g,, Section 3.5), the curve in Figure 5 -9 that represents the void fraction history of a perfectly sealed room shows that the void fraction decreases to a minimum value (closure stops and the void volume within the room begins to increase) at about 150 years.
- The gas pressure corresponding to this reversal is approximately 10 MPa, well below lithostatic stress of 14.8 MPa.
3.6.1,1 REGULATORYIMPLICATIONSOF GAS GENERATION
- "['he conceptual model of void volume creation or equivalently the question of "where does the gas goT' is of great interest in regard to the performance of the repository (e.g., Lappin et al., 1991) .
- In terms of the radioactive standard (40 CFR 191, US EPA, 1985) , the issue can be redefined to whether gas escaping from the repository can accelerate or retard flow of radioactive brine from the disposal rooms.
- Gas, by opening fractures or rendering marker beds more permeable, could create short circuits through which radioactive brine flows.
- Three factors tend to moderate this concern: 3-21.
While void volume creation to accommodate
- Compliance with this standard is necessary because many of the CH TRU wastes planned for disposal in the WIPP repository are expected to contain small amounts of hazardous components and therefore are classified as mixed waste.
- The nature of pressure induced volume change is important because VOCs are expected to become entrained in corrosion and/or decompositional gas.
- Therefore, the burden of performance assessment with regard to the RCRA is to demonstrate that the rates of release of soluble or gaseous VOCs beyond the WIPP boundaries are acceptable, and this, in turn, depends on how much gas moves away from the repository.
- The effects of gas mi_,ration must also be considered when addressing compliance with the National Environmental Policy Act of 1970 (NEPA).
- NEPA documentation does not focus on specific regulatory guidelines but involves the use of "best estimate" and "degraded property" sets of input parameters to investigate repository safety.
3.6.2 Fracture Concepts
- A simple fracture model has been developed to explore the behavior of fractures in the Salado and has been applied to two types of response, Case i represents an existing fracture in halite or an interbed extending beyond the DRZ ",at can open, given sufficient tensile stress, but is not penetrated by gas in its closed state.
- Because the portion of the crack beyond the DRZ is not penetrated by gas, the pore pressure in this region is zero.
- Thus, the term "irnpermeable" is used in the sense that the fracture beyond the DRZ does not transmit gas and create a gas pore pressure unless stress conditions cause the crack to open.
- This case is, in fact, a test of whether new fractures will initiate in the salt, as discussed further in Section 220.127.116.11.
Case 2 represents a (permeable)
- Fracture in a nonhalite interbed or clay seam, or a discontinuous interface between an interbed or clay seam and halite, where gas can penetrate even when closed.
- Fracture opening occurs when stress normal to the fracture layer becomes f tensile.
- For the permeable fracture model, gas can diffuse into the fracture prior to its opening and, like the impermeable fracture, it too opens if tensile stress is present.
- This assumption is justified from results of two-phase flow calculations that show that the pore pressure gradients behind the gas penetration front are small near the room .
- In addition, the impermeable fracture model has been applied using the initial conditions and boundary conditions for a single room in an infinite array of rooms; whereas the permeable fracture model has been appli,,d to both a single room in an infinite array and an isolated room boundary.
3.7 Thermal Effects
- CH TRU waste does not generate large amounts of heat, and any heat produced is expected to be rapidly dissipated because of the high thermal conductivity of salt (i.e., localized hot spots within the repository are not expected).
- Nevertheless, heat loads from RIt waste were not included in the disposal room models becausethe complexity introduced by inclusion of heat flow in the analysis did not justify the small increase in closure rate caused by elevated temperatures in the salt.
- One of Arg0ello's conclusions was that although the temperature of the salt adjacent to the canisters increased by at least 3.50K during the 6-year simulation (2.5-year thermal load duration), changes in closure results caused by thermal effects were almost imperceptible.
- Furthermore, more rapid closure is considered beneficial to the disposal room performance in terms of 40 CFR I_IB because decreases in permeabilities of the backfill and waste proceed more rapidly.
