ML17298A683

Review of Research on Performance Assessments for Disposal of High-Level Waste and Spent Nuclear Fuel in a Generic Geologic Bedded Salt-Hosted Repository

ML17298A683
Person / Time
Issue date:	09/30/2017
From:	Jin-Ping Gwo, Self S Division of Spent Fuel Management
To:
Gwo J
Shared Package
ML17271A120	List: ML17298A683
References
Download: ML17298A683 (22)

Text

Review of Research on Performance Assessments for Disposal of High-Level Waste and Spent Nuclear Fuel in a Generic Geologic Bedded Salt-Hosted Repository Prepared by Stephen Self and Jin-Ping Gwo Division of Spent Fuel Management Office of Nuclear Material Safety and Safeguards U.S. Nuclear Regulatory Commission September, 2017 Preamble This report has three parts: 1) Past work on (generic) salt repositories for high-level waste in the US and elsewhere, concentrating on performance assessments; 2) Recent work done by the Center for Nuclear Waste Regulatory Analyses (CNWRA) staff on creep of salt under pressure and heat; 3) How adaptions might be made to the Scoping of Options and Analyzing Risk (SOAR) model (Markley et al., 2011), NRCs generic performance assessment model, to accommodate future investigations of salt as a repository medium.

The acronym HLW is used for heat-producing high-level radioactive waste, and in this report the term incorporates spent nuclear fuel (SNF). Details of the types of waste to be disposed are not covered in this report, although they are important and, in a non-generic case for a repository, would be a driver for performance assessment (e.g., Clayton, 2011). Waste inventories and heat/thermal management must be assumed in a deterministic way for generic repository models (Clayton, 2011; Clayton and Gable, 2009; Robinson, 2013).

1. Past work in the US and elsewhere on (generic) salt repositories for high-level waste, concentrating on performance assessments 1.1. Introduction Natural salt formations have long been considered for permanent disposal of hazardous waste.

The National Academy of Sciences (1957) proposed that high-level nuclear waste (HLW) could be permanently disposed in salt cavities. The U.S. stopped considering salt as a potential host rock in 1987 when the amendments to the Nuclear Waste Policy Act of 1982 restricted consideration of geologic disposal of HLW to the site at Yucca Mountain, Nevada, only. More recently, permanent geologic disposal of HLW in salt (e.g., Hansen and Leigh, 2011) has been one of several options considered by the U.S. Department of Energy (DOE; e. g., DOE, 2014a) since the Blue Ribbon Commission report (BRC, 2012) on disposal of HLW was published.

1

In the report by DOE (2014b) on various host-rock options, salt comes out very favorably as a repository medium for HLW compared to other host media. This is because of its attributes (see below) and because it can accommodate all types of nuclear waste due to its high thermal conductivity and relative ease of mining.

The advantages of salt as a repository host-rock have been well documented for 60 years (Hansen and Leigh, 2011), and include:

• Salt has a relatively high thermal conductivity.

• Salt is essentially impermeable - mostly impervious to water flow - with low porosity and permeability. This leads to a greater reliance on the natural barrier in salt-hosted repository models.

• Salt rock has a very low water content. This, and the previous attribute, generally leads to a reducing environment in a mined repository.

• Salt undergoes plastic deformation, and fractures are self-sealing.

• Salt can be mined easily, and it has the strength to support the roof of excavations, attested to by long-open (long life-time) mines.

• Salt has been geologically stable for millions of years. Further, salt deposits are located in seismically stable geographical areas of the United States.

• There is a wide geographic distribution of salt deposits, and thus many potential sites in the U.S.

By contrast, salt as a host-rock has few disadvantages, one in the U.S. being the associations with resources such as oil and therefore the related uncertainty of future inadvertent human intrusion into a repository.

Salt rock occurs as bedded or domal (salt domes), the former being the original post-depositional state. Both are recrystallized from the original state, and both have been proposed as a promising repository host-rock. They share the same general advantages and disadvantages with respect to suitability as a repository. Dome salt has been deformed by diapiric rise and spread sometime in its history and has less pore space and fewer, if any, continuous layers of different lithology (clayey salt, for example). Bedded salt has more brine pockets and other fluid inclusions and this is possibly important to its performance as a repository host. Differences and similarities exist for bedded and domal salt and these are important at different scales when applied to nuclear waste disposal (Hansen, 2016).

1.2. Scope of the report and state of the problem 1.2.1. Definition and scene-setting A performance assessment (PA) is also called a safety assessment in some literature. A reference case (for a geologic repository, which is equivalent to a generic repository model) is the basis for the PA. Except for the Waste Isolation Pilot Project (WIPP) site in New Mexico 2

(DOE, 1996; Helton, et al., 1998), all reference cases for salt repositories are generic, i.e.,

based on general salt principles but not site-specific. The PA for WIPP is not based on HLW.

The German ISIBEL project was site-specific but did not reach completion (see below, Weber et al., 2011), and reference cases vary with the type of salt considered (dome vs bedded) and the type of waste being considered. It is assumed here that a PA should be a safety assessment involving probabilistic estimates for dose to a receptor or radionuclide releases, but no generic cases for HLW go into this aspect. Those that estimate doses are based on deterministic estimates of parameters.

While many technical studies exist on salt as a HLW repository host rock, relatively few full performance assessments have been completed. Most authors acknowledge that generic assessments are necessarily limited, as site specific information is needed for a full PA. As previously noted, the most site-specific PAs that have been prepared are for the U.S. WIPP facility and the German ISBEL project (e.g., Weber et al., 2011; see comparisons and summaries in Camphouse, 2015, and DOE, 2015). The recent work sponsored by DOE is detailed and supported by modelling on some thermal-hydrologic-mechanical (THM) aspects.

The DOE-sponsored studies have been strongly influenced by the DOE experience with the WIPP site and specifically concern bedded salt.

Hansen and Leigh (2011) summarize a number of the technical studies on salt, and provide an overview of the PA model approach for a salt-hosted repository developed at Sandia National Laboratories. They acknowledge that there is currently no performance standard for disposal of HLW in salt, (apart from 40 CFR 191, EPA for a generic, non-specific host-rock case), and stated that at a minimum, consideration of HLW disposal in salt would require changes to the legal framework specified in the National Waste Policy Act. The lack of a specific performance standard and current governing legislation (and regulations) for a HLW repository in salt places some constraints on PA development. Nonetheless, even in the absence of a complete regulatory framework, a PA can still provide useful risk insights as long as the uncertainties and limitations are recognized.

