ML19248C647

Enclosure 4 - Staff Report Documenting Technical Basis for the Alternative FHA Model

ML19248C647
Person / Time
Issue date:	12/04/2019
From:	Michael Salay Office of Nuclear Regulatory Research
To:
Dickson E
Shared Package
ML19114A117	List: ML19248C647 ML19248C668 ML19248C704 ML19270E335
References
Download: ML19248C647 (76)

Text

Enclosure 4 Re-evaluation of the Fission Product Release and Transport for the Design-Basis Accident Fuel Handling Accident Elijah Dickson Michael Salay

TABLE OF CONTENTS

1. ABSTRACT......................................................................................................................... 4

2. INTRODUCTION ................................................................................................................ 5

3. EVALUATION ..................................................................................................................... 8 3.1 Evaluation history ........................................................................................................ 8 3.2 Re-evaluation of WCAP-7518-L (non-proprietary version is WCAP-7828) ................... 9 Small-Scale Tests and Results ............................................................................11 Full-Scale Tests and Results ...............................................................................15 3.3 Re-evaluation of AEC Mass Transfer Method .............................................................19 Gas Bubble Parameters (size, db, and rise time, t) ...............................................20 Iodine Partitioning Factor .....................................................................................22 Iodine Speciation .................................................................................................24 Iodine Mass Transfer Coefficient Method .............................................................25 AEC Method to Calculate Iodine Decontamination Factors ..................................26 3.4 Analysis Computing Flexible Iodine Decontamination Factor ......................................28 Iodine DF Models .................................................................................................28 Iodine Mass Transfer Coefficient..........................................................................28 Gas Bubble Size Model (d) ..................................................................................29 Gas Bubble Rise Time Model (t) ..........................................................................30 Parametric Bootstrap Sampling and Mean Parameter Values..............................30 Results of WCAP and Burley Iodine DF Model Re-analysis as a function of Release Pressure ............................................................................................................................34 Re-Analysis Observations ....................................................................................36 Conclusion of Re-analysis Results .......................................................................42

4. ALTERNATIVE DESIGN-BASIS ACCIDENT FUEL HANDLING ACCIDENT MODEL........44 4.1 Summary ....................................................................................................................44 4.2 Behavior of Iodine under FHA Conditions ...................................................................44 4.3 Iodine Re-Evolution calculations .................................................................................47 Assumptions ........................................................................................................47 Evaluation of iodine volatile fraction .....................................................................49 Evaluation of the Pool Mass Transfer Coefficient .................................................55 Calculation with concentration-dependent volatile fraction ...................................57 Simplified Calculation with concentration-dependent speciation ..........................59 Evolution model limitations...................................................................................66 2

4.4 General Description of the Alternative DBA FHA Method ............................................67 4.5 RadTrad Models .........................................................................................................70 Pilot Plant Control Room ......................................................................................71 RadTrad Notes and Input Parameters .................................................................71 4.6 Results........................................................................................................................73 4.7 Recommendation and Guidance .................................................................................74

5. REFERENCES ..................................................................................................................75 3

1. ABSTRACT This study re-evaluates the design-basis accident (DBA) fuel handling accident (FHA) described in Regulatory Guide 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors, (USNRC, 2000), and Regulatory Guide 1.195, Methods and Assumptions for Evaluating Radiological Consequences of Design Basis Accidents at Light-Water Nuclear Power Reactors, (USNRC, 2003). FHAs are analyzed to assess the risk to public health and safety resulting from the operation of the facility and to demonstrate compliance with the various numerical radiological criteria set forth in regulation and subsequent guidance. The primary purpose is to evaluate the design basis of ventilation and penetration closure times and filter efficiencies to mitigate releases to the environment. As of late, the FHA has become one of the most common DBA radiological dose licensing actions, can require significance staff resource, and in some cases due to currently acceptable staff modeling assumptions described in Regulatory Guide 1.183, has become a limiting accident despite its low safety and risk significance.

This study revisits the original studies using modern data analysis tools to confirm results and conclusions. Errors in computation were corrected, and certain parameters were updated to be consistent with current NRC staff assumptions and practices with other DBAs. This study confirms and elaborates on the available experimental data, various reports, and staff reviews while recognizing identified limitations. This study concludes there is considerable design margin regarding the scrubbing effects of iodine in the spent fuel pool and that the current staff DBA FHA model should be updated to reflect an updated understanding of iodine behavior. Historically, iodine released from the fuel pin gap has been primarily considered to be in the form of gaseous iodine as I2 and methyl iodide and released to the environment instantaneously. In reality, most of the iodine in the gap is likely a solid as CsI at the time of the postulated FHA and therefore not available for instantaneous release. Rather, it is readily absorbed in the pool water and slowly re-evolved over a long period of time. Subsequent research has shown that re-evolution can potentially be significant. Should a spent fuel rod break under water it would be expected that the soluble contents in the gap be released to the pool. Therefore, a new FHA is proposed based on the environmental conditions in which fuel handling operations are taking place.

The proposed model makes several improvements based on the analyses described and incorporates our current understanding of reactor fuel pin physics and iodine chemistry under FHA conditions. Under these conditions, a time period is considered between power operation and the movement of recently irradiate fuel to account for both radioactive decay, less decay power, and the use of pool water temperature to determine internal gas temperature and chemical form of iodine. This time period is controlled by the facilities Technical Specifications. The fraction of gap activity in the damages rods is assumed to be released in two stages. The first stage is the instantaneous gaseous release from the fuel gap in rising bubbles where I2 and organic iodine are conservatively assumed to be in vapor form and subsequently decontaminated by passage through the overlying pool of water into the building atmosphere. This activity is then vented to the environment over a 2-hour period. The second stage is the protracted release initiated 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> (following the initial gaseous release) following the fuel bundle drop. The CsI in the fuel gap of the damaged assembly is conservatively assumed to completely dissociate into 4

the pool water then slowly re-evolve into the building atmosphere as I2 due to the low pool water pH. This activity is vented directly to the environment for a period of thirty days.

A case study was performed to analyze the radiological consequences of the proposed DBA FHA using the alternative analysis methodology. The purpose was to determine the impact of the proposed alternative methodology for the DBA FHA by comparing the computed radiological consequences to the licensing basis FHA analyses and whether the revised results would exceed the radiological accident dose criteria of 10 CFR 50.67 and the FHA-specific dose acceptance criteria listed in Regulatory Guide 1.183 and Standard Review Plant Chapter 15.0.1. A survey of operating plant Updated Final Safety Analysis Reports and recent Alternative Source Terms license amendments was conducted to review the various facility licensing- and design-bases and identify important modeling parameters. Models were developed based on the updated FHA transport model using the NRC maintained Symbolic Nuclear Analysis Package/RADionuclide Transport, Removal And Dose Estimation (SNAP/RADTRAD) dose analysis computer code.1 The code is used to estimate transport and removal of radionuclides and determine DBA radiological doses at the exclusion area boundary, the low population zone, in the control room, and other locations of interest. Estimated doses at all three receptors are 91-98% lower than the licensees current licensing bases. The impact of credit for various pool cooling and cleanup systems was not explicitly modeled since such systems were not credited in the surveyed licensees DBA FHA licensing basis analyses of record.

It is recommended that the NRC staff consider the new DBA FHA model. It is anticipated the new iodine DF model would maintain defense-in-depth, increase operational flexibilities, reduce over conservatisms and decrease the staffs limited time and resources on licensing actions concerning the FHA.

2. INTRODUCTION Use of regulatory source terms in design basis accident assessment is deeply embedded in the regulatory policy and practices of the NRC, even as the licensing process has evolved over the past 50 years. The source term refers to the magnitude and mix of the radionuclides released from the fuel, expressed as fractions of the fission product inventory in the fuel, as well as their physical and chemical form, and the timing of their release. It is based upon the concept of defense-in-depth in which power plant design, operation, siting, and emergency planning comprise independent layers of nuclear safety. This approach encourages nuclear plant designers to incorporate several lines of defense in order to maintain the effectiveness of physical barriers between radiation sources and materials from workers, members of the public and environment in operational states and, for some barriers, in accident conditions. It centers on the concept of DBAs, assessment of which aims to determine the effectiveness of each line of defense. The DBAs establish and confirm the design basis of the nuclear facility, including its safety-related structures, systems and components, and items important to safety; ensuring that the plant design meets the safety and numerical radiological criteria set forth in regulation and 1 ADAMS Accession Number ML16160AA019 5

subsequent guidance. From this foundation, specific safety requirements have evolved through a number of criteria, procedures and evaluations, as reflected in regulations, Regulatory Guides, standard review plans, technical specifications, license conditions, and various regulatory technical information documents.

FHAs are analyzed to assess the risk to public health and safety resulting from the operation of the facility and to demonstrate compliance with various regulatory requirements. The primary purpose is to evaluate the design basis of systems, structures and components that mitigate radiological releases to the environment, to include such items as ventilation system design, filter efficiencies, and primary or secondary containment penetration closure times, for example. An illustrative accident sequence consists of the dropping of a nuclear fuel assembly during refueling operations, resulting in the non-mechanistic assumptions of complete and instantaneous shearing off of all fuel pins on the dropped assembly, release of a portion of the volatile fission gases from the damaged fuel rods, transport of soluble and insoluble gases through the water of the spent fuel pool, absorption of soluble gases in the pool water, and release to and transport through the environment.

The technical basis for the current FHA fission product transport model is largely contained in two studies dating back to the 1970s. The first is a proprietary topical report by Westinghouse, WCAP-7518-L, (1970), reporting on a series of large- and small scale experiments to measure the iodine scrubbing effect. The non-proprietary version of WCAP-7518-L is WCAP-7828 (Westinghouse, 1971), Radiological Consequences of a Fuel Handling Accident. The primary purpose was to measure the iodine scrubbing effect from gas bubbles to the water, commonly referred to as the iodine decontaminant factor (DF). The report recommended a generic DF of 500 which covered about 90% of the data over a range of fuel pin bundle pressures. Please note that both WCAP-7518-L (proprietary) and WCAP-7828 (non-proprietary) are hereinafter referred to as the WCAP, unless specifically identified. At that time, the Atomic Energy Commission (AEC) staff reviewed and ultimately chose to not adopt the WCAP recommendations. Instead, the staff reported (Burley, 1971) (hereafter referred to as the Burley study) on the development of a theoretical iodine DF model which incorporated select measurement inputs from the Westinghouse study.

The purpose was to independently confirm the WCAP results and conclusions. The staff report recommended an effective iodine DF of 100 for fuel pin pressures up to 1200 psig with a minimum pool water level of 23 feet.

In November 1999, the staff performed an evaluation in response to a request from NRC Region II in Task Interface Agreement, TIA 99-03, Potential Non-Conservative Assumptions for Fuel-Handling Accidents, of whether the assumptions used in the Burley analysis were conservative for high burnup fuel. (USNRC, 1999) The evaluation concluded that adequate conservatism was provided by the analysis assumptions. In particular, the evaluation found that an effective pool iodine DF of 100 is likely overly conservative; therefore, the staff subsequently qualitatively increased it to a value to 200 in Regulatory Guide 1.183. The staff also updated the previous iodine speciation gap fraction break down of 99.75% elemental and 0.25% organic from the original Safety Guide 25 which is now superseded by Regulatory Guide 1.25 Assumptions Use for Evaluating the Potential Radiological Consequences of a Fuel Handling Accident in the Fuel Handling and Storage Facility for Boiling and Pressurized Water Reactors. (USNRC, 1972) The staff elected to specify the iodine gap fraction released to the pool as effectively 99.85% elemental 6

and 0.15% organic without re-computing the effective iodine DF based on the Burley method.

The recomputed effective iodine DF equates to 667. For a pool depth of 23 feet or greater, to give an overall iodine effective DF of 200, the assumed pool elemental iodine DF is 500 and organic DF is 1, with resulting iodine speciation of the release from the pool as 57% elemental and 43% organic. These speciation values are currently recommended in Regulatory Guide 1.183. Regulatory Guide 1.195, Methods and Assumptions for Evaluating Radiological Consequences of Design Basis Accidents at Light-Water Nuclear Power Reactors, (USNRC, 2003) which is based on methods prior to the publication of 10 CFR 50.67, retained the Regulatory Guide 1.25 assumption of 99.75% elemental iodine and 0.25% organic iodine species in the fuel gap, but revised the overall effective iodine DF to 200 if the depth of water above the damaged fuel is 23 feet or greater, consistent with the pool DF guidance in Regulatory Guide 1.183.

During the 473rd Advisory Committee on Reactor Safeguards (ACRS) Full-committee meeting, held June 7th, 2000, the committee members expressed concern with the staffs treatment of iodine speciation and the continuation of modeling the FHA as a puff release. 2 A staff member explained that the chosen DF of 200 represented not only the iodine immediate released following the postulated fuel rupture but also represented re-evolution of iodine. This re-evolution release was assigned to the immediate release for the purposes of determining the decontamination factor for simplicity. However, current research had shown that iodine re-evolution can potentially be significant and thus should be explicitly modeled. Paraphrased from the meeting transcript:

- MR. KRESS: It'll have much lower quantities [of iodine], but once again, they got the conservatism that they're assuming it comes in instantaneously, when actually it takes a considerable amount of time for this stuff [source term] to come out [of the SFP].

- Mr. Kress: [the staff] introduced completely unrealistic [assumptions by] going away from the intent of the new source term to put a little realism into it.