4.2 Code Oesorlptlons
- WIPI _ di_po_al realm behavior have consistently required the m_st advanced analysis techniques.
- Development ot' new il_, new capabilities, new constitutive equaiic)ns, and/or new m,dels has been needed for e=_chnew WIPP problem.
- An additinal complication is that experimen!,l verification of prediction_ i_ unlikely because the calcuiations extrapt)late di_p,)_al _¢)_)mresponse.
- _zndthe bottom rnv_ .f n(_es repre_entinll the hori_orll_l line of _mmelr_, through the nlid_point f the room i_ =1he fi_ed allam.tl metier1 in the _erlleal dire¢li.n.
- Fhe pressure boundary ¢.ndili.n i_ applied t)nl_ to the n.de_ ahmg the lop of the mesh.
4.3.3 Other Assumptions
- An important assumption for simplifying closure calculations was that closure of the disposal room is not greatly affected by the gravitational forces acting on the material in the immediate vicinity of the room.
- This assumption is supported by early studies showing that the near-field gravitational forces and the local variation in the initial stress field with depth have little effect on closure rates.
4,4 Isolated Room Initial and Boundary Conditions
- While the infinite array of rooms is the easiest to model, closure results from such calculations are limited to approximations of the response of the center room of the seven-room panel, where adjacent rooms influence closure.
- In addition, the array approximation does not represent the end rooms of the panel because these rooms have no adjacent rooms on one side of them.
- Itowever, in contrast to the array room mesh, the isolated room mesh configurations were larger because there are no vertical symmetry boundaries corresponding to the pillar centerline.
- The assumptions were also made that (I) closure of the isolated disposal room was not greatly affected by the gravitational forces acting on the material in the immediate vicinity of the 4-7 r_)c)n), ( 2) their the qti_|liKrL|ph', _ L'_uld als be iI, im, t'd.
5,1.1 Empty Room Calculations
- Ttae simplest closure calculations addressed the creep response of empty rooms in salt over short periods of simulated closure.
- These results were used for comparisons with early room closure measurements that were becoming available from underground tests at the WIPP Site.
- Thus, information about the rate that an empty room closed and the amounts and theoretical solid densities of waste and crushed-salt backfill within the room was sufficient to compute void fraction variation with time (Lappin et al., 1989) .
- Results for void fractions below about 20% were ignored because the assumption of zero backstress was clearly questionable at such dense states.
Additional insight into resolution of the discrepancy has recently been obtained from calculations
- With a new code called SANTOS by Stone and Argtiello , which were undertaken to investigate determining the difference between emptyroom closure results computed using small and large deformation formulations.
- The SANTOS finite-element code was modified for this study so that it could generate both types of solutions.
- In addition, the effects of incorporating contact surfaces to prevent overlapping of the corner elements of the room configuration were determined.
- The contact surface feature is important because once overlapping occurs, as illustrated in Stone and Argiaello's memo , the computational results no longer have any physical meaning.
- Ehgartner's (1990) SPECTROM-32 curve shown in Figure 5 -1 was also obtained using small deformation behavior.
All SANCHO calculations
- After and including the results described by Weatherby et a!. , Figures 5-4 and 5-7, were obtained using the new consolidation curve.
- All SPECTROM-32 results after and including the results described by Callahan and DeVries , Figures 5-5 , 5-6, and 5-8, were also obtained using a similar curve, although the method of incorporating it in SPECTROM-32 differed from the way in which consolidation data were incorporated in SANCtlO (e.g., Section 3.3).
- Gas generation was assumed to be zero for all calculations, Several conclusions are evident from the closure estimates for rooms filled with waste and backfill ( .
- Fisure 5-8, SPECTROM-32 results for a disposal room filled with crushed salt/bentonite backi'ill and TRU waste (Callahan and DeVries, 1991),.