An important area of technical uncertainty that bears on a PA is the deformation behavior of rock salt under thermal loads. Many studies recommend reducing the thermal load by management of the types, sizes, and packing density (placement in the mined repository) of waste packages (e.g., Robinson et al., 2013).

1.2.2. Earlier work Some PA-relevant work for a salt-hosted repository was conducted pre-2008 in the US and Germany, including WIPP. The 1980s-1990 NRC-supported work set the scene for later PA models (e.g., Pepping et al., 1983) but most of the information and scenario development is now superseded by post-2008 models for a generic salt repository PA (see below). The ISIBEL project in Germany (summarized by Weber et al., 2011) was a mine in dome salt at Gorleben that was going to be developed as a repository for HLW. Various choices of deterministic 3

parameter values were made for the study so that features, events, and processes (FEPs) could be chosen, which used the attributes of the salt dome to constrain some parameters. There was a 10-year moratorium imposed on the project in 20001, but some work continues. In 2014, GRS (a government-funded German technical support organization) published a report on human intrusion, the only disruptive scenario considered for the Gorleben site (GRS, 2014).

Relative to PA, the ISIBEL study looked at normal behavior of the repository, and at a special scenario: A big brine pocket was envisaged by each emplacement hole (this may be the emplacement site for a single waste package). The project designers saw this as an event that would be considerably below the limit of acceptable probability. In this special scenario, as convergence presses the brine out, the sealing salt behavior is critical. If this has been modelled taking into account all processes that can adsorb radionuclides or hinder radionuclide movement, then radionuclide releases are small. If not, then there could be issues in the performance of the salt. This underlines that the behavior of crushed salt used as a seal in repository drifts is a property of great interest (see below), and it also shows that special circumstances must be invoked in order to have significant releases from a model of salt repository behavior.

1.2.3. More recent work Since 2008 there has been a U.S.-German cooperative agreement on salt repository research, design, and operation. This has produced the Sandia Laboratory-led, DOE-supported models which subsume the German work (e.g., Camphouse, 2014), as well as a series of workshops (e.g., DOE, 2013a). Some of the work has been aimed at dose estimates, mainly deterministic, but much can be cast as developing the generic repository concept or reference salt repository case, and explores issues such as thermal loading. The Nuclear Energy Agency (NEA) of the Organisation for Economic Co-operation and Development (OECD) also formed the Salt Club in 2011. Its mission is to develop and exchange scientific information on rock salt as a host rock formation for deep geological repositories for HLW and long-lived radioactive waste.

For non-heat-producing waste, there is the Waste Isolation Pilot Project (WIPP), New Mexico, which is a mined repository in bedded Permian salt at ~ 650 m depth (2150 ft) (WIPP Energy, 2016). The repository is designed to contain radionuclides from trans-uranic waste that is not significantly heat-producing. It began accepting waste in 1999. A probabilistic procedure has 1 A 10-year moratorium later imposed by the German government on the Gorleben underground research laboratory (URL) and repository site ended in 2010. In July 2013, the German Parliament passed the Repository Site Selection Act, which called for a new science-oriented and transparent site selection process, with the involvement of multiple stakeholder groups in determining the site selection criteria. The Act calls for the proposed plan for site selection and exclusion criteria by the end of 2015.

The Repository Site Selection Act, effective on 27 July 2013, halted exploration work in the Gorleben salt dome, although it is still eligible for consideration as a disposal site during the newly legislated site selection process (GRS, 2014). According to the rules and criteria of the Repository Site Selection Act, operations in the Gorleben salt mine are to be reduced to only what is absolutely necessary.

4

been used for estimating the radionuclide releases to the accessible environment associated with each possible future scenario that could occur at the WIPP site over the next 10,000 years.

No releases to the accessible environment are associated with the undisturbed scenario. After a FEP analysis, exploratory drilling for natural resources and mining of potash were identified as the only likely significant disruptions at the WIPP site with the potential to affect radionuclide releases to the accessible environment. Results of a PA confirm that direct releases from drilling intrusions are the major contributors to radionuclide releases to the accessible environment. The WIPP PA is a demonstration that potential cumulative releases of radionuclides to the accessible environment over a 10,000-year period after disposal are less than specified limits based on the nature of the materials disposed.

Prior to undertaking a PA for HLW, there is need to set up a reference salt repository case (Vaughn et al., 2013; Freeze et al., 2013) - as stated above, DOE models are generally influenced by the WIPP experience. An example is the DOE study (Freeze et al., 2013) entitled Generic Deep Geologic Disposal Safety Case. In this study, salt is considered along with other potential host media and a mined repository is modelled, for most cases at an equivalent depth to WIPP (600-700 m). However, it is not site-, nor waste-type-, specific like the WIPP case.

In general, these reports describe steps in the models but do not give results based on probabilistic estimates of properties and rates. Key processes and features included in the DOE models are the creation of the mined drifts and alcoves, the engineered damage zone (EDZ),

the use of crushed salt backfill, the inclusion of impure layers in the host halite rock (with higher permeability), and, usually, an aquifer in the strata above the salt-host-rock layer.

A common question is: when does large-scale deformation of the salt start, such as room closure? It starts immediately an opening is made by mining, and could be well-developed within 100-200 years, or sooner. The rate of deformation increases with increased temperature of the salt (e.g., Ghosh and Hsiung, 2013; Hansen and Leigh, 2011). Some studies claim that the EDZ will recover back to host-rock properties within a few years and that the crushed salt backfill will reconsolidate within 20 years, especially if the salt is heated by the waste (Rutqvist et al., 2014.). One advantage is that salt properties lost due to mining (the EDZ) will be replaced as the salt heals. If this process takes on the order of a century or less, then the repository may evolve as early-mined alcoves are closed, perhaps while the repository is still in the pre-closure stage. Such a procedure needs to be pre-planned before and during the filling of the repository (Robinson, 2013; Hansen and Leigh, 2011).

1.2.3.1. Steps in the PA process Steps are usually defined (Freeze et al., 2013) as: 1) FEP analysis; 2) Scenario Development;

3) Model implementation (in the NRC case, SOAR). One consideration is the type of waste considered, and this can affect steps 2 and 3. Waste types to be disposed in any future salt-5

hosted repository cannot be predicted at present, as several options are being considered (DOE, 2014a). The Sandia PA process is illustrated on Figure 1.