- DR. POWERS: I guess I wonder -- you know, if I'm sitting around, trying to figure out how I'm going to respond to an accident, and you tell me, okay, you've got a puff release and everything's over. And I pick up one set of actions. If instead you're telling me I've got a protracted release -

- MR. KRESS: Over a long period of time.

DR. POWERS: -- over days and days and days, I think I'd come up with a different set of actions.

- MR. KRESS: Yeah, and that's the other thing that bothers me. Part of the source term specification is the time. And here we've gone back to the puff release, just for convenience, when we know it's not a puff release, but we're saying it's conservative and I'm not sure it is, because you, you have one set of actions versus another and I'm not sure which is the right things.

2 https://www.nrc.gov/reading-rm/doc-collections/acrs/tr/fullcommittee/2000/ac000607.html 7

- Mr. KRESS: as an alternative, I would have I would have said, .25 percent of it goes in immediately and only a DF of 500 of the element goes in immediately, and the rest of it comes out protracted over time. And I would have made a calculation for what that protracted time release is Mr. Kress recommended 100% CsI in the fuel pin gap as opposed to assuming the iodine is a vapor. It would be an improvement to account for the reduction of amount of radioactive iodine available to be released as gas due to limited I2 vapor pressure since I2 is solid at FHA temperatures. In the Letter from ACRS member Dr. Powers to Dr. Travers, dated June 20, 2000, concerning the proposed final regulatory guides and standard review plan sections associated with the AST, the ACRS suggested changes to the staff for its consideration, in part:

a need for both minor editorial changes and to purge the Regulatory Guide of several carryover items from previous regulatory guides that are not appropriate for implementation of alternative source terms. Examples include the lack of adequate technical justification for the speciation of iodine in the fuel pin gap. 3 In 2004 (USNRC, 2004), the staff found that the peak assembly average pressure is an acceptable method rather than using the maximum fuel rod pressurization specified in the regulatory guidance. It was noted that the pressurization criterion was inadvertently omitted from Regulatory Guide 1.183 and is expected to be restored in a future revision for fuel that does not incorporate zirconium diboride. In effect, this decision approved the iodine DF of 200 for pressures up to 1,300 psig.

To date, these recommended changes have not been incorporated into the appropriate staff guidance. However, this work, as discussed, provides part of the technical basis and analysis to now update the staff guidance with an acceptable alternative method to the FHA. This alternative model modifies the existing FHA boundary conditions. It is based on the environmental conditions in which fuel handling operations are taking place; incorporating several improvements of our current understanding of reactor fuel pin physics, iodine chemistry and re-evolution while maintaining conservatism. When coupled with the latest cycle-specific gap fraction source term methodology for non-LOCA DBAs other than reactivity insertion accidents, the computed radiological doses are generally 91-98% lower than the originally computed. 4 The new model provides the same level of safety while providing some regulatory relief during fuel handling operations.

3. EVALUATION 3.1 Evaluation history By Informal Assistance Request (IAR) dated August, 17, 2018, from the Office of Nuclear Regulatory Regulation (NRR), Division of Risk Assessment (DRA), Radiation Protection and Consequence Branch (ARCB), requested the Office of Nuclear Regulatory Research (RES),

3 https://www.nrc.gov/reading-rm/doc-collections/acrs/letters/2000/4731896.html 4 ADAMS Accession Number ML090360256 8

Division of Systems Analysis (DSA), Fuels and Source Term Code Development Branch (FSCB),

to perform an independent review of an NRR-staff re-evaluation of the fission product release and transport model for the FHA DBA described in Regulatory Guide 1.183, Appendix B. The main objective of this re-evaluation was to revisit the original studies forming the technical basis for the FHA dose analysis methodology and seek input in updating the model with current information and practices consistent with other DBAs. RESs IAR response confirmed the staffs re-evaluation and recommended improvements to the fission product transport model. These improvements would be established from our current understanding of reactor fuel pin physics and iodine chemistry under the environmental conditions in which fuel handling operations are taking place.

The RES and NRR staff then developed the updated FHA fission product transport model under these conditions, with transcriptions from the 473rd ACRS Full-committee meeting, held June 7, 2000 and subsequent recommendations in a letter from ACRS member Dr. Powers to Dr. Travers, dated June 20, 2000. A follow-on Research Assistance Request dated June 5, 2019, requested RES to finalize the FHA transport model, have it peer-reviewed, and formally submit it to NRR through an inter-office memo as an acceptable method for staff use. RES staff contacted Sandia National Laboratory and other experts in the field to aid their work in developing the technical basis for this alternative FHA analysis methodology.

The evaluation and subsequent documentation are divided into four parts as follows:

1. Re-evaluation of the WCAP study using modern data analysis tools to confirm results and conclusions (see Section 3.2).

2. Re-evaluation of the Burley study using modern data analysis tools to confirm results and conclusions (see Section 3.3).

3. Analysis to developed flexible iodine DF models based on fuel pin pressure (see Section 3.4).

4. Alternative FHA analysis methodology (see Section 4).

3.2 Re-evaluation of WCAP-7518-L (non-proprietary version is WCAP-7828)

Westinghouse topical report WCAP-7518-L (1970), Topical Report: Radiological Consequences of a Fuel Handling Accident, reports on a series of experiments to measure the spent fuel pool water scrubbing effect of iodine and to estimate a generic iodine DF. The non-proprietary version of WCAP-7518-L is WCAP-7828 (Westinghouse, 1971), Radiological Consequences of a Fuel Handling Accident." Both reports are referred to as the WCAP. The tests were performed at conditions simulating those expected during a FHA. The program sought to verify the efficiency of iodine mass transfer through a detailed experimental program to simulate closely the actual conditions of the FHA. The test program consisted of two principal efforts to measure the mass transfer of iodine from gas bubbles to solution to yield the iodine DF as a function of bubble size and bubble contact time in the pool, both of which are a function of pressure.

Small-Scale Tests - to obtain quantitative measurements of the iodine and carbon dioxide absorption from gas bubble to surrounding liquid at the design basis solution conditions 9

(temperature and chemistry). Data was collected for various bubble diameters and solution depths.

Full-Scale Tests - to identify bubble patterns and gas bubble behavior upon their rise through the column of water in a deep pool, where gas released would be typical of a postulated damaged fuel assembly in a spent fuel pool.

From the full-scale tests, in conjunction with small-scale carbon dioxide tests, the effective bubble diameter of large-scale gas releases was ascertained. The test program did not directly measure the iodine DF in the full-scale pool. Instead, mass transfer coefficients for iodine measured from the small-scale tests were applied to the full-scale test data characterizing the sizes, patterns behavior of gas bubble traveling through the pool water column over a broad range of fuel pin pressures. Figure 1 provides a flow diagram of the WCAP test program process. It provides a high-level view of each experiment, the collected data, and the resulting data-driven predictive DF models.

Small-Scale CO2 Large-Scale CO2 Small-Scale Iodine

Purpose:

Measure CO2

Purpose:

Design basis test

Purpose:

Measure I2 mass transfer coefficient mass transfer coefficient Data: Before/after CO2 Data: Before/After CO2 concentrations, bubble size, Data: Before/After Iodine concentrations, bubble rise time concentrations, bubble size and rise time size, rise time Result: Measured CO2 DFs Result: Predictive CO2 Result: Predictive I2 DF DF Model (Equation 3-5 Model (Equation 3-4 below) below) 2_ = 1.54 0.1396 = 81.046 0.305 Predicted CO2 DFs CO2 predictive model Predicted Iodine DFs correlation verifies Model: 2_ Model:

the applicability for lnput: Large -scale bubble the small-scale iodine lnput: Large -scale bubble size and rise time. to predict iodine size and rise time.

decontamination in a Result: Figure 5 below which deep pool. Result: Figure 6 Below demonstrates correlation.

Figure 1: Flow Diagram of WCAP Test Program Process Since the full-scale tests did not measure iodine DF directly, Westinghouse analytically correlated the small-scale tests results with those obtained from the full-scale tests to estimate an iodine DF.

Westinghouse estimated an expected iodine DF of 760 for a 26-foot pool and applied a factor of 10

66 percent to obtain a conservative iodine DF of 500 for a 26-foot pool. Overall, the WCAP results demonstrated iodine is readily removed from the gas rising through the spent fuel pool water.

Small-Scale Tests and Results A series of small-scale tests were performed to obtain measurements of the iodine and carbon dioxide absorption from gas bubble to surrounding liquid. The purpose was to collect the necessary data to compute carbon dioxide and iodine DFs. The test assembly was composed of a 9-inch-diameter by 8-foot-high glass column, in which temperature, bubble size and solution chemistry could be carefully controlled. Tests explored the effects of pool depth (up to seven feet) and bubble diameter on the absorption of tracer gas species (iodine and carbon dioxide) from inert nitrogen or helium carriers. The solution used for the iodine tests was held at the design value of 120°F and consisted of demineralized water containing 2000 parts per million (ppm) boron as boric acid with a pH that ranged between 4.3 and 5.0. The carbon dioxide tests were conducted with a solution chemistry corresponding to the deep pool, large-scale tests, pH = 7.0 at 70°F.

All collected data pertaining to the small-scale test and resulting DFs can be found in WCAP Tables 3-1 and 3-2, for the iodine and carbon dioxide respectively. Selected data and important WCAP figures are provided in the following.

3.1.1.1 Estimation of Gas Bubble Parameters: Size and Rise Time Estimations of the gas bubble diameters were performed using the small-scale test assembly.

Measurements were performed for each inlet orifice (3/8, 1/2, and 1) by counting the number of bubbles delivered and measuring the volume collected. The volume delivered was measured with a wet test meter accurate to +/- 10 cm3. The rise times were measured as a function of water depth for each bubble size. Table 1 presents observed averaged bubble diameters and rise times.

Table 1: Measured Bubble Diameters - Small Scale Test Bubble Diameter Bubble Volume Bubble Rise Orifice Diameter Times Average Computed D (in) or (cm) T (sec/ft.) or (in) or (cm) obs. V (in3) or (cm3)1 (sec/cm) 0.375 or 0.9525 0.336 or 0.854 1.33E-2 or 0.219 1.10 or 0.036 0.5 or 1.27 0.393 or 1.00 1.78E-2 or 0.293 1.03 or 0.033 1 or 2.54 0.870 or 2.12 3.57E-2 or 0.586 0.89 or 0.029 1 - WCAP reported mixed SI and British units, both are included for consistency.

The data are well represented in linear fashion as:

. = 0.1479 + 1.2036, 2 = 0.9466. (3-1)

Where:

T = bubble rise time, 11

d = average bubble diameter (in).

The inner diameter of Westinghouse PWR fuel pins is 0.405 in (1.029 cm), equating to a bubble rise velocity of 1.03 sec/feet (0.033 sec/cm) is applied to the small-scale iodine and carbon dioxide DF calculations.

Staff Notes: Small-scale bubble sizes seem to be the volume-average over the duration of the bubble-size tests. Since the system depressurizes in the tests and it is considered that bubble size depends on pressure, not only would bubble size at any one time have distribution rather than a single size at any time but bubble size also varies throughout the small-scale test. The mass transfer for a bubble size determined as the total volume divided by total counted bubbles may not be representative of that where the bubble size distribution was carefully monitored.

3.1.1.2 Measured Mass-Transfer Coefficients and Resulting Iodine DFs WCAP Tables 3-1 and 3-2 provide the necessary information to compute tracer gas DFs as a function of bubble rise time and diameter. The data were found to fit a relatively simple analytical model describing the iodine DF of a pool of water. This model assumes constant pressure and does not consider simple expansion of the bubble which occurred in the experiments and the DBA FHA scenario. Considering a single bubble of trace component and carrier gas thoroughly mixed as the bubble rises to the pool surface, mass transfer from gas to liquid can be described by:

= = . (3-2)

Where:

C = concentration of trace species in the bubble gas phase at time t; Co = initial concentration of trace species in the bubble gas phase at t=0;

= effective area across which mass transfer occurs (cm2/cm3) 5

= overall mass transfer rate constant, sec-1; and, t = time of bubble travel.

The rate constant, , contains the mass transfer variable, which is composed of the trace species deposition velocity, Vt (cm/sec), and the effective area across which mass transfer occurs (cm2/cm3). For a perfect sphere, which is assumed for the analysis of all test data, the surface area is given as 6/d, where d (cm), is the spherical bubble diameter. Thus, the overall expression becomes:

5 Staff Note: This is not the effective cross-sectional area that is part of the Beta-term, it is simply a pre-exponential representing early decontamination.

12

6

= . (3-3)

Staff Notes: It should be noted that Co/C is not an adequate way to calculate decontamination for a depressurization/expansion event such as an FHA, the WCAP experiments, or bubble rise since simple expansion results in a concentration reduction even when no decontamination occurs. However, it is okay to characterize by considering the total mass captured and the total mass transmitted as the WCAP did for the iodine tests. It is also okay to consider the concentration reduction relative to other gases not being absorbed by water.

The WCAP used relative concentrations for both the large scale-and small-scale carbon dioxide tests.

Figures 2 and 3 plot iodine and carbon dioxide DFs as a function of bubble rise time, (),

on a semi-log plot. The data was fitted with a power function where the slope equal, 6 , for a given bubble size, assuming the deposition velocity, , is constant with the range of bubble sizes of interest. The overall expression which correlates the iodine data, , for all bubble diameters and solution depths tested is:

= 81.04 0.3050 , 2 = 0.91 . (3-4)

The small-scale iodine tests are well represented by a straight line with a slope equal to 0.305 cm/sec. Thus, the iodine mass transfer deposition velocity is taken as 0.305 cm/sec and the mass transfer rate is 0.305/6 cm = 0.0508 s-1.