- , Weatherby et al. (1991b) , and Brown and Weatherby that investigate the behavior of a disposal room when both the total gas potential and gas-generation rates are reduced.
- In all cases, sufficient brine was assumed present to allow the corrosion reactions to go to completion.
18.104.22.168.1 Void Fraction Resultsfor Varying Amounts of Gas
- While gas is being generated, the value of f reflects the net amount of gas existing within the room at a given time, i.e., the amount of gas that has been produced up to that time less the amount of gas that has leaked out of the room.
- The effects of gas generation on disposal room closure .
- Weatherby et al., 1991b) was that a different value of the solid volume of the waste was used to compute void fractions (.
- In all cases, pressures eventually rose above and then gradually decayed to lithostatic pressure.
6.2.1 Reeuite of Caloulatlone Examining the Role of Fraoturee on Room Pressurization
- The finite-element mesh representation used for both these analysesis shown in Figure 5-!5 (Argt_ello et al., 1992) .
- These calculations predated introduction of a waste compaction relationship in SANCHO based on experimental data and instead used the "old" SANCHO consolidation relationship for waste .
- Becausethe old compaction relationship assumed little backstress until low porosities were reached, these results, like the gas expansion results described in Section 5.1.4, predict lower minimum void fractions than are currently expected.
- This discrepancy is considered to have little effect on the objective of" the calculations, which was to explore two different conceptual models of the role of' fractures during disposalroom pressurization.
- The reason for the fracture-related calculation described as Case 2 is that the geological formation surrounding the repository is not a monolithic region of salt but rather is a layered structure o£ salt, horizontal anl'=ydrite interbeds, and clay seams.
- These conclusions suggested two major simplifications in the observed two-phase response: (I) gas migration within the marker bed could be approximated as a moving boundary, and ( 2) the gas in back of the moving boundary could be assumed to be at the same pressure as the gas pressure within the disposal room.
- Opening was assisted by the increased pore pressure: once the fracture opened, the open part of the fracture was assumed to have the same pressure as the gas pressure in the room.
- The slower closure rate resulted in higher void volumes within the isolated room, which provided greater gas-storage volumes than for the array room at a given time.
- In addition, supporting evidence may come from planned experimental work to show that the marker beds in the undisturbed region outside the DRZ contain pre-existing fractures, will readily accept fluid pressurization, and will dilate as pressures approach lithostatic.
6.1 Human Intrusion
- The computational methods described in the previous sections can be used to predict the nonlinear behavior of the subsequent consolidation process.
- The procedure appears straightforward, Closure calculations, such as th," _as calculations described in Section 5.1.4, are stopped at the time of human intrusion and the amount of gas within the room redefined, This process requires assumptions about how much gas escapes instantaneously up the borehole and what the subsequent borehole-leakage rate is, which can be obtained by independent calculations.
- Some approximations of closure after depressurization that can be used in lieu of the availability of actual closure histories are discussed in Chapter 7,0.
6.2 Coupled Flow-Closure Calculations
- Thus, the race is on between brine/waste interaction, gas-pressure buildup, and brine rejection by gas.
- The outcome of this competition cannot be predicted without including brine flow in the calculations, I Any change in the current assumption that microbial decomposition occurs without brine consumption would also influence brine availability.
- Equations3 ,rod4 apply after P reachesthe conJt,u_t UthoJtaticpressureof 14.8 MPa.
- In both of these expressions, the pressure is held constant at 14.8 MPa and I/, a 1/-Ys is the void volume.
Influence of Gas Potential
- In these analyses, the room wu assumed to be _led and sealed immediately after excsvltion.
- For some waste forms, this may not be achievable under'lithostatic stress conditions.
- Tkts.rein/_atlon of the room occurred before the room pressureexceeded the lithostatic stress level.
- There was a significant v_iation of the minimum void fraction with change in gu generation potential.
- This variation is shown in Figure 5 ; note that all of the predicted peak values are |rester than the lithoJtatic pressure (14.8 MPa).