The analysis of FEPs leads to those FEPped in, usually site specific, and FEPped out, usually on the basis of probability and consequence; this is also referred to as screened in and out.

The FEPs that are considered in are used to make up scenarios (scenario development). For salt, the host rock is the dominant feature; it has low moisture and permeability (leading to fewer perceived leaks from the host rock than other media). Scenario Development consists of the construction of combinations of FEPs after screening - and should include the nominal scenario, failure of engineered barrier system components, and others for disruptive events.

Human intrusion is usually the only disruptive event considered for salt repository models (Freeze et al., 2013), and this is the case for WIPP. Of course, disruptive events will vary with specific salt-repository sites, and could include seismicity. FEPs should be independent but may overlap; however, the consequences of scenarios should be distinct from each other.

1.2.3.2. PA results To effectively use performance assessment to evaluate a disposal system, the three common risk triplet questions must be answered (e.g., as given by Hansen and Leigh, 2011):

• What can happen to the disposal system?

• What are the chances of it happening?

• What are the consequences if it happens?

These should be considered before the steps in the PA process for a salt repository that will contain heat-producing waste.

Hansen and Leigh (2011) go on to state: The phenomena caused by heat from HLW would add some FEPs. Elevated temperature in a salt repository will enhance deformation upon placement of the waste in the rooms. Elevated temperatures and deviatoric stress states near the waste will enhance dry-out and promote encapsulation. Thermally induced salt plasticity is a constant-volume process. As stress equilibrium is re-established by accelerated salt creep, permeability will be eliminated.

The damaged zones around the disposal room release the accessible moisture by flow down the stress gradient and evacuation by the ventilation air.

Room closure will be accelerated by thermal activation of crystal plasticity (flow without damage). If the room is backfilled with crushed salt, the granular material will reconsolidate.

Stresses will drive toward equilibrium, which effectively heals damaged rock, and the waste is expected to be entombed in dry halite. Once the EDZ is healed, the permeability will be similar to that of undisturbed halite.

6

Figure 1. Sandia National Laboratory long-term performance assessment methodology (from Hansen and Leigh, 2011).

These considerations lead to a performance-based set of research goals (the Directed Science Program on Figure 1) that can be followed based on the generic reference salt repository case or PA model, which include mechanical behavior and creep of salt at elevated temperatures (see part 2 of this report).

Most studies of the normal or undisturbed case show that a salt-hosted repository will not leak (e.g., Freeze et al., 2013; Weber et al., 2011) unless extreme conditions are chosen for the reference case (e.g., the study of Lee et al., 2013). As stated above, human intrusion is the only disturbed case usually considered (in US and German work). Other disturbed cases can be FEPped out by site selection. These considerations lead to the expected or nominal (undisturbed) and disturbed scenario cases.

Another set of steps or stages (also called four primary model components, see later) in the geometric (proximal to distal) consideration of how a salt repository would perform is that of:

Waste packages - Near field - Far field - Biosphere (Freeze et al., 2013; Sevougian et al.,

2012, 2013). This represents a sequence in PA from source to distal regions. Results of studies are general because conditions are forced by assumptions for a generic case. A PA is 7

worked through in Freeze et al. (2013) using the above 4 stages or steps. The approach is deterministic in all conditions selected (including a 10 million-year time scale), and the result is very low doses in the biosphere at 5 km from repository, on the order of 10-8 to 10-13 millirem per annum for the undisturbed case at long time scales (> 1 Ma) and as low as > 10-20 millirem per annum at 10,000 years, even with early corrosion and all waste packages failing. This salt baseline scenario assumes an undisturbed transport pathway, but takes only minimal credit for the engineered barrier system; 95 % of waste form degradation occurs in 150,000 years, all waste packages fail instantaneously, and there is no sorption in the near-field salt EDZ between the repository and the underlying interbed. I and Cl isotopes are the only significant releases at long time-scales, due to long half-lives and no sorption.

The enhanced performance of salt (Freeze et al., 2013) is considered to be due to:

- Very low permeability (to water) and very low diffusivity (to radionuclides) in the engineered barrier (crushed salt backfill) system and in the near-field salt EDZ.

- Slower waste form degradation due to reducing chemical conditions.

- Relatively high thermal conductivity and thermally enhanced creep closure and dry-out of the salt.

- Removing the model assumption that all radionuclides diffuse downward to an underlying interbed; the dose would be further halved if upward diffusion were assumed and there was no equivalent overlying interbed.

This last bullet reflects the influence of the WIPP site on the geologic conditions of the safety case. The study also refers back to results for Gorleben, which drew similar conclusions.

1.2.3.3. Other aspect Hansen and Leigh (2011) consider aspects of the mine design, alcoves, etc., and alcove, drift, and shaft seals. The impact of catastrophic situations (political/natural, due to the length of time needed to fill a repository) can be minimized by sealing the emplacement drifts in modular compartments during the course of disposal operations. Thus modules will be far-enough separated to be sealable and closable and will not interfere with each other. This mining technique is also recommended by others from Sandia National Laboratories; the alcove-fill and move-on scenario for filling the repository is supported by thermal calculations (Clayton, 2011).

This would give tens of years before the whole alcove (and hundreds of years for the whole mine/repository) area reaches an elevated temp (say 100 °C) despite the surfaces of disposed WPs being hotter (> 300 °C) before this time. Desirable maximum temperatures of 140-160°C for a whole drift can be achieved by waste package spacings of 12 m (40 ft) and panel spacing of 60 m (200 ft) (Clayton, 2011; Freeze et al., 2013).

1.3 Discussion Various relevant aspects of the group of reports and papers from DOE/Sandia National Laboratory are discussed in the following paragraphs.

8

The Lee et al. (2013) results, mentioned above, seem too extreme to form the basis of a realistic PA. They chose conditions that are, by their own admission, too conservative, e.g.,

WPs are all leaking at start of analysis and no credit is given for engineered barriers, including crushed salt. The study is based on WIPP scenario but for HLW; the results show, in the far field, immediate access of radionuclides to the aquifer above the generic repository, but still the far-field appearance of radionuclides is delayed and modest (below 10-3 millirems per annum for all radionuclides after 7 x 104 years for the undisturbed case).