The corresponding data for the carbon dioxide data, 2 , for all bubble sizes and solution depths tested is:

2_ = 1.54 0.1396 , 2 = 0.97. (3-5)

The small-scale carbon dioxide test data plots linearly with a slope equal to 0.1396 cm/sec. Thus, the carbon dioxide mass transfer deposition velocity is taken as 0.1396 cm/sec and the mass transfer rate is 0.1396/6cm = 0.0233 s-1.

The WCAP found the assumption that is reasonable when compared to Nate and Himmelblau (Nate & Himmelblau, 1967) who reported on measurement of carbon dioxide transfer from a gas bubble to surrounding water and compared their results to those of other work. Within the range of bubble diameters used in iodine and carbon dioxide tests in the present study, Nate and Himmelblau reported a variation in Vt of between 0.017 and 0.022 cm/sec, for equivalent diameters between 0.85 and 2.1 cm.

13

Staff Notes: This also suggest that the mass transfer characteristics using the volume average bubble size is somewhat close to that for the time-averaged bubbles size distribution throughout the test if the same gases were used in those experiments and if the diffusivities in both gases are either not limiting relative to that in the liquid or similar to each other.

10000 Iodine Vapor Bubble Decontamination Factor in Helium Solution pH 4.5, 120F 1000 Expon. (Iodine 100 Vapor in y = 81.046e0.305x Helium Solution pH R² = 0.9067 4.5, 120F) 10 0.00 2.00 4.00 6.00 8.00 10.00 12.00 t / d (sec/cm)

Figure 2: Recomputed WCAP Bubble Decontamination - Small Scale Tests with Iodine.

10.00 2% Carbon Dioxide in Bubble CO2 DFs - Samm Scale Helium Solution, pH 7, 70F Expon. (2%

Carbon Dioxide in Helium Solution, pH 7, DFCO2 = 1.5411e0.1396x 70F)

R² = 0.9747 1.00 0 2 4 6 8 10 t / d (sec/cm)

Figure 3: Recomputed WCAP Bubble Decontamination - Small Scale Tests with Carbon Dioxide.

14

Full-Scale Tests and Results The series of full-scale tests were performed to identify bubble patterns and gas bubble behavior upon their rise through the column of water over a series of release pressures to simulate various fuel pin pressures. This series of tests were set up in a 25-foot deep pool utilizing equipment which simulated the cross-section of a full-scale 14 x 14 fuel assembly. The mock-up assembly was fitted with a gas pressurization and safety relief network which permitted the instantaneous release of gas from each of the 179 fuel tubes. The test assembly conservatively assumed that damage to the fuel assembly resulted in complete and instantaneous shearing of all fuel rods, releasing the contained gas in the upward facing direction. Slightly soluble carbon dioxide was used as the tracer component to simulate fuel assembly gas. The time for gas bubbles to reach the pool surface, traveling through 23 feet from the test assembly, was measured at each pressure from 100- to 1400 psig. Tracer gas concentration measurements were made in the injection vessel prior to the test then from escaping gas at the pool surface, within the sealed floating blanket assembly. To insure a representative final concentration was obtained, repeated samples were collected over a 30-minute period.

3.1.2.1 Measured Gas Bubble Rise Time The bubble rise time was measured as a function of the initial test pressure. Table 2 presents bubble rise time data for each test pressure. Figure 4 (copied from WCAP Figure 3-6) more clearly shows the measured bubble rise time with respect to test release pressure.

Table 2: Test Pressure and Bubble Rise Time Data Test Case # Test Pressure (psig) Bubble Rise Time (s) 1 100 8.8 2 300 7.8 3 300 7.8 4 600 6.6 5 600 6.6 6 900 5.6 7 900 5.6 8 1200 4.7 9 1200 4.7 10 a 1300 4.2 11a 1400 4.5

a. Bubble rise time extracted from Figure 4.

15

Figure 4: Copy of WCAP Figure 3-6: Bubble Rise Time - Large Scale Test These measured bubble rise times are characteristic of the release of gas from the assumed damaged fuel assembly and rise through the spent fuel pool. Thus, they serve as the basis for the calculations of iodine DFs. The data are well represented in exponential fashion as:

() = 9.2261 (64) , 2 = 0.9866, (3-6) where:

x = test release pressure (psig).

Note: Section 3.4 will apply Equation 3-6 as the Bubble Rise Time model to compute new iodine DFs as a function of fuel pin pressure.

3.1.2.2 Effective Gas Bubble Size and Carbon Dioxide DFs Computed for a Deep Pool For each test release pressure, Figure 5 plots the averaged measured large-scale carbon dioxide DFs. The overall expression which correlated the large-scale carbon dioxide DF data,2_ ,

for all release pressures are represented well as a power function:

2_ = 12.002 0.162 , 2 = 0.89, (3-7) where:

16

x = test release pressure (psig).

Figure 5 also plots Equation 3-5 to predict CO2 DFs using the measured large-scale rise times and effective bubble diameters for each release pressure.

Bubble Decomtamination - Large Scale CO2 DF 7.00 Measured Large Scale CO2 6.00 DF 5.00 Ave. CO2 DF 4.00 Power (Measured Large 3.00 y = 12.002x-0.162 Scale CO2 DF)

R² = 0.89 2.00 Power (Predicted with 1.00 Equation 3-5) 0 200 400 600 800 1000 1200 1400 Release Pressure (psig)

Figure 5: Reproduction of Figure 3-9: Bubble Decontamination - Large Scale Test with Carbon Dioxide, and predictive CO2 DF Equation 3-6 Given the analytical expression for carbon dioxide removal, results from the large-scale tests can be analyzed to yield the effective bubble diameter of the larger gas volume releases. Since bubble rise times and DFs were measured for carbon dioxide, the effective bubble diameter can be obtained by rearranging Equation 3-3 to yield:

6

= 2_ (3-8)

The averaged data points for a given test release pressure, shown in Table 2, is substituted in Equation 3-5 and solved for effective bubble diameter, shown in Table 3 below.

Table 3: Test Pressure and Effective Bubble Diameter Release Effective 100 0.94 300 0.96 600 0.91 900 0.82 1200 0.73 1300a 0.74 1400a 0.72 (a) values were computed from Equation 3-8 based on measured bubble rise time selected from Figure 4.

These effective bubble diameters are therefore characteristic of the release of gas from the assumed damaged fuel assembly and serves as the basis for the calculations of iodine DFs.

These effective bubble diameters are well represented in linear fashion as:

17

() = 0.0002x + 1.0009 (3-9) where:

x = test release pressure (psig).

The WCAP concludes the measured and predicted CO2 DFs correlate well, lending to Westinghouses method of applying Equation 3-4 to predict iodine DF in a deep pool.

Staff Notes: This is simply calculating the effective bubble diameter that will return the DF using the small-scale test carbon dioxide DF model and is not truly a measurement of bubble size. This would correlate well even if completely wrong since the inverse of the equation was used to determine input for the equation where the DF was calculated by Equation 3-5, 2_ = 1.54 0.1396(), but d was calculated from rearranging Equation 3-3 to yield 2_

Equation 3-7, = 6 . Therefore, this equation would still correlate well even if bubble size was off.

This is a calculation that assumes the DF in the small-scale experiment matches that of the large-scale experiment where the diffusion of carbon dioxide in helium either matches that of carbon dioxide in nitrogen or that the diffusion of carbon dioxide in water is limiting. Enhanced mass transfer upon bubble formation can occur at higher velocities (higher pressures) thus changing the pre-exponential term. As mentioned above, we need to check how helium and nitrogen gas absorption in water compare to that of carbon dioxide as water may already be at saturation in nitrogen gas.

3.1.2.2 Iodine DFs Computed for a Deep Pool of Water Direct application of Equation 3-4, knowing the bubble rise times (see Table 2 and Equation 3-6) and effective bubble diameters (see Table 3), the iodine DF is obtained for a water depth of 23 feet. All data was fitted with a power function showing a decreasing slope, suggesting the pool DF does not change as rapidly with increasing pressure. The calculated iodine DFs are 1360 and 536 for release pressures of 100 and 1200 psig respectively. For pool depths greater than 23 feet., Equation 3-6 was used to extrapolate longer bubble travel times to calculate iodine DFs for each release pressure. Under 26 feet. of water, extrapolated iodine DFs are 1964 and 685 for 100- and 1200 psig respectively. Figure 6 presents computed iodine DFs for a deep pool computed at 23- and 26 feet. depths as a function of pressure. Table 4 compares recomputed iodine DF presented in the WCAP and those computed with this analysis where general agreement is seen between the two analyses. Finally, the WCAP recommended a generic iodine DF of 500 to be applied in DBA FHA radiological consequence analyses. Westinghouse indicated this value covers about 90% of the data over a broad range of fuel pin bundle pressures.

This re-evaluation confirms this recommendation.

18

2000 1800 1600 23 ft deep Pool 1400 With equation DF =

1200 81.046e^(0.305*(t/d))

Iodine DF y = 1821e-9E-04x 1000 R² = 0.9937 26 ft deep Pool 800 600 With equation DF =

81.046e^(0.305*(t/d))

400 y = 1271.7e-8E-04x R² = 0.9937 200 0

0 200 400 600 800 1000 1200 1400 Release Pressure Figure 6: Iodine Decontamination Factors for a Deep Pool Table 4: Iodine Decontamination Factors as a Function of Test Pressure Based on Test Data for 23 ft. of Extrapolated to 26 ft. of Water Release Water Pressure Recomputed Recomputed (psig) WCAP Equation Equation 3-4 WCAP Equation Equation 3-4 3-3 3-3 100 1221 1208 1738 1718 300 846 985 1148 1365 600a 653 753 856 1007 900a 554 597 711 775 1200 500 490 633 620 1300b --- 462 --- 580 1400b --- 437 --- 545 (a) Recommended range of release pressures to consider for alternative FHA iodine DF based on RES recommendations and a sufficient understanding of fission gas generation and fuel pin pressures. It seems reasonable to consider a decay time, thus less decay power, and use pool temperature to determine gas temperature and thus pin pressure. (Further discussion is found in Section 5 under Observation 2)

(b) values were extrapolated based on measured bubble rise time selected from Figure 4.

3.3 Re-evaluation of AEC Mass Transfer Method Burley developed an analytical model to calculate an overall effective iodine DF. The purpose was to independently confirm the WCAP results and conclusions. A parametric analysis was performed by varying both the bubble diameter (1.20 to 2.50 cm) and the iodine partition factor (PF) while applying the WCAP bubble rise time measurement at the 1200 psig test pressure.

These bubble sizes were based on a short literature review of observed test data. The staff qualitatively selected a central value on the axis between the most and the least conservative 19

values, such that an effective iodine DF considering both elemental and organic iodine would be 100 (elemental DF = 133). Since the DF for organic iodine is assumed to be unity because of its low solubility, the Safety Guide 25 (USNRC, 1972) assumption that 0.25 percent of the gap iodine activity is organic effectively limits the overall DF to 400 regardless of the value of the inorganic iodine DF.

This study adjusts the Safety Guide 25, and Burleys, assumption of 0.25% organic iodine to 0.15% to be consistent with Regulatory Guide 1.183. This change increases the overall iodine DF from 400 to 667. This adjustment numerically equates an DF value with the WCAP test data.

Gas Bubble Parameters (size, db, and rise time, t)

Three methods were cited and assessed to calculate bubble diameters under a variety of initial conditions. Table 5 lists Burleys assessment of each model and the WCAP measurements. The range of calculated bubble diameters show a broad spread between 0.85 cm and 2.5 cm.

The inside diameter of the Westinghouse PWR fuel pins is 1.02 cm. For small bubbles formed at an orifice under equilibrium conditions, an approximate relationship correlating the volume of the bubble (V) and diameter of the orifice (D) is V/D = 0.231. This relationship correlated well with the Westinghouse data; yielding diameters of about 0.85 cm under equilibrium conditions. Burley notes the "effect of changes in release pressure should be minimal for these bubbles the bubble diameters appears to remain relatively constant for nozzle diameters of about 0.4 cm or greater."

However, other experiments seemed to show an independence of bubble size with orifice size and for bubble size distribution to mostly depend on flow rate.

Based on the literature review, the Burley report stated the most probable range of effective bubble diameters is 1.2-1.6 cm, with greater than 99% certainty that the value will not exceed 2.00 cm. It is unclear how this uncertainty estimate was made. Burley stated, rather inconspicuously, the Westinghouse data yielded an effective measured bubble diameter of 1.21 cm with a confidence of 99%, taken as 3-sigma, not to exceed 1.37 cm. This discussion, or result, was not mentioned in the WCAP. With this information, the staff chose to perform their parametric analysis with larger bubble sizes than those measured by Westinghouse presumably to account for uncertainties and their current state-of-knowledge on the subject. In general, a larger bubble size results in decreased predicted DF values. However, the Burley model never correlated with the WCAP data.

The staff based the bubble rise time on the observed WCAP full-scale carbon dioxide tests (see Table 2). The minimum observed rise time for the bubble rising through 23 feet of water was reported as 4.7 sec, equivalent to an average velocity of 147 cm/sec. A rise time value of 4.7 sec, at the 1200 psig test pressure, was deemed suitably conservative and was used for all following calculations. Using rise time as a function of pressure should also be conservative for nominal or greater-than-nominal gas quantity (moles) since it results from an upward-directed jet.