Influence of Gas Generation Rate
- To investig,_tethe e_ect of varying the gas generation rate (_',).
- This value w,, reduced by a factorof fivefrom the bueIine case.
- Two differentcases wereevAluAted fora simulation September17, J990 time of 4000 years.
- The pressure built up st a much slower rate, which was consistent with, the slower rate of lpU generation.
- The room pressure was LUowedto vary accordinlLto Equation I until re,chin8 the lithostatic level (14.8 MPa).
- At 500 ye_s, both rates dropped, and then the ra',e of loss slowly increMed.
- T With the second usumption, the authors see that the TRU wute is three times stiffer than that obtained usin| the first assumption.
- In Figure 1 Although there is a significant difference between the curves obtained using the two assumptions, the procedure for adopting Assumption (2) to produce the stiffer TRU waste model is simple.
- Therefore, the same material model adopted for the TRU waste and Included in SPECTROM-32 can be used to represent the stiffer TRU waste and obtain the hilher values of porosity.
- Two different methods were used to zenerallse the TRU wute functional form (Equation 1) to three-dimensional states of'stress.
- The two methods produce TRU waste stlffnmNs that vary by a factor of 3.
- The firet tenersllsation (Assumption (1)) producm com,erv_tlve results with rmpect to the backfill msterlal; whereM, the second zenersJlsatlon (Auumptlon ( 2)) produces conservstlve results with rmpect to the TRU waste when porosity is the vLrlable belns considered Also similar to anhydrite "b", once gas has penetrated the interbed, its pressurewithin the interbed follows room pressurequite closely .
- Also, once penetration occurs, gas pressurein the interbed tracks gas pressurein the room quite closely.
Did you find this useful? Give us your feedback
Cites methods from "A summary of the models used for th..."
...A Typical Porosity Surface Used for the 1992 Comparison of Predicted WIPP Performance With 40 CFR Part 191, Subpart B (Butcher and Mendenhall, 1993). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 Figure 2....
...A Typical Porosity Surface Used for the 1992 Comparison of Predicted WIPP Performance With 40 CFR Part 191, Subpart B (Butcher and Mendenhall, 1993)....
Cites background from "A summary of the models used for th..."
...Fracturing and dilation in response to excavation is expected to create a zone of enhanced permeability, porosity, and interconnectivity that decreases with distance from the excavation (Stormont, 1990), However, fractures in the DRZ are expected to close and heal as room closure and consolidation reach their maximum extent, returning the zone to its original, undisturbed state (Butcher and Mendenhall, 1993)....
Cites methods from "A summary of the models used for th..."
...…Figure 1: A typical porosity surface used for the 1992 comparison of predicted WIPP performance with 40 CFR Part 191, Subpart B (Butcher and Mendenhall, 1993) ...13 Figure 2: A model and data flow diagram for the WIPP CRA PA. .................................14 Figure 3:…...
...Figure 1: A typical porosity surface used for the 1992 comparison of predicted WIPP performance with 40 CFR Part 191, Subpart B (Butcher and Mendenhall, 1993)...
...Figure 1: A typical porosity surface used for the 1992 comparison of predicted WIPP performance with 40 CFR Part 191, Subpart B (Butcher and Mendenhall, 1993) ....
Cites background or methods from "A summary of the models used for th..."
...Ideally the Disposal Room Model should address the configuration of the entire repository, which is three-dimensional....
...Additional information about closure surface configurations can be found in Butcher and Mendenhall (1993)....
...The flow diagram for this part of application of the Disposal Room Model is shown in Figure 2....
...This information must be provided for all regions of the configuration, and represents the bulk of the developmental effort for the Disposal Room Model....
...Major changes involved dividing the Young's modulus value of the reference creep law by a factor of 12.5, and greatly simplifying the stratigraphy of the Disposal Room Model, eventually changing it to a uniform formation of 100% halite (Morgan, 1993b, pp. A-92 to A-94, and conclusions)....
Related Papers (5)