The generic salt repository reference case (Vaughn et al., 2013) shows what is needed for generic approaches, pre-PA, and highly useful for the salt case (see step 2 on the right side of Figure 2, which provides a FEPs analysis methodology for developing a salt-hosted repository PA model). Important aspects are the waste types and the configuration in the repository, the engineered barriers, including crushed salt backfill (which will consolidate in 200 years to a very low permeability), the seals, and the natural barrier host-rock. Repository operations and thermal management can keep whole repository temperatures below 200 °C. The Vaughn et al.

(2013) case study gives no result but provides a solid scenario on which to base some PA estimates.

Regarding the near field - far field connection, radionuclides are usually modelled being transported from the repository to the biosphere via an aquifer. They have to rise to that level as the aquifer is prescribed to be above repository depth (WIPP-like), thus in the Vaughn et al.,

2013) case the only access is via human intrusion. Details of this are not well worked out (when it occurs, etc.). The undisturbed-scenario FEP case is estimated for 10,000 years in the Vaughn et al. study; if one million years is considered, then there is a need to introduce disruptive cases, such as seismic, as well as human intrusion. However, a case can be made to consider human intrusion (as in the WIPP study) and seismic effects in the 10,000-year time frame.

Papers and reports by the SNL group (e.g., Sevougian et al., 2013; Sevougian et al., 2012; and Freeze and Wolf, 2015, the latter in the DOE 2014 workshop report, DOE, 2015 ) represent the most evolved treatment of FEPs for PA work on a generic salt-hosted repository. These studies present a slightly different approach to the methodology for development of a quantitative safety assessment model (= PA model) for the safety case, summarized on Figure 2, but are based on the reference case of Vaughn et al. (2013).

Steps in the recent Sandia PA models are:

(1) FEPs identification specific to salt host rock, (2) definition of a salt repository reference case, (3) preliminary FEPs screening based on past salt research and development, and safety assessments, (4) specification of quantitative sensitivity analyses and/or reasoned arguments necessary to support FEPs screening, and 9

(5) implications of FEPs screening for PA model construction.

Figure 2. FEPs analysis methodology for developing a salt PA model (from Sevougian et al., 2013)

The abovementioned five steps are shown in Figure 2 on the right, with respect to the rest of the PA process figure, which is basically the same as in Figure 1.

For the PA, four primary model components (see Figure 3) of inventory and source-term, near-field, far-field, and biosphere, are considered (these are the same as the steps mentioned before).

Figure 3. Features and components of the generic salt disposal system (from Sevougian et al., 2013).

10

Reports by these authors identify 208 disposal system FEPs that are potentially relevant to a repository for permanent disposal of SNF and HLW at a generic salt site with the engineered and natural features shown in Figure. 2. Preliminary FEP screening (step 3, right side of Figure

2) examines included, excluded, site-specific, and design-specific FEPs, plus FEPs that need to be evaluated. From the included FEPs, a key step in the development of the generic salt PA model is the identification and evaluation of coupled processes (physical-chemical processes in the included FEPs) important to overall system performance, and how these coupled processes should be represented in the PA model.

1.4 Other considerations These are considerations mentioned in some studies, and considered in some FEPs and scenarios, and which appear to warrant discussion here.

Consideration of brines in the host rock of salt-based repositories is important for canister corrosion. If there is more salt-bound water than envisaged and more brine pockets, there could be migration upward over time or migration towards the hot waste. German workers saw contact with brines as the only way radionuclides could leave the model repository (Weber et al., 2011), but a very low probability for such an occurrence is stressed. Moreover, after repository closure as salt creep closes a disposal room and the stress gradient decreases, preexisting fractures in the excavation disturbed zone (EDZ) and crushed salt backfill will reconsolidate. Thus, conditions in a repository would evolve to significantly limit brine flow to the waste disposal areas. The elevated temperatures from high-temperature waste on the processes occurring in a salt repository are not expected to negatively impact the capability of the salt natural system and may result in improved capabilities (more rapid consolidation and healing of damaged salt) (Freeze et al., 2013). Modelling of such processes has been conducted by several groups (e.g., Rutqvist et al., 2014; Ofoegbu and Dasgupta, 2017, and references therein).

A significant aspect in the near-field in bedded salt is that brines dry out at 107 °C at atmospheric pressure, thus a dry-out zone around hot waste could develop. If it is possible to keep the repository temperature at > 107 °C for a while, then this could help slow down corrosion (with appropriate waste package design to withstand the heating). The dry state, and possible low water content, of rock salt implies that the role of gases in deep geologic salt repositories is minimal (e.g., NEA, 2015).

Heat production from waste packages can be limited by loading, or by design, or by thermal management in repository. Several reports discuss thermal management by the design of excavated rooms and alcoves. The packing distance (separation between waste packages) is important; some calculations show how heat is dissipated through salt around canisters (from crushed salt, into the host rock). Some also estimate the likely whole repository heating scheme (e.g., keeping repository drifts to 140-160 oC maximum by optimum spacing of alcoves for waste packages and panels or drifts, e.g., Clayton et al., 2011).

11

In situ heat tests on rock salt are needed (e.g., Robinson et al., 2013; Hansen and Leigh, 2011).

It was concluded that most of the simple models do not adequately simulate the closure of an excavation due to salt creep or the consolidation of crushed salt (see next section, 2). The estimated creep-induced closure rate of an excavation in salt could be underestimated, especially under thermal loads (Ghosh and Hsiung, 2013). In addition, studies have shown that the estimated time required for crushed salt to seal an emplaced waste package may be substantially different, and underestimated, from that observed. Papers presented at the 3rd U.S./German Workshop on Salt Repository Research, Design and Operations, held in Albuquerque on October 8-10, 2012, support this conclusion (DOE, 2013; Ghosh and Self, 2013).

The behavior of salt during seismic events has been considered as a disruptive event (has been FEPped in - e.g., see studies by Freeze and Wolf, 2015 or Freeze et al., 2015). No specific results are available, but salt deforms plastically and should absorb energy and be self-healing.

Some recent publications concerning salt Jordan et al. (2015) examine experiments conducted at Los Alamos National Labs on salt containing hydrous minerals, and look at the effect of heat-generating nuclear waste on the mobility of water vapor as the minerals dehydrate. The effects are limited to hydrous mineral contents (up to 10 volume %), higher than those seen in most target salt-host horizons.

Movement of moisture away from, and later towards, waste packages is theoretically possible, according to this study.