20

Table 5: Assessment of Bubble Diameters, WCAP vs. Empirical Methods Source Assessed Bubble Diameter Westinghouse WCAP Data Method: Calculated Full-Scale Experiment - CO2 Small Scale Tests Release Pressure Eff. Diam.

(psig) (cm) 100 0.99 300 0.92 600 0.85 900 0.79 1200 0.76 Method: Computed from Full-Scale Experiment - CO2 Ave Diameter: 1.21 cm 99% Conf. Level: 1.39 cm Equilibrium Release from Orifice 6 Method: Empirical Vol. (V) to Dia. of the orifice (D)

= 0.231 Calculated Diameter: 0.85 cm Pressurized Source - Rise Velocity 7 Method: Empirical Rise velocity (vb) to Vol. (V) 1

= 29.86 6 Calculated Diameter: 2.0 - 2.5 cm 1

= () 2 Calculated Diameter: 1.5 - 2.0 cm 6 Datta (1950) 7 Taylor (1950) and Cole (1948) 21

Staff Notes: Bubble behavior has been studied extensively in the nuclear field (after these experiments) for decontamination in BWR suppression pools. The sizes that are being used for FHA do seem large. RES is looking at the different models for bubble size distribution and have found some are involved and a lot of references exist. These may better represent the sizes that one gets from these experiments.

It may be more accurate to consider external work on bubble size distribution. Bubble sizes (and distributions) for the WCAP-based model are not well known since they were back-calculated from DF. Considered at in depth in the nuclear industry for suppression pool decontamination modeling and is Important for other fields.

It would be good to explain the third velocity-based bubble-size estimate. The large-scale WCAP rise times were taken, fit, and the curve extrapolated to zero pressure to eliminate the flow induced by pressure resulting in a rise time of 9.4 s, a velocity of 73 cm/s, and a volume of 216 cm3. Further assuming flow and bubble conditions (shape factor of 2/3) a radius of curvature was obtained. Burley obtained an effective (migration-distance, not volume equivalent, based) diameter of 2-2.5.

This method seems to apply only to a single orifice and a single bubble. This should represent an upper limit because, in the event of multiple failures and in the FHA experiments, there exists a bubble swarm that increases the effective velocity and reduces rise time.

It is not clear way Burley seemed to have not considered smaller bubble sizes from multiple orifices and instead to opted for a higher size of 1.6-1.7 cm. The method of calculation or the basis for the 99% certainty was not explained.

Iodine Partitioning Factor The AEC staff derived an instantaneous iodine PF of 10 instead of utilizing the measured iodine PF of 26 under equilibrium conditions presented by Postma (1970). The staff felt the contact time for bubbles released from the pressurized fuel assembly was too short for the equilibrium value to be established before the bubble reached the surface of the pool. However, the way Burley used the iodine PF differs from the way it is currently considered for iodine.

Burley and Postma consider the iodine PF to be the ratio of concentration in the liquid phase to the concentration in the gaseous phase, PF = [Iinorganic,liquid] / [Iinorganic,gas]. In both reports, the iodine PF was considered a function of not only temperature, but also a function of concentration, pH, as well as being time-dependent. Additionally, Postma referred to the ratio as inorganic, Burley considered the ratio to be of I2 itself. Burley mentioned equilibrium partition factors but at the same time used the Postma instantaneous concentration measurements as instantaneous partition factors, as Postma did. The concentration and pH effects seen by Postma and adopted 22

by Burley are not considered to significantly affect the partition factor of I2. Rather, these parameters affect the aqueous speciation of iodine in water which is the concentration of I2 in water available for evolution and subject to partitioning. Other forms of iodine in water that are prevalent at different concentrations and pH are not volatile and do not evolve/partition into the gas space.

Staff Notes: Today, the iodine partition coefficient is considered differently. It is considered an equilibrium concentration ratio and not an instantaneous concentration ratio where iodine PF = ([I2liquid] / [I2gas]) under equilibrium conditions. The instantaneous concentration ratio will not necessarily correspond to the equilibrium concentration ratio but will tend towards that value. The iodine PF is only considered to apply to I2, not to the concentration of all iodine in water. The concentration and pH effects seen by Postma and adopted by Burley are not considered to significantly affect the partition factor of I2. Rather, these parameters affect the aqueous speciation of iodine in water which the concentration of I2 in water is available for evolution and subject to partitioning. Other forms of iodine in water that are prevalent at different concentrations and pH are not volatile and do not evolve/partition into the gas space.

Today, the iodine PF factor is considered solely a function of temperature. This is as used by NUREG/CR-5950 (ORNL, 1992) (also referred to by the primary author, Beahm, et al.),

MELCOR, and in other international iodine models.

It is recommended that the MELCOR PF be used. The NUREG/CR-5950 PF provides values similar to the MELCOR PF at the temperature range of the DBA FHA. Equilibrium relationships (partition coefficients) can be affected by concentration and pressure. At higher concentrations and pressures, equilibrium curves do not necessarily follow Henrys law. At low concentrations and modest pressures, the relationship generally follows Henrys law.

Since the FHA scenario involves low iodine concentrations and pressures Henrys law is considered to apply.

23

Staff Notes (continued): According to conversations with RES and Dr. Powers, the non-equilibrium assumption was presumed as a conservatism in the absence of knowledge of the equilibration rate. Whether equilibrium is reached or not depends on bubble size and mass transfer rate. Beahm, et al. (ORNL, 1992) explains the mass transfer between gas and water is sufficiently quick and that equilibrium can be assumed in steam suppression pools which is similar in geometry to a spent nuclear fuel pool. As mentioned above, the current iodine models consider the iodine PF is solely a function of temperature. The equilibrium distribution of iodine is represented by the iodine partition factor, PF(I2):

(2 )

(2 ) = (2 )

(3-10)

The partition coefficient for iodine can be obtained from:

10 (2 ) = 6.29 0.0149 () (3-11)

Where:

T = temperature in Kelvin.

Equation 3-8 yields higher iodine PF values as temperature decreases.

The WCAP full-scale experiments were conducted at temperatures between 65°F and 70°F.

Typical spent fuel pools are operated between 100°F and 115°F F. Equation 3-8 yields iodine PFs of 45 and 34 for temperatures of 100°F and 150°F respectively.

This analysis selects to apply the more conservative iodine PF of 35.

Iodine Speciation The AEC staff assumed the maximum equilibrium faction of organic iodine in the fuel pin gap should be 0.25%. This was in agreement with the observed releases from single fuel pins reported by Parker (1967) at Oak Ridge National Laboratory. Since the solubility of organic iodine in water is extremely low, assigning a factional non-soluble iodine species to the total iodine available for release from the gap and plenum regions imposes an upper numerical limit to the overall attainable iodine decontamination factor in water. For 0.25% organic iodine, the limiting decontamination factor has a numerical value of 400.

24

Regulatory Guide 1.183 gives an updated iodine speciation of 99.85% elemental and 0.15%

organic iodine. For 0.25% organic iodine, the limiting decontamination factor has a numerical value of 667.

The reanalysis should apply an organic iodine speciation of 0.15%.

Iodine Mass Transfer Coefficient Method The AEC staff computed the iodine DF using a theoretical treatment of iodine mass transfer without reference to those measured by Westinghouse (see Section 3.2). However, certain parameters from the WCAP were applied as follows.

Equations applicable to the evaluation of the specific mass transfer coefficient are as follows:

1

= 1 1 , (3-12)

+

+

Where:

= 1.646 , (3-13)

= 3.75103 ( ), (3-14)

= 4.48 ( ), (3-15) 1 2

= 1.13 , (3-16)

DG = molecular iodine, I2, diffusivity in He = 0.278 (cm/sec) db = bubble diameter (cm)

Vb = bubble velocity (cm/s)

DL = I2 diffusivity in water = 1.27E-5 (cm/sec)

P = I2 partition factor The bubble rise time for a simulated damaged fuel assembly was measured by Westinghouse and reported in the WCAP (see Table 4 above). The minimum observed residence time for the bubble rising through 23 feet of water was reported as 4.7 sec at a test pressure of 1200 psig, equivalent to an average velocity of 147 cm/sec. To compute a mass-transfer coefficient, Burley assumed a bubble diameter of 1.6 cm and iodine partition coefficient of 10, computing a mass-transfer coefficient of 0.26.

The mass transfer coefficient was re-computed under turbulent and laminar flow conditions with an effective bubble diameter of 1.21 cm (based on the WCAP) and an iodine PF of 35 (based on Beahm, et al. (ORNL, 1992)). Computed turbulent- and laminar mass-transfer coefficients are 0.52 and 0.59, respectively. These coefficients are rather high. When applied to the Burley iodine DF model, inorganic DF approaches infinity and the effective DF reaches the limiting numerical value of 667 for nearly all bubble sizes.

25

Staff Notes: Burley does not list sources for equation 3-9. The general form and dependencies are similar to correlations that have been validated against a large quantity of data for different gases using carefully controlled and characterized bubble sizes. Burleys inclusion of the partition coefficient to determine the effective mass transfer coefficient from that of the liquid and gas is standard in the two-film mass-transfer model and takes the general form:

1

=

1 + 1H kl An improved model (surface renewal) exists that accounts for the interaction processes of the two diffusivities. It does not change the results and do not remove the dependence on equilibrium concentration ratio (partition coefficient). Other phase-specific mass transfer models include diffusivity, bubble size, and velocity in a similar manner to that used by Burley. Other models consider other effects such as whether water is pure or contaminated.

Parameters such as these are typically used to account for bubble shape and recirculation within a bubble dependent on size. Either an original source (with comparison against data) should be found for Burleys relation with regions of applicability or an alternate model should be used. It would be good to have as a constraint that the combination of bubble size (distribution) and mass transfer coefficient tested against WCAP data.

Staff Notes: The bubble sizes in the WCAP experiments and analysis werent carefully controlled or characterized. In the small scale tests there was a bubble size distribution that changed as the injection vessel depressurized. The effective bubble sizes for the large-scale test and model were back-calculated from the large-scale test DF and rise time using the small-scale carbon dioxide test DF curve.

The mass transfer coefficient derived from fitting the DF data using the volume-averaged would not return the same result than if determined considering the transient bubble size distribution. Although WCAP tests measure decontamination they dont provide an adequate estimate of the mass transfer coefficient since bubble sizes were not characterized.

AEC Method to Calculate Iodine Decontamination Factors The AEC staff performed a parametric analysis to evaluate the overall effective iodine DF by varying both the bubble size and the iodine PF. The AEC staff considered a combination of a bubble diameter of about 1.6-1.7 cm and an iodine DF of 10 to represent a probable lower bound estimate for the assumed FHA spent fuel pool condition (as mentioned before, this is low compared to the current understanding and modeling approaches). One this basis, the staff chose a central iodine DFeff to be of the order of 100.

26

This analysis adjusts the bubble sizes and iodine PF based on data developed in the previous sections. Bubble sizes ranged from 0.86 cm (the mean value of the Westinghouse small-scale experiments) to 1.39 cm (the maximum "effective" bubble diameter from the Westinghouse large-scale experiments. Equilibrium iodine PFs ranged from 26 (Postma 1970) through 35 to 45 for spent fuel pools at operating temperatures between 100 and 115 F.

The iodine DF is defined as the ratio of initial to final concentrations of the species of interest within the bubble.

=

(3-17)

Where:

Co = initial concentration Cf = final concentration The DF for elemental iodine is given by the following:

6

= (3-18)

Where:

db = bubble diameter (cm) keff = Mass transfer coefficient Vb = bubble velocity (cm/s)

H = bubble rise height (cm)

This is the same formulation used in WCAP (except for the pre-exponential term that accounts for early decontamination). The 6/d represents the surface to volume ratio. The residence time, t, is represented by t = distance (i.e. height) / (bubble rise rate).

The overall effective iodine decontamination factor represents a composition for the several different iodine species, and is given by:

1

=  %  % (3-19)

+

1 For 0.15% organic iodine, the limiting decontamination factor has a numerical value of 667. (The organic faction could likely be the limiting factor on releases.)

Table 6 presents results of the parametric analysis. As the inorganic DF approached infinity, the effective DF reaches the limiting numerical value of 667 for nearly all combinations of bubble size and iodine PF.

Table 6: Reproduction of Burley, Table IV, Iodine Decontamination Factors 27

Bubble PF = 26 PF = 34 PF = 45 DFInorg DFeff DFInorg DFeff DFInorg. DFeff Diameter (cm) 0.86 547467392 667 7215344363 667 55215200809 667 0.945 39312060 667 370700801 667 2179058300 667 1.03 4874244 667 35324464 667 168808663 667 1.115 900672 666 5282518 667 21378578 667 1.2 224456 665 1107046 666 3910832 667 1.285 70349 660 300407 665 947882 666 1.37 26373 650 99734 662 286122 665 3.4 Analysis Computing Flexible Iodine Decontamination Factor Four iodine DFs models were developed; two applying the WCAP method and two applying the AEC method. The purpose was to understand how each model computes iodine DFs as a function of bubble size and rise time through the water column, both of which are functions of fuel pin pressure. Sensitivity analyses were performed for each model to determine how certain independent variables impact the new models. Modeling uncertainties were analyzed for the most sensitive variables using parametric bootstrapping. Iodine DF results provided below are an estimated simple mean of 1000 simulations.

Iodine DF Models The WCAP-based expression which correlates the iodine data, , for all bubble rise times and diameters as a function of release pressures tested is:

= 81.04 0.0508/6 , 2 = 0.91 (From Equation 3-4)

The Burley-based theoretical expression to calculate iodine DFs for all bubble rise times and diameters as a function of release pressure is:

6

= (From Equation 3-18)

Both formulations are essentially the same. Except the WCAP accounts for the pre-exponential term presumably accounting for early decontamination. The 6/d represents the surface to volume ratio. The residence time, t, is represented by t = distance (i.e. height) / (bubble rise rate).

Iodine Mass Transfer Coefficient The WCAP measured the iodine mass transfer deposition velocity to be 0.305 cm/sec which corresponds to a mass transfer rate of 0.305/6 = 0.0508 cm/sec.

The Burley calculated a theoretical iodine mass-transfer coefficient based on first principles is 0.2642 cm/sec. The re-computed Burley model under turbulent and laminar flow conditions with an effective bubble diameter of 1.21 cm (based on WCAP measurement) and an iodine PF of 35 (based on Beahm, et al. (ORNL, 1992)) yields turbulent- and laminar mass-transfer coefficients are 0.58 and 0.74, respectively.

28

This analysis selects to apply the measured the iodine mass transfer deposition velocity of 0.305 cm/sec which corresponds to a mass transfer rate is 0.305/6 = 0.0508 cm/sec.

Gas Bubble Size Model (d)

The effective bubble diameter is a characteristic of gas release pressure from the assumed damaged fuel assembly. Thus, the bubble size as a function of pressure serves as the basis for iodine DF calculations.

This analysis selects to apply two Bubble size models:

1. An averaged bubble diameter of 1.21 cm with a 99% confidence interval not to exceed 1.37 cm based on observed data.

2. The effective bubble diameter using Equation 3-7 fitted to the large-scale carbon dioxide results (Equation 3-8) as a function of test pressure (see Table 3 for computed effective bubble diameters).

29

Staff Notes: There is quite low confidence in the bubble sizes used by both WCAP and Burley.

However, they are generally in the right range.

The bubble size distribution changes with distance from orifice, orifice size, and with flow rate (which depends on orifice size and pressure). Pressure varied throughout tests so distribution changed throughout tests. For the small-scale tests different experiments were used to measure DF and estimate bubble size. The tests varied in pressure from 100-900 psig and used different liquid compositions and temperatures. The nominal bubble sizes from the bubble tests were used as input to evaluating DF for the small-scale tests. What is uncertainty? What was variation? Was bubble-size consistent from test-to-test and within a given test?

Bubble size was measured in the bubble-size tests by capturing released gas, counting bubbles, dividing total volume by number of bubbles to get average bubble size, and then calculating an effective diameter assuming a sphere. The surface/volume based diameter, not the volume average diameter, scales with mass transfer.

For the large-scale tests and model bubble size was back-calculated from the small-scale CO2 test DF. This approach implicitly involved a lot of assumptions. Furthermore any errors in the CO2 decontamination propagates to the calculated bubble size for the large-scale tests and for model.

Burley didnt explain in detail how bubble sizes were obtained. Bubble swarms also behave differently than single bubbles. Does it affect scalability in determination of initial bubble size?

Bubbles have subsequently been looked at in the nuclear field for the purposes of evaluating suppression pool decontamination.

Gas Bubble Rise Time Model (t)

The bubble rise time is dependent on release pressure and is a characteristic of the release of gas from the assumed damaged fuel assembly. Thus, the bubble rise time as a function of pressure serve as the basis for iodine DFs calculations.

This analysis selects to apply bubble rise times as a function of release pressure using Equation 3-6, () = . () , from 100 psig to 1400 psig based on observed data.

Parametric Bootstrap Sampling and Mean Parameter Values A parametric bootstrap sampling method was performed on each of the four models. The purpose was to understand how each model computes iodine DFs as a function of bubble size and rise time through the water column, both of which are functions of fuel pin pressure. Table 7 provides details of each iodine DF model.

30

The Bubble Size models are considered to have a normal distribution around the mean as a function of release pressure. The Bubble Rise Time model computes discrete rise times as a function of release pressure. Random sampling was performed on the Bubble Size. The calculation was repeated 1000 times and an averaged iodine DF was computed. Five hundred iterations was sufficient for convergence. Table 8 provides a summary of modeling assumptions.

Table 7: Flexible Iodine DF Models Model Model-Type Bubble Size Model Bubble Rise Time Model (60.0.508 1 )

µ = 1.21 cm 1 - WCAP-based = 81.04 () = 9.2261 (64) 3 = 1.37 cm 1 ) µ = () =1.0359e3E04x 2 - WCAP-based = 81.04 (60.0.508 () = 9.2261 (64) 3 = function of pressure 6 µ = 1.21 cm 3 - AEC-based = () = 9.2261 (64) 3 = function of pressure 6 µ = () =1.0359e3E04x 4 - AEC-based = () = 9.2261 (64) 3 = 1.37 cm 31

Table 8: Summary of Modeling Assumptions PARAMETERS WCAP (1970) Burley (1971) Chosen Parameter (2018)

IOINDE MASS TRANSFER COEFFICIENT Deposition Velocity =

Deposition Velocity =

Deposition Velocity = 0.2925 (cm/sec) 0.52 (cm/sec), turbulent flow 0.305 (cm/sec) 0.59 (cm/sec), laminar flow Basis: Small-Scale Tests - Quantitative Basis: First principles calculation. Assumes Basis: Directly measured. Does not include measurements of the iodine absorption from turbulent flow. additional assumptions (e.g. PF, Bubble gas bubble to surrounding liquid. Diameter, Rise Time). Value is similar to Conditions: PF = 10, Bubble Diameter - 1.6 calculated results produced under laminar Conditions: Iodine vapor in the helium cm, Rise time = 4.7 sec. flow assumption.

carrier gas bubbled through boric acid solution (ph=4.3-5.0) at 120 F. Application: Utilized to compute iodine DFs under both WCAP and Burley 0.2925 (cm/sec) / 6 = 0.0488 (cm/sec) 0.52 (cm/sec) / 6 = 0.0867 (cm/sec) 0.305 (cm/sec) / 6 = 0.0508 (cm/sec) 32

Table 8 (continued)

BUBBLE PARAMETERS Rise time: Release Pressure Rise Time Release Pressure Rise Time The data are well represented in linear and (psig) (sec) (psig) (sec) exponential fashion.

100 8.8 1200 4.7 300 7.8 Exponential model applied:

600 6.6 t (sec)= 9.2261e^(-6E4x), R=0.9866 900 5.6 1200 4.7 1300 4.25 1400 4.5 Exponential fit:

t (sec)= 9.2261e^(-6E4x), R=0.9866 Basis: Full-Scale Tests - At varying test Basis: Measurements from WCAP full- Basis: Fitted equation 3-7, based on direct pressures, identified gas bubble patterns scale tests. measurements.

and behavior upon their rise through a column of water. Conditions: Temperature range from 65 to Application: Utilized to compute iodine DFs 70 F, pH~7.0, 23 ft. of water. Minimum under both WCAP and Burley models as a Conditions: Temperature range from 65 to observed rise time at peak release pressure. function of release pressure.

70 F, pH~7.0, 23 ft. of water.

Effective diameter: Small Scale Tests Effective Diam. = 1.6cm The data are well represented in linear and Release Pressure Eff. Diam. exponential fashion.

(psig) (sec) The range of bubble diameters calculated 100 0.99 show a spread between 0.85 cm and 2.5 cm. Linear fit:

300 0.92 Concludes with "greater than 99% d (cm) = -2E-4x(psig)+1.0009 600 0.85 certainty that the value will not exceed 2.00 900 0.79 cm."

1200 0.76 Linear fit:

d (cm) = -2E-4x(psig)+1.0009 Basis: Fitted equation, based on direct Basis: Literature review and qualitative Basis: Fitted equation, based on direct measurement of CO2 tests. From the full- judgements. measurement of CO2 tests. From the full-scale tests, in conjunction with small-scale scale tests, in conjunction with small-scale carbon dioxide tests, the effective bubble Application: Parametric analysis between carbon dioxide tests, the effective bubble diameter of large-scale gas releases were 1.2 and 2.5cm. diameter of large-scale gas releases were ascertained. ascertained.

Application: Utilized to compute iodine DFs under both WCAP and Burley models as a function of release pressure.

IODINE PARTITIONING FACTORS No numerical value. PF = 26 (equilibrium), No numerical value.

PF = 10 (probable lower bound)

Note: Burley reports a PF of 13 following PF = 5 (conservative) Note: Burley reports a PF of 13 following conversations with Westinghouse. It is conversations with Westinghouse. It is unclear how this value was determined. PF = 35 and 45 (Beahm) unclear how this value was determined. PFs were not explicitly measured by Westinghouse. The value was most likely inferred through back-calculations.

Basis: Implicitly measured through small- Basis: Derived an instantaneous iodine Basis: Implicitly measured through small-scale iodine tests. PF of 10 from data. Assumes equilibrium scale iodine tests.

PF would not be reach during the bubble rise time.

Application: Parametric analysis with PF of 5, 10 , 26, 35 (Beahm), and 45 (Beahm).

33

Table 8 (continued)

IODINE SPECIATION Iodine in the various solutions was Iodine speciation of 99.75 elemental and Iodine in the various solutions was determined as iodine ion by use of specific 0.25% organic iodine. determined as iodine ion by use of specific ion electrode and calibration solutions. ion electrode and calibration solutions. The The iodine in boric acid solution in the iodine in boric acid solution in the column column was reduced to iodine with was reduced to iodine with hydrazine hydrazine before its determination to before its determination to complete its complete its hydrolysis. hydrolysis.

All test with iodine vapor in helium carrier All test with iodine vapor in helium carrier gas bubbled through boric acid solution gas bubbled through boric acid solution (ph.=4.3-5.0) at 120 F. (ph.=4.3-5.0) at 120 F.

The WCAP analysis didnt really address The WCAP analysis didnt really address organic iodine or other forms. They organic iodine or other forms. They worked worked with I2. (as they should have). No with I2. (as they should have). No Cs was Cs was present for I to be attached to form present for I to be attached to form CsI. It CsI. It would be solid at FHA would be solid at FHA temperatures in any temperatures in any case. The carbon in case. The carbon in the prototypic FHA the prototypic FHA comes from impurities comes from impurities in fuel fabrication.

in fuel fabrication.

Basis: Implicitly measured through small- Basis: Safety Guide 25. Assuming 0.25% Basis: Regulatory Guide 1.183. Assuming scale iodine tests. organic iodine, the limiting 0.15% organic iodine, the limiting decontamination factor has a numerical decontamination factor has a numerical value of 400. value of 667.

Application: Iodine Deff calculation. Application: AEC Iodine Deff model.

Results of WCAP and Burley Iodine DF Model Re-analysis as a function of Release Pressure The collected data and developed models in the preceding sections have been applied to compute iodine DFs as a function of release pressure. A brief description of the results is provided below. Figure 7 plots the computed iodine DF as a function of fuel pin pressure for each model as well as the historical models discussed above.

1400 1200 WCAP (1970) Data 1000 Linear (WCAP (1970)

Recommendation)

Iodine DF Linear (AEC (1971) 800 Recommendation)

Linear (RG 1.183 (1999) 600 Adjustment)

Log. (Model 1 (2018), WCAP) 400 Log. (Model 2 (2018), WCAP) 200 Linear (Model 3 (2018), AEC) 0 0 200 400 600 800 1000 1200 1400 Release Pressure (psig)

Figure 7: Comparison of Various Iodine DF Models 34

Model 1 - Derived from the WCAP data; applies an 'effective' mean bubble diameter of 1.21 cm for all release pressures and the Bubble Rise model. When compared to the WCAP results, computed iodine DFs results are under estimated, from -40%-difference at 100 psig to -49% difference at 1200 psig. A simple logarithmic function has been fitted to the computed results for comparisons with the WCAP. The model slope as a function of test pressure, is consistent with the WCAP. It is suspected an effective mean bubble diameter of 1.21 is too large.

Model 2 - Derived from the WCAP data; applies an 'effective' mean bubble diameter using the Bubble Size Model, and the Bubble Rise Model. When compared to the WCAP data points, computed iodine DFs results over estimate, from +13%-difference at 100 psig to

+9% difference at 1200 psig. A simple logarithmic function has been fitted to the computed results for comparisons with the WCAP data points. The model slope as a function of test pressure, is consistent with the WCAP data.

Model 3 - Based on the Burley model; applies an 'effective' mean bubble diameter of 1.21 cm for all release pressures, and the Bubble Rise Model. A simple linear function has been fitted to the computed results for comparisons with the WCAP data points. When compared to the WCAP data points, computed iodine DFs results are under estimated, from -45%-difference at 100 psig to +32% difference at 1200 psig. The fitted iodine DF model slope, as a function of test pressure, is not consistent with the WCAP data. The slope is essentially constant for all test pressures. This is due, in part, to the modeling assumption of assigning a factional non-soluble iodine species to the total iodine available for release from the fuel pin gap. This imposes an artificial upper numerical limit to the overall attainable iodine decontamination factor in water. For 0.15% organic iodine, the limiting decontamination factor has a numerical value of 667. The mass transfer coefficient should be adjusted as a function of bubble size.