A late 2015 paper in Science (Ghanbarzadeh et al., 2015) caused some media attention at the time. The results were based on experiments on synthetic salt and examination of salt samples from oil wells, which included deformation-assisted percolation of fluids. The authors warned that the design of nuclear waste repositories in salt should guard against deformation-driven fluid percolation, because, in general, static percolation thresholds may not always limit fluid flow in deforming environments. However, the depths from which the salt samples used in the study came are much deeper (> 1 km) than depths being considered in HLW salt-repository models (500-700 m). The samples also came from domal salt, which is not currently being considered in the US as a repository host option. How the results of this study apply to the generic bedded salt repository reference models, discussed above, is not clear.

1.5 Summary Probability weighted predictions of dose relevant to HLW/SNF repository models do not occur in the papers/reports summarized herein; those that occur are based on deterministic arguments.

Probability-based estimates must wait for the site-specific cases to arise, as too many of the vital parameters cannot be defined in the generic cases. Moreover, if human intrusion is the only path for radionuclide release, many organizations would state that probability-based 12

estimates would not be possible due to difficulties in predicting future human actions over such timescales. However, the time, location, depth, and what may happen at the repository horizon can all be treated stochastically, either in an epistemic (e.g., waste package and waste characteristics) and aleatory (e.g., time, location) manner. Future regulations may prescribe the time, but stochastic treatment will remain. To quote the findings of IAEA project HIDRA (Human Intrusion in the context of Disposal of Radioactive Waste): In some cases, probabilities have been considered and the approaches have been accepted by the regulators. However, it is recognized that estimates of the probability of intrusion are uncertain. It is generally recommended in ICRP Publications 60 and 122 that safety assessment should seek to evaluate the doses associated with human intrusion that may occur, but should not attempt to use a risk based concept that uses as a basis for assessing the product of the probability of intrusion and the dose arising from the intrusion. (From IAEA, 2014.)

2. CNWRA work on creep and deformation of rock salt, and other work In Fiscal Year 2012, CNWRA conducted a literature survey of constitutive models to simulate the creep behavior of intact and crushed salt under thermomechanical (TM) loads. It was concluded that most of the simple models do not adequately simulate the closure of an excavation due to salt creep or the consolidation of crushed salt. This work (Ghosh and Hsiung, 2013) followed an earlier CNWRA report (Winterle et al., 2012) which dealt with barrier functions, salt geology, and the geomechanics of salt deformation, including deformation around waste packages.

The 2013 report included the creep of salt (both intact and crushed salt as backfill) interpreted as excavated room closure rates. Various recommendations were made including suggestions that the closure rate in rooms excavated in salt under various TM loads be investigated. A similar suggestion was made for crushed salt, and salt crack formation and propagation.

The above work was followed by an on-going CNWRA investigation describing an implementation of a model in FLAC (a computer software code, Itasca Consulting Group, 2011) for the mechanical behavior of salt rock (Ofoegbu and Dasgupta, 2015). The report describes the model implementation, use of the implementation to conduct simulated triaxial compression creep testing of WIPP salt to determine the material creep parameters for the salt, and numerical modeling of two in-situ experimental rooms (one heated and one at ambient temperature) at the WIPP site (DOE, 2012) using the implementation.

Results show that the measured convergence of the unheated room from the field test continued to increase throughout the 1500-day test duration whereas the calculated (modelled) convergence suggests the room should have attained a stable configuration after a much shorter period. The difference perhaps arises because creep is the only inelastic deformation mechanism included in the model, whereas the measured convergence indicates that inelastic deformation due to material failure also likely occurred during the field test. Moreover, the calculated temperatures for the heated room are higher by about 20 °C than the in situ 13

measurements at comparable times. Part of the difference can be attributed to greater effective heating in the two-dimensional model compared to the heating in the experiment. The Centers modeling reveals a discrepancy between the measured convergence of the unheated room from the field test and the calculated (modelled) convergence at long time duration A more recent modeling effort based on coupled creep and plasticity resulted in improvements in the calculated convergence relative to the creep-only model (Ofoegbu and Dasgupta, 2017).

Compared to creep-only, the new model showed improvements in the shape of the calculated convergence history and the magnitude of calculated convergence improved to approximately 75 percent of the measured room convergence approximately 1,200 days after excavation. The authors also investigated contributions to convergence due to rock creep from small deviatoric stress conditions. Although increased creep rates from small deviatoric stress conditions resulted in increased convergence, the shape of the calculated convergence history departed from the shape of the measured history enough to indicate that the calculated effect is not consistent with rock behavior around the openings.

Deformation of clay seams or clayey interfaces between salt rock layers could contribute to overall rock behavior around the openings. Although this is a potential improvement, interface parameters of the clay seams need to be calibrated in order to obtain the measured convergence. In the new approach, potential creep in the clay seams is not included.

Regarding thermal effects, the calculated convergence history for heated Room B shows an increase in convergence at the start of thermal loading, consistent with the measured convergence history. However, the calculated magnitude of convergence is still smaller than the measured magnitude. The authors expect that the accuracy of the calculation can be improved by increasing the strength of coupling of the two deformation mechanisms by representing some of the controlling parameters as functions of distortional strain. In the model used (Ofoegbu and Dasgupta, 2017), creep and plastic deformations are coupled, somewhat weakly, through deviatoric stress. However, additional thermal effects could result from thermal strain affecting mechanical behavior, and these need to be taken into account.

While the above work on deformation of salt is useful and is showing some not-unexpected differences between models and in situ field measurements, it is at present difficult to see how it will inform any PA models for a generic salt repository, as the reference cases are generic. It may help us understand the deformational closing of rooms and alcoves once they are loaded with waste packages. The main issue is that the type, and heat-production, of waste packages to be disposed is undecided (DOE, 2014a), thus it is difficult to start realistic, detailed modelling due to the dependence of deformation rates on heating rates of the salt. One must be mindful though that heat generation is the early phase, and PA usually deals with long term performance. If we assume that high temperature in general results in faster room closure and seal of drifts, then this discussion is more relevant to the operational phase than the post-closure phase.

14

In a series of papers, other work on TM effects of high temperature waste on salt have been investigated with FLAC (and TOUGH-FLAC, a computer software code, see Rutqvist et al.,

2014) by a group at Lawrence Berkeley National Laboratories. Rutqvist et al. (2014) report on THM modeling and field test planning activities in support of disposal of heat-generating waste in salt. Overall, the generic repository simulation results suggest that the excavation disturbed zone (EDZ) around an emplacement tunnel is healed within the first few years and that the backfill reconsolidates within the first two decades. This occurs along with the temperature mediated salt creep of the host rock; greater compaction occurs in the roof and floor areas, and lesser compaction in the sidewalls, and the backfill reconsolidation process takes between 6 and 20 years to complete. These results suggest that a field test involving a heater in a salt tunnel should last at least 10 years to realize meaningful results for reconsolidation processes.