Model 4 - Based on the Burley model; applies an 'effective' mean bubble diameter using the Bubble Size model, and the Bubble Rise model. A simple polynomial function has been fitted to the computed results for comparisons with the WCAP data points. When compared to the WCAP results, computed iodine DFs results are under estimate, from -

45%-difference at 100 psig to -24% difference at 1200 psig. The fitted iodine DF model slope, as a function of test pressure, it is not consistent with the WCAP data which trend well with a power function. This is due to the modeling assumption of assigning a factional non-soluble iodine species to the total iodine available for release from the fuel pin gap as well as a constant mass-transfer coefficient for all test pressures. This imposes an artificial upper numerical limit to the overall attainable iodine decontamination factor in water. For 0.15% organic iodine, the limiting decontamination factor has a numerical value of 667.

The mass transfer coefficient should be adjusted as a function of bubble size.

All models can be extrapolated proportionally, given the observed bubble rise times from the large-scale experiments to longer bubble rise times to calculation iodine DFs for actual spent fuel pool depths of 26 and 40 feet. However, these values would be outside the WCAP data tested under 23 feet of water.

35

Re-Analysis Observations A few observations are as follows:

Observation 1: Inconsistency between the WCAP iodine DFs and the AEC model.

Results between the two studies are inconsistent. It can be inferred from the Burley report that the staff had concerns about the analytical method used by Westinghouse. The cause for this discrepancy was found with Burleys assumed bubble size, the use of an instantaneous iodine PF, and the original 1970s model, limit of a numerical iodine DF value of 400.

Models 3 and 4 were adjusted with the updated bubble sizes, an iodine PF consistent with SFP conditions, and updated organic iodine speciation. This new model has a limiting iodine DF of 667. These corrections do not correct the models slope to trend with the WCAP data.

The WCAP analysis didnt really address organic iodine or other forms. No cesium was present for iodine to be attached to form CsI. It would be solid at FHA temperatures in any case and re-evolution would be important given the low pH for spent fuel pools. Additionally, if iodine is completely insoluble a limit would be provided on the WCAP-based models which only considered molecular iodine, I2.

Solubility of iodine has been characterized in other studies for suppression pool scrubbing. Concentrations are low, and volumes are large.

Observation 2: Iodine DFs as a function of fuel pin pressure under spent fuel pool temperatures.

Historically, the staff has assumed the FHA DBA fuel pin pressures are at operating pressure within the reactor core. However, there is now a sufficient understanding of fission gas generation and fuel pin pressures under various conditions. At spent-fuel-pool temperatures with decay heat rather than at reactor conditions at power, the gap pressures are much lower. Therefore, it would be more realistic to use the upper limit of gas generation models rather than assuming fuel pin pressures are at operating pressures. It would be reasonable to consider a wait time (e.g. as required by a plant-specific Technical Specification) for less decay power and use spent-fuel-pool temperatures to determine gas temperature and thus pressure.

For example:

The Office of Research performed an analysis in support of a recent alternative source term license amendment request to estimate a bounding end-of-life rod internal pressure for fuel being moved in the reactor cavity or spent fuel pool. The purpose was to compute best-estimate rod internal pressures as a function of time 36

post shutdown. To do so, a series of FRAPCON models were developed to model a typical PWR fuel rod with an aggressive power history that maximizes at-power rod internal pressure at power with a target around 3000 psia. The case assumed a Westinghouse 17x17 PWR fuel assembly that operated at high power for two cycles to achieve a rod average burnup of 55.2 GWd/mtU. Three simulations modeled reactor shutdown and decay heat up to 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> post shutdown. In additional, a fourth case was added that assumed the reactor conditions remained consistent with the conditions immediately before shutdown. The fourth case was designed to be an extreme case to demonstrate the impact of coolant conditions on rod internal pressure.

The results of rod internal pressure as a function of time post shutdown are presented in Table 9. As expected, rod internal pressure decreases as both a function of time (due to less decay heat) and with lower spent fuel pool temperatures. The pool temperature has a much stronger effect than decay heat at longer times post-shutdown. Case #2 best represents an upper temperature limit under a fuel candling conditions where temperatures can range between 100ºF and 115ºF and movement of recently irradiated fuel is controlled by the facilities Technical Specifications which generally does not begin until at least 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> post-shutdown. Under these conditions, estimated rod internal pressures are about 775- to 770 psi (760- to 755 psig) at 24- and 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> respectively.

Table 9: Coolant Conditions and Results of Fuel Rod Internal Pressure Temp. Pressure Model Rod Internal Pressure (psi)

Case #

(ºF) (psi) type 1 min 24 hrs. 48 hrs. 72 hrs. 100 hrs.

1 70 14.7 1974 709.7 671.3 669.3 668.7 668.4 2 150 14.7 1974 816.8 774.1 772.1 771.6 771.2 3 238 24.7 1974 935.7 888.1 886 885.4 885.1 4 555 2250 1974 1373.8 1354.2 1352.6 1352.1 1351.8 Figure 4 above presents iodine DFs as a function of test pressure. At test pressures of 600 and 900 psig, the computed iodine DF is 752 and 490 respectively. By applying Equation 3-4 (with Equations 3-6 for bubble rise time and 3-8 for bubble diameter) for a rod internal pressure at 760 psig, the computed iodine DF is 662.

Observation 3: Applicability of the current DF model to shallower water columns.

It is not uncommon for a spent fuel pools to have less than 23 feet of water above the damaged FHA assembly, typically between 21 and 22 feet. In practice, the staff have required licensees to reduce the iodine DF from 200 to some lower value to account for this reduction in the water column depth. For water depths less than 23 feet, Regulatory Guide 1.183 directs staff to the exponential factor described by 37

Burley on page 26: = (6 )() , where H is the bubble rise time (cm), and recalculate for the applicable distance. This is the general approach for vapor mass transfer from bubbles. However, as currently modeled, the Burley equation is highly, perhaps overly, sensitive to water level due to input assumptions. For instance, a reduced water level of 1 foot computed an iodine DF of 167, or a percent-difference of -17%. In addition, it is based a PF of 10 and an organic iodine speciation of 0.25%, both of which are highly sensitive parameters and no longer recommended values.

The ( ) term is simply the bubble rise time which is described in WCAP report, page 3-14, as time of bubble travel, i.e., time of mass transfer. This parameter can be adjusted directly rather than recalculating. Therefore, new Iodine DF are computed for each test pressure and by applying the new parameters discussed above where the:

• Bubble Size Model: db (cm) = -0.0002x(psig) + 1.0009;

• Bubble Rise Time Model: vb (sec) = 9.2261e(-6E-4x(psig));

• Mass Transfer Coefficient keff (cm/sec) = 0.305; and,

• Iodine Speciation = 0.9985 elemental, 0.0015 organic.

Table 10 presents the computed iodine DFs vs. water depth with corresponding percent-differences from the 23 ft of water base case. Using the test pressure case of 900 psig as an example, the base case iodine DF at 23 feet of water is 664, at 22 feet of water the iodine DF is 662 or a percent-difference of -0.3%.

This sensitivity analysis shows little change in iodine DF with respect to water level and range of pressures due to the limiting assumed iodine speciation. It may be more practical (to reduce request for additional information) to assume a single iodine DF being applicable to a range of water levels.

38

Table 10: Computed Iodine DFs with Burley Model vs. Water Depth Water Depth Test Pressure (psig)

Ft 100 300 600 900 1200 1300 1400 23 667 667 666 664 658 654 649 22.5 667 666 666 663 655 651 645 22 667 666 665 662 653 647 640 21.5 667 666 665 661 649 642 634 21 667 666 664 659 644 636 626 20.5 666 666 664 657 639 629 617 20 666 666 663 654 632 620 605 19.5 666 665 661 650 623 609 592 19 666 665 660 645 612 596 576 Water Depth Percent-Difference from 23ft Base Case (Above)

Ft 100 300 600 900 1200 1300 1400 23 (based case) 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%

22.5 0.0% 0.0% 0.0% -0.1% -0.4% -0.5% -0.6%

22 0.0% 0.0% -0.1% -0.3% -0.8% -1.1% -1.4%

21.5 0.0% 0.0% -0.1% -0.5% -1.4% -1.8% -2.4%

21 0.0% -0.1% -0.2% -0.8% -2.0% -2.7% -3.6%

20.5 0.0% -0.1% -0.3% -1.1% -2.9% -3.9% -5.0%

20 0.0% -0.1% -0.5% -1.5% -4.0% -5.2% -6.7%

19.5 -0.1% -0.2% -0.7% -2.1% -5.3% -6.9% -8.8%

19 -0.1% -0.3% -0.9% -2.8% -6.9% -8.9% -11.2%

Observation 4: Applicability of the iodine chemistry PF model.

Burley derived an instantaneous iodine PF of 10 instead of utilizing the measured iodine PF of 26 under equilibrium conditions presented by Postma (1970). The staff felt the contact time for bubbles released from the pressurized fuel assembly was too short for the equilibrium value to be established before the bubble reached the surface of the pool.

In 2018, conversations with RES and Dr. Powers suggested that the presumed non-equilibrium assumption was made as a conservatism in absence of knowledge of the equilibration rate. Dr. Powers and RES confirm the Beahm, et al. (ORNL, 1992) model could be used to update the FHA iodine DF model.

The Burley iodine DF model is highly sensitive to the iodine PF value and the Beahm, et al. (ORNL, 1992) model behaves as a power function at low spent fuel pool operating temperatures. Where the Burley assumed an instantaneous iodine PF of 10, the Beahm model predicts iodine PFs between 45 and 34 at spent fuel pool temperatures of 100ºF and 115ºF respectively. An iodine PF of 10 results in a total effective iodine DF of 88. A plus or minus of 10% in the iodine PF (PF of 9 and 11) yields calculated iodine DFs of 67 (-24%) and 112 (27%) and respectively.

39

It is recommended that the FHA iodine DF model be updated with an iodine PF value consistent with Beahm, et al. (ORNL, 1992).

Observation 5: Applicability of the current DF model and Iodine Speciation.

The WCAP and the Burley reports assume the inorganic iodine released from the damaged fuel is in the form of elemental iodine is 99.75% and 0.25% organic iodine as vapor, I2. Current insights on fission product behavior indicate that the predominant chemical form would be CsI. However, CsI is less volatile than I2 and both are solid at FHA temperatures. Currently, RG 1.183 gives an updated iodine speciation of 99.85% elemental and 0.15% organic iodine. For 0.15% organic iodine, the limiting decontamination factor has a numerical value of 667. It is recommended that the FHA iodine DF model be at least updated with an iodine speciation consistent with RG 1.183 or with a new model that considers CSI.

On another note, TIA 99-03 (USNRC, 1999) reports on a review of certain assumptions of the FHA DBA. The staff discussed research performed by Collins et al. (1987) who reported on experiments to establish the amounts of fission products and the chemical species released from the gap of fuel rods irradiated in commercial reactors. The maximum test fuel temperature of 2200 F was selected to be less than that associated with fuel melting such that only the gap inventory would be measured. The test results indicated that neither elemental cesium nor elemental iodine was released from the gap.

Observation 6: Applicability of BWR fuel assemblies to PWR-based experiments.

The large-scale experimental bubble rise tests involved tubes with internal diameters comparable to those of current fuel rod designs for both PWR and BWRs. The large-scale experimental assembly was design with an inner diameter typical of Westinghouse PWR fuel pins of 0.405. A typical BWR fuel pin has an inner diameter of 0.357 (PNNL, 1979) and would produce smaller bubbles. As discussed above, the mass transfer is more efficient for smaller bubbles. Based on this analysis we see less decontamination at higher pressures, as seen in Figure 7 above, due to the effect of the shorter bubble rise time which outweighs any enhanced decontamination due to bubble size. If bubbles are indeed smaller at the higher pressure, the bubble-size trend would not match that from other experiments and modeling (e.g. suppression pool decontamination modeling):

greater pressure -> greater volumetric flow -> greater initial average bubble size and greater (or similar) average stable bubble size.

Thus, it is reasonable to apply an iodine DF base on PWR-type fuels to a BWR-type FHA.

On another note, TIA 99-03 (USNRC, 1999) discusses how the experiments are conservative simulations in that actual fuel rods contain pellets would significantly reduce the free cross-sectional area available for gas flow. In extended-burnup 40

fuel, expansion of the pellet further reduces the space between the pellet surface and the inner surface of the clad. As such, there would be resistance to the expansion of the gases in the gap region. Initial bubble sizes would be less than those observed in the tests and resulting DF would be larger.

Observation 7: WCAP test assembly versus action accident conditions.

The WCAP experiments demonstrated that severance of the fuel pins, when the bundle is in the vertical position, is more limiting than in the horizontal position. The vertical position experiments is analogous to a complete and instant shearing of all fuel pins resulting in an unobstructed ejection of fission product gases toward the pool water surface. This, in turn, reduces the buddle rise time through the water column and subsequent mass transfer of iodine from the gas phase to the water phase, resulting in lower iodine DFs and is thus considered conservative. Although fuel damage such as that modeled is possible, administrative controls over the movement of heavy loads and the configuration of structures and equipment in the path of irradiated fuel during movement reduce the likelihood of its occurring. A more likely situation would involve the dropping of a fuel assembly to an oblique or horizontal position in which the gas jet might not be upward-directed. Therefore, for the purposes of computing the FHA DBA radiological consequences, it is reasonable to assume the shorter bubble travel times reported in the WCAP maintains conservatism under actual conditions.

Observation 8: Iodine Re-evolution from the pool water The staff currently does not directly model iodine re-evolution form the pool. If the RG 1.183 iodine DF model were adjusted, it could be appropriate to include a re-evolution mode since they are two separate physical processes. The staff indicate the Beahm, et al. (ORNL, 1992) study could address this uncertainty with the updated iodine PFs.

At high pH, iodide, I-, is the preferred form of iodine whereas I2aq is at a very low concentration so it doesnt evolve. At low pH I2aq becomes dominant and the I-concentration is low so there is more available to evolve into the gas space. For instance, in the RG 1.183 LOCA analyses, CsI does not evolve into gas in the containment sump if a minimum pH>7 is maintained. However, if the pH is less, evolution of elemental iodine is assumed. Since the spent fuel pool pH generally ranges between 4 and 5, evolution of gaseous elemental iodine could be expected.

For modeling simplicity, the FHA chemical form of radioiodine released from the fuel to the spent fuel pool is assumed to be 95% CsI, 4.85% elemental iodine, and 0.15% organic iodide. As discussed above, the CsI released from the fuel is assumed to completely dissociate in the pool water but due to the low pH of the pool water, the iodine re-evolves as elemental iodine instantaneously.

On a side note, TIA 99-03 (USNRC, 1999) discusses how the thermal and hydraulic conditions in the spent fuel pool are substantially less energetic than those in a containment sump during LOCA 41

conditions. It is therefore reasonable to assume that much of this dissociated iodine would be retained by the pool and a re-evolution model would not be necessary from a dose-consequence perspective (though this has not been documented). Burley discusses a revolution of 0.1%-0.5%

per day however no reference is provided. Initial calculations by RES that assumed the full spent-fuel-pool size and a concentration dependence on the volatile iodine fractions an effective re-evolution was not too dissimilar to that reference by Burley. RES indicated additional work would be done in the area.

Conclusion of Re-analysis Results A detailed re-analysis of the iodine DF under DBA spent fuel conditions has been completed, where the scrubbing effect accounts for the iodine mass transfer processes from the gas bubble size to the surrounding interaction with the boric acid solution of the pool water. The WCAP and Burley studies were re-analyzed using modern data analysis tools to confirm results and conclusions. This new analysis developed a series of flexible iodine DF models based on information from the WCAP and Burley studies to compute iodine DF as a function of fuel rod pin pressure.

The new iodine DF models, found to fit a relatively simple analytical model, involving the most obvious mass transfer processes calculate iodine DFs consistent with the test data. The Burley theoretical models include bubble characteristics developed by the WCAP as well as the iodine PF under spent fuel conditions. The Burley models do not calculate iodine DFs consistent with the test data and are limited to an artificial upper numerical limit of 667, assuming 0.15% organic iodine. As previously discussed, this is due to the modeling assumption of assigning a factional non-soluble iodine species of 0.15% for organic iodine to the total iodine available for release from the fuel pin gap as well as a constant mass-transfer coefficient for all test pressures.

This study confirms and elaborates on the previous staff review in TIA 99-03. It was found that the preceding analyses applies to a wider range of conditions than that tested by Westinghouse or analyzed by Burley.

It is recommended a new FHA model be developed with the following considerations:

• Account for the reduction of the amount of radioactive iodine available to be released as gas due to limited I2 vapor pressure since I2 is solid at FHA temperatures.

• Assume initial gas pressure under FHA conditions:

• It would be more realistic to use the upper limit of gas generation models rather than assuming additional gas to reach limit. There is a sufficient understanding of fission gas generation.

• Consider a wait time (24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />) for less decay power and use spent fuel pool temperatures to determine gas temperature and thus pressure. In assuming higher pressures initially results in greater cooldown (decontamination) effects upon depressurization.

• To match WCAP observations, it is necessary to include models that account for the observed early decontamination.

42

• Use well referenced and data-backed mass transfer model applicable to situation:

• If reference cant be found for Burleys mass transfer coefficient and found to be applicable, switch to another model.

• Beahm, et al. (ORNL, 1992) or MELCOR iodine partition factors should be used for mass transfer evaluation.

• Uncertainties in measurement, scaling, and the bubble size distribution need to be evaluated for the WCAP-based model and this uncertainty considered in final recommendations.

• It should be verified that CO2 solution concentrations are not of a magnitude that would affect net mass transfer

• Consider external work on bubble size distribution:

• Bubble sizes (and distributions) for the WCAP-based model are not well known since they were back-calculated.

• Considered at in-depth in the nuclear industry for suppression pool decontamination modeling.

• Important for other fields.

• Re-evolution of iodine from pool should be considered.

• It would be good to be recalculated and documented.

• It would be good if summary report included technical basis and sample calculation.

43

4. ALTERNATIVE DESIGN-BASIS ACCIDENT FUEL HANDLING ACCIDENT MODEL 4.1 Summary The alternative FHA in-pool fission product transport model modifies the existing FHA boundary conditions. It is based on the environmental conditions in which fuel handling operations are taking place; incorporating several improvements of our current understanding of reactor fuel pin physics, iodine chemistry and re-evolution while maintaining conservatism.

The iodine gap activity in the damaged rods is assumed to be released in two stages. The chemical form of radioiodine released from the fuel to the pool water should be assumed to be 95 percent CsI, 4.85% I2, and 0.15% organic iodide. The first stage is the instantaneous gaseous release from the fuel gap in rising bubbles where I2 and organic iodine are conservatively assumed to be in vapor form and subsequently decontaminated by passage through the overlying pool of water into the building atmosphere. This activity is then vented to the environment over a 2-hour period. The second stage is the protracted release initiated 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> (following the initial gaseous release) following the fuel bundle drop. The CsI in the fuel gap of the damaged assembly is conservatively assumed to completely dissociate into the pool water then slowly re-evolve into the building atmosphere as I2 due to the low pool water pH. This activity is vented directly to the environment for a period of thirty days.

A case study was performed to analyze the radiological consequences of an alternative DBA FHA analysis methodology to Appendix B of Regulatory Guide 1.183. The purpose was to determine the impact of the proposed DBA FHA radiological consequences on the licensing basis FHA analyses and whether the control room, exclusion area boundary, and low population zone dose limit exceed the radiological accident dose criteria of 10 CFR 50.67. Results are intended to provide background information for staff and management. The computed radiological doses are generally 91-98% lower than the originally computed. The new model provides the same level of safety while providing some regulatory relief during fuel handling operations.

4.2 Behavior of Iodine under FHA Conditions The representative scenario for the design-basis FHA is that the most active bundle is dropped with rupture of all rods. This is assumed to occur near the minimum time allowed for fuel movement after shutdown and near the highest pool temperature allowed. During fuel movement, the pressure and temperature in the fuel rod gaps have decreased from those during reactor operation due to both the reduced power generation and lower coolant temperature.

Of initial interest regarding iodine behavior for FHA scenarios are the initial conditions of iodine during the postulated accident. This is specified as the isotope-specific fraction of iodine inventory that resides in the gap, the chemical form of this iodine, and its state. Gap fractions of iodine have typically been in the 5-10% range.

Current guidance recommends that the form of the iodine in the gap is predominantly, and perhaps nearly completely as, CsI with perhaps some fraction of molecular iodine (I2) (around 5%) and a small fraction (around 1/10th to 1/5th percent) as organic iodides, which is expected to be primarily methyl iodide, methyl iodide. The organic fraction is based on the results of 44

experiments conducted in the 1960s with supplemental thermodynamic calculations. The source of organic materials for organic iodide formation is organic impurities that remain in the fuel following manufacturing. It would not be surprising if modern fuel fabrication methods are substantially cleaner and contain far less organic contamination than the fuel used in these experiments. The proportion of CsI and I2 in the gap was not reviewed during this analysis.

Should Cs preferentially react with something else in the fuel (e.g. some additive used in the manufacturing process) over iodine and a sufficient amount of that material is present iodine could be left to form I2.

Near spent fuel pool temperatures, conservatively assumed for this analysis to be 150°F, CsI would be solid, I2 is solid, and methyl iodide as a liquid. Methyl iodide and molecular iodine have significant vapor pressures at this temperature. Lumped-parameter heat transfer analyses indicate gap temperatures of a few degrees higher than the pool for the decay power levels. The assessment of state (and fuel-clad gap pressure) should be based on natural-convection heat transfer analysis that considers the bulk pool temperature entering along the bottom and heated along the length of the rods.

Since the rods are heated internally, by fission during operation then by decay of fission products following shutdown, the clad is cooler with fission-product vapors preferentially depositing on it over the hotter fuel. Iodine can react chemically to bond with zirconium that it comes into contact with. Access to the clad can be blocked by other vapors that have previously deposited at higher temperatures. Both I-131 and other shorter and longer-lived iodine isotopes reach the gap. Most of the iodine in the gap consists of longer-lived iodine isotopes. Since longer-lived isotopes are present at higher concentrations they preferentially react with surfaces over I-131. Even in cases where I-131 or shorter-lived iodine isotopes interact with a surface, these atoms will most likely decay only to be replaced by another specie or by longer-lived iodine isotopes. It seems unlikely that I-131 will be substantially chemically attached to zirconium surfaces.

The considered iodine initial conditions for the FHA event are therefore mostly solid CsI, some I2 that is mostly inert or long-lived with a smaller quantity of radioactive iodine 131I and shorter-lived isotopes that is split between solid and gaseous forms. This depends on physical conditions (mainly temperature) and quantity, and a small amount of methyl iodide that is split between liquid and gaseous forms depending on physical conditions and quantity.

Following the postulated fuel rupture, the gases, liquids, and perhaps some of the solids are swept out of the rods. Solids that had previously deposited on the clad could potentially remain with the clad. Large solids or droplets completely deposit if not broken up and become trapped in the water. Smaller particles would also be substantially trapped.

The gap gases (helium and noble gases) rapidly cool as they expand. This cooling greatly reduces the vapor pressure of molecular iodine to temperatures at which nearly all of it is in solid form. The vapor pressure of methyl iodide is also greatly affected by the temperature drop. This cooling results in condensation of both molecular iodine and methyl iodide to small particles.

Small particles are substantially trapped both during bubble formation with additional trapping occurring during bubble rise. As the bubble rises to the surface, continued expansion further cools the bubble, but to a lesser extent than the initial expansion. This cooling is countered by 45

being heated by the warmer water and by the evolution of water vapor into the bubble. Nearly all of any molecular iodine that remains gaseous and a moderate fraction of any methyl iodide that remains gaseous will get trapped in the pool. The amount of gaseous iodine that is trapped by the pool can be estimated by the bubble rise time using gas phase mass transfer relations. The amount of initial trapping of particles and gases depend on bubble dynamics. Factors that affect this initial trapping include the bubble size distribution and the bubble rise time.

Note: For the purpose of the analysis methyl iodide has been assumed to be completely gaseous and assumed to not get trapped in the pool even though it is a liquid at these temperatures.

Iodine present in the gap is expected to end up in the pool. It is assumed to occur over the first two-hour period for modeling simplicity. Cesium iodide, the predominant expected form of iodine in the gap, is not only hygroscopic (readily absorbs water even from atmosphere) but also deliquescent (absorbs water vapor from the atmosphere to the extent that it forms a solution).

Most other forms of iodine also readily dissolve in water and many are hygroscopic and deliquescent. It would be expected that any iodine-containing particles that get trapped in the pool would readily dissolved. If liquid water penetrates the fuel-clad gap after rupture, which could be expected, solid CsI and I2 remaining on clad/fuel surfaces in the gap would be likewise readily dissolved. Even if only water vapor penetrates into some parts of the fuel-clad gap a solution will form, the solid CsI and I2 would dissolve, and the iodine still would be expected to eventually end up in the pool if it does not decay first. Whether substantial iodine originally in the fuel pin gap is in solid form transfers to pool before decay or before the end of the calculation period depends on the rate of release.

The iodine originally in the gap may not be the only source of iodine to the pool. High burnup fuels can fragment into small particles and these fragments can be expelled upon fuel rupture.

The small particles have large surface to volume ratios, which would result in much faster leaching of radionuclides than from intact fuel. It is not expected that leaching rates are such that transfer of iodine to the pool would be significant before I-131 decay even with the high surface to volume ratios. Parenthetically, radiation-enhanced leaching models developed for geological repositories can be used to verify whether leaching rates may be significant.

Inorganic iodine that gets trapped in the pool establishes equilibrium between different species depending on pH. The primary species are I- at high pH or low iodine concentrations and I2 at low pH and higher concentrations. Other iodine species exist but at lower concentrations. Iodine in I2 form can evolve into the gas above the pool where it can transport to, and contribute to doses at, other locations.

Volatile organic iodide species such as methyl iodide trapped in the pool likewise evolve to the gas above the pool. Due to the lower solubility in water this evolution occurs at a greater rate than that of molecular iodine.

Iodine can also adsorb on surfaces. Different iodine forms adsorb on different materials at different rates depending on pH. Some adsorption is reversible, and some adsorption is irreversible due to chemical reactions with surfaces. The adsorption can depend on surface history (e.g. oxidation or deposition of other substances). In some cases, normally irreversible 46

iodine adsorption can be reversed by radiation. Adsorption of iodine on surfaces was not considered during this analysis.

The presence of ionizing radiation affects the behavior of iodine in pools. One of the ways is by iodine reacting with water radiolysis products to convert some I- to I2 leading to the presence of a small amount of I2 even at low concentrations and/or elevated pH. Ionizing radiation also affects iodine behavior in pools by the formation of organic iodides. Radiation converts some I2 to methyl iodide in the presence of methane and also decomposes methyl iodide. Methane is not expected persist long in pools since it is volatile and not very soluble.