In Blanco-Martin et al. (2015a), the TOUGH-FLAC simulator - (mechanical-flow) - is applied to heat-producing waste in a generic salt repository with crushed salt backfill (drift in-fill type at 600 m depth). Results show that when considering THMC processes, allowing for halite dissolution and precipitation, then a) the reconsolidation time and final porosity of crushed salt is not much affected, b) for a while (10-200 year), the effects of dissolution / precipitation on the salt and crushed salt is to increase the porosity and thus permeability slightly over distances of 2-3 m around the drifts. Temperatures are < 160 °C between 10 and 200 years after closure of the repository, and only 80 °C 25 m from the repository margin.

Blanco Martin et al. (2015b) investigated the capabilities of two simulators, TOUGH-FLAC (US) and FLAC-TOUGH (German), to predict the long-term thermal-hydraulic-mechanical (THM) response of a generic salt repository for heat-generating nuclear waste. The simulators are based on sequential coupling and include state-of-the-art knowledge for salt. Their main difference is the sequential method used. The paper presents a benchmark between Lawrence Berkley National Lab (LBNL) and Clausthal University of Technology (TU Clausthal). The scenario studied assumes heat and gas generation from the waste packages, and crushed salt backfill. The comparison of results is very satisfactory, providing increased reliability and confidence in the capabilities of the simulators to evaluate the geological and engineered barriers in the long-term, with modelling extending to 100,000 years.

3. Adaptions to SOAR In the generic salt reference case, in the literature, the nominal (expected) case is successful in keeping radionuclides within the salt for the long term. Radionuclides do not leak to the biosphere without exceptional conditions and/or a disruptive event.

SOAR is the Scoping of Options and Analyzing Risk (SOAR) model (Markley et al., 2011). The buffer in SOAR is presently bentonite. This will need to be replaced by crushed salt, which is used as backfill in all studies of modelled salt repositories. In the long term, we can assume the 15

backfill has been reconsolidated and nearfield properties have returned to the state of the host rock.

There are a few aspects in SOAR that either can be readily changed or may need to be structurally modified to accommodate a generic salt repository performance assessment.

Below, we focus on three areas that require most attention: (1) host rock, (2) waste and waste packages, and other engineered barriers, and (3) release scenarios, including human intrusion.

Host Rock Consolidated salt rocks are known to possess low permeability, but moisture in salt rocks may also move along pressure and thermal gradients. The latter is particularly important for heat-generating SNF and HLW and requires the conceptualization of the engineered barrier system (EBS) and near field behaviors during early postclosure periods. Thermohydrology and geomechanics technical bases including the associated model parameters are required for SOAR near field modeling to address the processes and likelihood that moisture, either in isolated brine pockets or reservoirs, may come into contact with waste packages and radioactive wastes.

Previous performance assessment for both WIPP and generic salt rock repositories indicated that human intrusion or drilling into the repository horizon may create two types of potential flow paths. The first one is a flow path through the borehole and into drinking water aquifers overlying the repository. The second type of flow path is direct brine release through the borehole to the ground surface (e.g., Clayton et al., 2010). In the present SOAR conceptualization, radionuclides can only be released into the geological formation connected directly to the EBS and the near field. SOAR will have to be structurally modified to accommodate the other two types of potential flow paths in order to adequately model human intrusion.

Previous salt rock repository performance assessment also considered impacts of mining activities to hydrologic properties of repository geological formations and to off-site radionuclides transport (e.g., Clayton et al., 2010). Currently SOAR assumes that flow fields, both near and far fields, are in steady state throughout the lifetime of the repository. Reported PA results, e.g.,

in Clayton et al. (2010), suggest that radionuclide moving off-site through a flow path interconnected by mining activities are not significant for a transuranic nuclear waste repository.

Should future risk insights suggest otherwise for SNF and heat-generating HLW, modification to allow transient flow path properties in SOAR is necessary to evaluate long-term mining activity impacts to repository safety.

Additional model parameters, such as coefficients of sorption on host rocks and on waste and waste form colloids, and solubility of SNF and HLW radionuclides in brines that may depend on other brine properties (e.g., temperature and pH) will also need to be entered into a salt rock flow path model in SOAR and/or the SOAR model parameter database.

16

Waste Inventory, Waste Package and the EBS Gas may be generated in the near field of geological repositories due to microbial activities or waste package material corrosion. For transuranic nuclear waste repository, the conditioning of waste by organic matters is particularly of interest to preclosure and postclosure repository safety. Currently, gas generation and the migration of radionuclides with gaseous species are not considered in SOAR. The fundamental difference between transuranic waste streams and SNF/HLW is the key radionuclides and their abundance in the waste inventory. SOAR, primarily designed for generic SNF/HLW repositories, has the capability of modeling repositories of differing waste inventories and waste packages (including copper, stainless steel, carbon steel, and titanium). The dissolution rates of SNF/HLW and waste package material corrosion modes and rates in brines, however, will need to be added to the existing waste package models and/or updated in the SOAR model parameter database.

Crushed rock salt is usually used as backfilling materials and concrete was designated as the drift sealing material for the operating WIPP transuranic nuclear waste repository. Additionally, magnesium oxide was placed on top of WIPP contact-handled waste packages to absorb gas generated through microbial decomposition of organic matters in the waste stream. Crushed rock salt will be reconsolidated into salt rock in approximately 100 years after the closure of repository, with its mechanical and hydrologic properties comparable to those of the pristine salt host rock. Modifications to SOAR buffer materials and their properties are thus minimal.

Release Scenario It is generally accepted that human intrusion is the dominant release scenario for salt rock nuclear waste repositories. Radionuclide releases through a borehole penetrating waste packages can be associated with well bore cutting materials (solid) or with brines directly upwelled from the repository horizon, in additional to releases into drinking water aquifers overlying and connected through the wellbore to the repository. This scenario calls for a separate solid and fluid release path in addition to the multi-layer flow path conceptualized in SOAR.

The WIPP performance assessment also includes a hypothetical brine reservoir next to the repository footprint, which may supply the repository sufficient brine to cause direct brine release to the ground surface if penetrated by a borehole. For salt rock repository modeling in SOAR, the brine flow rates into and out of the repository will need to be properly calculated with offline analysis to provide the necessary technical basis.

Human intrusion can be modeled as a one-time disruptive event in the SOAR model. As human intrusion is the only disruptive event normally considered (see part 1 of this report), a drillhole through an overlying aquifer and access for radionuclides to the aquifer from the area of the repository will be required. The drillhole could penetrate both the aquifer and repository horizon.

The release of radionuclides resulting from human intrusion may need to be implemented separately. If human intrusion is the only reasonable failure mode, the PA could assume that 17

since the salt has been by-passed via intrusion, transport calculations could start from the salts edge. Alternatively, for a salt repository which might be entirely within the SOAR near-field domain, the near-field may have to be divided into two fields (the very near and the near): the former would include the damaged salt (EDZ) and the latter would consist of intact salt (if we use salt state to define domains), or hot temperatures and ambient temperatures (if we use temperature to define domains).

Further, when (after permanent closure) the disruptive event happens (pre- or post-deformation of salt; or corrosion of waste packages) controls a lot of issues. Another issue is how many waste packages might be intersected by a drill hole (it will depend on alcove design) and how much of the waste stored in the repository might be accessible to this intrusion/disruption. Also, how many waste packages are corroded and the state of waste packages regarding salt deformation at the time of the disruptive event need to be considered. All of these and other specific issues related to abstraction of the EBS FEPs and scenarios in a salt rock repository will have to be addressed in the SOAR model if application of SOAR to salt rock repository is foreseeable.

4. References Blanco Martín, L., Rutqvist, J., Birkholzer, J.T., and Battistelli, A., 2015a. Long-term modeling of coupled processes in a generic salt repository for heat-generating nuclear waste: preliminary analysis of the impacts of halite dissolution and precipitation. ARMA 15-440, American Rock Mechanics Association, 2015.

Blanco Martín, L., Wolters, R., Rutqvist, J., Lux, K-H., and Birkholzer, J.T., 2015b. Comparison of two simulators to thermal-hydraulic-mechanical processes related to nuclear waste isolation in saliferous formations. Computers and Geotechnics 66, p 219-229.

BRC, 2012. Report to the Secretary of Energy. Washington, DC: Blue Ribbon Commission on Americas Nuclear Future. 2012.

Camphouse C, 2014. Reflections on the ISIBEL project and perspectives on modeling salt reconsolidation. In, DOE, 2015. Proceedings of the 5th US/German Workshop on salt repository research, design, and operation. SAND2015-0500R. DOE Used Fuel Disposition Campaign; Sandia National Laboratories (Jan., 2015). (Workshop 9/2014.)

Clayton, D.J., 2011. Thermal performance of a generic high-level waste salt repository.

International High Level Radioactive Waste Management Conference (IHLRWMC) 2011, Albuquerque, NM, April 10-14, 201, pg. 977-983.

18

Clayton, D. J., R. C. Camphouse, J. W. Garner, A. E. Ismail, T. B. Kirchner, K. L. Kuhlman, and M. B. Nemer, 2010. Summary Report of the CRA-2009 Performance Assessment Baseline Calculation, Sandia National Laboratories, Waste Isolation Pilot Plant.

Clayton, D.J. and Gable, C.W., 2009. 3-D Thermal Analyses of High Level Waste Emplaced in a Generic Salt Repository; Sandia National Laboratories Report, SAND2009-0633P.

DOE, 1996. Title 40 CFR Part 191 Compliance Certification Application for the Waste Isolation Pilot Plant. DOE/CAO-1996-2184.

DOE, 2012. Revisiting Historic Numerical Analyses of the Waste Isolation Pilot Plant (WIPP)

Room B and D in-situ Experiments Regarding Thermal and Structural Response. SAND2012-7525, 177 p. (Sept. 2012). [By Rath J and Arguello JG.]

DOE, 2013. Proceedings of the 3rd US/German Workshop on Salt Repository Research, Design, and Operation. SAND2013-1231P (2/14/2013).

DOE, 2014a. Assessment of Disposal Options for DOE-Managed High-Level Radioactive Waste and Spent Nuclear Fuel. Department of Energy, October 2014, 31 p.

DOE, 2014b. Evaluation of Options for Permanent Geologic Disposal of Spent Nuclear Fuel and High-Level Radioactive Waste in Support of a Comprehensive National Nuclear Fuel Cycle Strategy. Volume I. DOE Used Fuel Disposition Campaign; Sandia National Laboratories, SAND2014-0187P, 91 p. (April 2014.)

DOE, 2015. Proceedings of the 5th US/German Workshop on salt repository research, design, and operation. SAND2015-0500R. DOE Used Fuel Disposition Campaign; Sandia National Laboratories (Jan., 2015). (Workshop 9/2014.)

Freeze, G., Voegele, M., Vaughn, P., Prouty, J., Nutt, W.M., Hardin, E., and Sevougian, S.D.,

2013. Generic Deep Geologic Disposal Safety Case. DOE Used Fuel Disposition Campaign; Sandia National Laboratories and Argonne National Laboratory; SAND2013-0974P, 356 p.

(August 2013).

Freeze, G., and Wolf, J., 2014. FEP development activities. In DOE, 2015, Proceedings of the 5th US/German Workshop on salt repository research, design, and operation. SAND2015-0500R. DOE Used Fuel Disposition Campaign; Sandia National Laboratories (Jan., 2015).

(Workshop 9/2014.)

Freeze, G., Sevougian, S.D., Leigh, C., Gross, M., Wolf, J., Monig, J., and Buhmann, D., 2015.

Development of a salt repository FEP catalogue. SAND2014-2423P. Sandia National Laboratories, Albuquerque, NM.

19

Ghanbarzadeh, S., Hesse, M.A., Prodanovi, M., and Gardner, J.E., 2015. Deformation-assisted fluid percolation in rock salt. Science 350, Issue 6264, p. 1069-1072.

Ghosh, A. and Hsiung, S.-M., 2013. Progress ReportEvaluation of Salt Under Thermomechanical Loading as a Host Repository Medium. Washington, DC: U.S. Nuclear Regulatory Commission. 2013. [Needs ML#.]

Ghosh, A. and S. Self. 3rd U.S./German Workshop on Salt Repository Research, Design and Operations. Sandia National Laboratories. October 9-11, 2012. Albuquerque, New Mexico:

Washington, DC: U.S. Nuclear Regulatory Commission. 2013. [Needs ML#.]

GRS, 2014. Studies Relating to Human Intrusion into a Repository - Preliminary Safety Case of the Gorleben Site (VSG). GRS-348, 124 p. Gesellschaft für Anlagen-und Reaktorsicherheit (GRS) mbH.

Hansen, F.D., 2016. A comparison of bedded and domal salt regarding heat-generating nuclear waste disposal. Abstract; 7th US/German Workshop on Salt Repository Research, Design and Operation, Washington, DC. September 7-9, 2016.

Hansen, F.D. and Leigh, C.D., 2011. Salt Disposal of Heat-Generating Nuclear Waste.

SAND2011-0161. January 2011.

Helton, J.C. et al., 1998. Uncertainty and Sensitivity Analysis Results Obtained in the 1996 Performance Assessment for the Waste Isolation Pilot Plant. SAND98-0365. Sandia National Laboratories, Albuquerque, NM. September 1998.

IAEA, 2014. From http://www-ns.iaea.org/projects/hidra/ - checked May 2016.

Itasca Consulting Group, Inc. FLAC Fast Lagrangian Analysis of Continua. Version 7.

Minneapolis, Minnesota: Itasca Consulting Group, Inc. 2011.

Jordan, A.B., Boukhalfa, H., Caporuscio, F.A., Robinson, B.A., and Stauffer, P.H., 2015.

Hydrous Mineral Dehydration around Heat-Generating Nuclear Waste in Bedded Salt Formations. Environ. Sci. Technol. 49, p. 67836790.

Lee, J.H., Siegel, M., Jove-Colon, C., and Wang, Y., 2011. A performance assessment model for generic repository in salt formation. IHLRWMC 2011, Albuquerque, NM, April 10-14, 2011, pg. 1-8.

Markley, C., Pensado, O., Gwo, J.-P., Winterle, J., Ahn, T., Benke, R., Cao, T., González, H.,

Gray, A., He, X., Janetzke, R., Jung, H., Oberson, G., Shukla, P., Sippel, T., Stothoff, S., and Tipton, L., 2011. SOAR: A Model for Scoping of Options and Analyzing Risk, Version 1.0, User Guide. U.S. Nuclear Regulatory Commission. ML113060286.

20

NAS (National Academy of Sciences), 1957. The Disposal of Radioactive Waste on Land.

Report of the Committee on Waste Disposal of the Division of Earth Sciences. Pub. 519.

Washington, DC: National Research Council.

NEA (Nuclear Energy Agency), 2015. Relevance of gases in the post-closure safety case for radioactive waste management. Radioactive Waste Management, NEA/RWM/IGSC(2015)1/REV1, March 2015.

NRC (National Research Council), 1996. The Waste Isolation Pilot Plant: A Potential Solution for the Disposal of Transuranic Waste; National Research Council: Washington, DC, 1996; http://www.nap.edu/openbook.php?isbn=0309054915.

Ofoegbu G and Dasgupta B, 2015 (2016). Implementation of a creep model in FLAC to study the thermomechanical response of salt as a host repository medium - progress report. San Antonio, Texas: Center for Nuclear Waste Regulatory Analyses, July 2015. ML16174A165 Ofoegbu G and Dasgupta B, 2017. Implementation of a creep model in FLAC to study the thermomechanical response of salt as a host repository medium - 2nd progress report. San Antonio, Texas: Center for Nuclear Waste Regulatory Analyses, (Mar. 2017). ML17074A119 Pepping, R.E., Chu, M.S.Y., Wahi, K.K., and Ortiz, N.R., 1983. Risk analysis methodology for spent fuel repositories in bedded salt: final report. NUREG/CR-2402 (SAND81-2409 GF).

Washington, DC: U.S. Nuclear Regulatory Commission.

Robinson, B.A., Nelson, R.A., and Weaver, D.J., 2013. Disposal of heat-generating nuclear waste in salt - recent activities and future plans. IHLRWMC 2013, Albuquerque, NM, April 28-May 2, 2013, pg. 793-800.

Rutqvist, J., Blanco Martín, L., Mukhopadhyay, S., Houseworth, J., and Birkholzer, J.T., 2014.

Modeling and Field Test Planning Activities in Support of Disposal of Heat-Generating Waste in Salt. DOE Used Fuel Disposition Campaign; Lawrence Berkeley National Laboratory; FCRD-UFD-2014-000622, 14 p. (Sept 2014.)

Sevougian, S.D., Freeze, G.A., Gross, M.B., Hardin, E.L., Lee, J., Leigh, C.D., MacKinnon, R.J.,

Mariner, P., and Vaughn, P., 2012. TSPA Model Development and Sensitivity Analysis of Processes Affecting Performance of a Salt Repository for Disposal of Heat-Generating Nuclear Waste. FCRD-UFD-2012-000320, Rev. 0, U.S. Department of Energy, Office of Nuclear Energy, Office of Used Nuclear Fuel Disposition, Washington, D.C., September 28, 2012.

Sevougian, S.D., Freeze, G.A., Gross, M.B., Hardin, E.L., Lee, J., Leigh, C.D., MacKinnon, R.J.,

Mariner, P., and Vaughn, P., 2013. Performance assessment model development for a bedded salt repository. IHLRWMC 2013, Albuquerque, NM, April 28-May 2, 2013, pg. 465-472.

21

Vaughn, P., Sevougian, S.D., Hardin, E.L., Mariner, P., and Gross, M.B., 2013. Reference case for generic disposal of HLW and SNF in bedded salt. IHLRWMC 2013, Albuquerque, NM, April 28-May 2, 2013, pg. 443-450.

Weber, J.R., Keller, S., Mrugalla, S., Wolf, J., Buhmann, D., Mnig, J., Bollingerfehr, W., Krone, J., and Lommerzheim, A., 2011. Safety strategy and assessment for a German HLW-repository in salt. IHLRWMC 2011, Albuquerque, NM, April 10-14, 201, pg. 924-931.

Winterle, J., Ofoegbu, G., Pabalan, R., Manepally, C., Mintz, T., Pearcy, E., Smart, K., McMurry, J., Pauline, R., and R. Fedors, R., 2012. Geological Disposal of High-Level Radioactive Waste in Salt Formations. ML12068A057. Washington, DC: U.S. Nuclear Regulatory Commission.

2012.

WIPP Energy, 2016.

Accessed, Sept 2016.

22