Longer chain organic substances can remain in pools for longer so they can be present to form organic iodides in the presence of ionizing radiation since they generally have lower vapor pressures and thus do not evolve as rapidly. Similarly, longer-chain organic iodides are less volatile and thus are not expected to evolve as rapidly and contribute significantly to offsite dose unless they somehow convert to more volatile forms. Experiments involving the irradiation of organic substances in aqueous solutions typically result in evolution of organic iodides indicating either the presence of or conversion to volatile organic forms.

To convert inorganic iodine to organic forms, some source of organic material must be present.

A low concentration of organic substances exists in pools by equilibrium with the steady-state organic concentration in air.

Boraflex, a neutron poison used in spent fuel pools, is an organic-containing material that is composed of 46% silica, 4% polydimethyl siloxane polymer, and 50% boron carbide, by weight.

(USNRC, 2010) Boraflex degrades by radiation in spent fuel pools until an equilibrium silica concentration develops with the pool water. Some organic concentration of polydimethyl siloxane and organic radiolytic reaction products likely also develop unless organic substances are filtered out from the pool water. Iodine will likely react with these organics in the pool in the presence of radiation to form organic iodides that can evolve from the pool surface.

4.3 Iodine Re-Evolution calculations Two sets of iodine re-evolution calculations were performed: The first set consists of numerical calculations that account for the change in the volatile iodine fraction based on the change in concentration. The second, simplified set of calculations, neglects this change so the evolution removal coefficient remains constant. This simplified method results in a simple exact transient solution. This second approach is recommended for estimating evolution from spent fuel pools.

The second set of calculations also considers the effect of potential filtration and illustrates how other removal mechanisms can be considered in an analysis.

The calculation method is detailed so the approach can be easily reproduced.

The example below is a simplified example calculation that uses nominal isotopic inventories, pool dimensions, temperatures, and decay time. Plant-specific inventories and conditions should be used for any licensing calculation.

Assumptions 47

The calculations involve the following assumptions:

• 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of decay assumed for radioactive inventory.

• Assuming a 23 feet (7m) bundle rupture depth.

• Assuming a 35 feet (10.7m) pool depth.

• Pool surface dimensions of 29.5 feet x 39.4 feet (9m x 12m)

• The iodine volatile fraction was estimated using a model from Beahm, et al. (ORNL, 1992).

• Neglecting the buildup of iodine in gas such as would occur with a high ventilation rate to the outside environment.

• Iodine is assumed to be evenly distributed in the water.

• Assumes a high initial DF for I2 (i.e. assumes all iodine is initially trapped in pool).

• Assumes no initial decontamination for methyl iodide.

• Assumes all iodine in gap ends up instantaneously in pool following rupture.

• Neglects iodine adsorption of on surfaces.

• Neglects radiolytic increase of I2 concentration.

• Neglects conversion of iodine to methyl iodide in the presence of organics in the water.

• Assumes a constant water level.

• The evaluation of the volatile fraction uses reaction rates valid at 25oC rather than actual temperature and does not account for the potential for iodate (IO3-) on the assumption that the concentration is negligible. 8 The calculation considered the speciation of iodine that reduces the fraction available for release to the environment, and radioactive decay. It does not consider potential surface deposition of iodine, organic iodide formation with organics present in water, or filtration of the pool water.

The transfer of iodine from the pool water is evaluated using a mass transfer coefficient, K. The depletion solely due to evolution to the gas would be:

131,

= 131, . (Equation 4-1) where NI131,P represents the number of moles of I-131 atoms in the pool, KL the liquid-phase-based overall mass transfer coefficient, XIV the fraction of I-131 that is in I2 form and thus volatile, and Vpool the volume of the pool.

The other considered depletion mechanism for I-131 in a pool is radioactive decay. The decay constant for I-131, , is quite close to 1E-6 s-1. The depletion of I-131 in the pool by both evolution and radioactive decay can be evaluated as:

8 Future calculations could probably account for temperature differences and the presence IO3- (if not also other iodine-containing species) and ensure that radiolytic I2 formation are not underpredicted).

Considering surface deposition in the calculation while accounting for the possibility that radiation can reverse deposition.

48

131,

= 131, 131, . (Equation 4-2) with the amount of iodine that has been evolved to the gas evaluated using the following coupled equation:

131,

= 131, 131, . (Equation 4-3)

Evolution is evaluated until nearly all the radioiodine has decayed.

The iodine volatile fraction, XIV, depends on the total iodine in the pool including non-radioactive iodine. This value can change if the iodine concentration in the pool changes.

Evaluation of iodine volatile fraction The I2 fraction in the pool depends on pH and the iodine concentration in the pool. To determine concentration both the amount of iodine released into the pool and the volume of the pool is needed. Figure 8 shows the concentration trends with varying pH using the model for volatile fraction from Beahm, et al. (ORNL, 1992). Figure 9 shows the same behavior on a log-fractions scale.

At high pH and at low iodine concentrations, most of the iodine in the pool is as I- and very little of the iodine is volatile. At higher iodine concentrations and low pH, a much larger fraction of the iodine is in I2 form and thus volatile and can evolve from the pool. This model was tested against data for irradiated solutions including the irradiated water test cases presented by Beahm, et al.

(ORNL, 1992). The model exhibited reasonable agreement at higher volatile fractions but seemed to miss some of the increase in I2 concentration due to radiolytic reactions at higher pH and lower concentrations. The radiolytic contribution seems to set a minimum volatile fraction based on dose rate and I2 concentration, neglecting the effects of any other substances in the pool.

49

Volatile I fraction 1

0.9 0.8 0.7 0.6 I fraction as I2 0.5 0.4 0.3 0.2 0.1 0

2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 pH @ 298 K (25oC,77oF) 1E-4 M I 1E-5 M I 1E-6 M I 1E-7 M I 1E-8 M I 1E-9 M I 1E-10 M I Figure 8 Fraction of I atoms that reside in I2 form, derived from equations from Beahm, et al.

(ORNL, 1992) 50

Volatile I fraction 1.E+00 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 1.E-01 1.E-02 I fraction as I2 1.E-03 1.E-04 1.E-05 1.E-06 pH @ 298 K (25oC,77oF) 1E-4 M I 1E-5 M I 1E-6 M I 1E-7 M I 1E-8 M I 1E-9 M I 1E-10 M I Figure 9 Fraction of I atoms that reside in I2 form, derived from equations from Beahm, et al.

(ORNL, 1992), log scale.

The chemical reaction set provided by Beahm, et al. (ORNL, 1992) can be used to describe the concentration- and pH-dependent equilibrium Iodine speciation that governs the volatile iodine fraction. The determined equilibrium speciation satisfies the following equation:

[I2] / [I-]2 = [H+]2 / (D + E * [H+]) (Equation 4-4)

Where:

[I2] represents the molecular iodine concentration in the solution (pool),

[I-] represents the iodide ion concentration in the pool,

[H+] represents the hydrogen ion concentration evaluated by

[H+] = 10-pH, where pH is known.

Both D and E are algebraic combinations of equilibrium constants that result from combining the different equilibrium equations into the above equation. These constants have the following values:

51

D: (6.06 +/- 1.83) x 10-14 E: 1.47 x 10-9 The equilibrium constants are valid at 25 oC and apply to concentrations specified in moles/liter.

The value used for D was obtained by directly evaluating the algebraic relation using Beahm, et al. (ORNL, 1992) equilibrium constants. The value differs from that in Beahm (6.05) by one digit (it is not a typo).

The reactions provided by Beahm consider that the equilibrium I2 and I- concentrations make up the total concentration of iodine atoms in the solution in any form, [Ie]:

[Ie] = 2 [I2] + [I-] (Equation 4-5)

This relation is used in the determination of iodine volatility for the FHA revolatization/re-evolution calculation. This simplifies the evaluation of the volatile iodine fraction to a quadratic equation once pH and total iodine in the pool is specified. There are some caveats on the validity of this relation which are discussed below.

Beahm, et al. (ORNL, 1992) explicitly states that below concentrations of 1e-6 mol of I atoms /

liter of solution, the iodine (IO3-) concentration may not be small enough to be neglected. The amount of iodine that is expected to end up in the in the pool for an FHA scenario can result in iodine concentrations less than this value.

The reaction set used by Beahm, et al. (ORNL, 1992) also considers reactions involving other iodine forms, HIO and IO-. Neglecting these other forms for evolution calculations implicitly assumes that the concentrations of these other iodine species are substantially less than the sum of the I2 and I- concentrations at pH and I concentrations. To verify this assumption, the equilibrium HIO and IO- concentrations were calculated using the Beahm equations for the iodine-concentration and pH ranges used in the above plots. At their most significant, the iodine in HIO and IO- made up 3/10,000 (or 0.03%) of the iodine by mole as determined using the reaction set and rates in Beahm. This confirms that their contribution is negligible for this reaction set.

The removal of these other reactions simplifies the problem solution.

If other species concentrations (e.g. IO3-) are significant, the volatile fraction should decrease so this simplifying assumption seems conservative (but also could be insignificant).

Although the reaction set used by Beahm, et al. (ORNL, 1992) should capture the major phenomena, the international severe accident community has since conducted substantial research into iodine chemistry and developed many different reaction sets, some of which may be too detailed for the FHA scenario. A simplified reaction set derived from this research would provide better accuracy of iodine behavior and would provide results that are valid under other conditions including validity at temperatures other than 25 oC and validity at different concentrations by accounting for the fact that chemical activities can deviate from proportionality with concentration.

52

Since the volatile I fraction depends on I concentration in the pool, the amount of iodine in the pool is needed to evaluate releases.

Additional detail is provided on assumptions relating to fraction evolved:

The total iodine that contributes to the speciation and thus the evolution rate depends not only on the radioactive iodine but also the non-radioactive iodine in the pool. Since the non-radioactive iodine makes up most of the available iodine, and that this fraction becomes more significant as the radioactive iodine decays, the evolution rate of radioactive iodine depends primarily on the non-radioactive iodine concentration.

The calculation considers that radioactive iodine in the pool can be depleted by both loss to gas and by radioactive decay and that non-radioactive iodine can be depleted only by loss to gas.

The radioactive iodine inventory 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> post shutdown is considered as the initial conditions. To simplify the calculation the 24-hour-decayed I-131 inventory is increased by 44.34% to account for the activity of the shorter-lived iodine isotopes at the postulated time of the FHA. Nominal bundle inventories from the WCAP report were used for illustrative purposes. An actual licensing calculation should use reactor- specific inventories, pool volumes, and shutdown times.

The entirety of the bundle gap iodine that does not immediately escape following rupture is considered to be immediately and uniformly distributed within the spent fuel pool water.

For the instantaneous release, we consider what radionuclides might immediately be released from the gap and which could possibly make it to the gas space above the pool in the event that fuel rods break: historically, Iodine has primarily been considered to be in the gaseous iodine form I2 or methyl iodide for the FHA. This amount of iodine considered available for release by evolution is significantly greater than that considered available for immediate release following rupture. Most of the iodine in the gap is expected to solid (primarily as CsI and as I2). While solids deposited on the clad and fuel may not come out during the initial release following fuel rupture they could dissolve in the pool once exposed to the water. Assumptions that represent expected behavior for instantaneous release for an accident scenario are not necessarily representative for calculating re-evolution release for the same scenario.

The rupture of a fuel rod can also release solid I2 or CsI if loosely bound to surfaces. Solids could possibly be tightly bound to surfaces (in case which they dont come out), loosely bound to surfaces (which come off as large pieces and could potentially break when ejected), and loose smaller particles. Possibilities are simply being listed here without detailed knowledge about behavior. If loosely bound a small fraction of small particulates could potentially make it to the surface large particles would almost-certainly be trapped in the water unless they can substantially be broken up upon release. If the solids adhere well to clad or fuel they do not get released during the initial rupture. As mentioned in the 473rd Advisory Committee on Reactor Safeguards (ACRS) meeting on June 7, 2000 on the review of Source Term guidance for fuel handling accidents, CsI (solid at pool temperatures) is the expected dominant form of iodine in the fuel-clad gap since the formation of molecular iodine (I2) was expected to be formed in the Reactor Coolant System (RCS).

53

Should Cs preferentially react with something else in the fuel (e.g. some additive used in the manufacturing process) over iodine and there exists enough of the other material iodine could be left by itself to form I2.

One of the points Steve LaVie made during the ACRS meeting was that the chosen DF of 200 represented not only the iodine immediate released following the postulated fuel rupture but also represented re-evolution of iodine. This re-evolution release was assigned to the immediate release for the purposes of determining the decontamination factor for simplicity.

The release mechanism for re-evolution differs from the instantaneous gaseous release. The iodine that can be released by re-evolution from the pool is the iodine that must first be dissolved in the pool. The relevant question then becomes: How much iodine gets dissolved in the pool?

In estimating the amount dissolves one naturally considers how readily a substance dissolves.

Cesium iodide is not only hygroscopic (readily absorbs water even from atmosphere) but also deliquescent (absorbs water vapor from the atmosphere to the extent that it forms a solution).

We have seen this in experiments: When we put CsI particles on a dry surface in a humid environment we end up with droplets of solution. Most other forms of I also readily dissolve in water. It would be expected that any iodine-containing particles that get trapped in the pool would readily dissolve. If liquid water penetrates the fuel-clad gap after rupture, which could be expected, solid CsI and I2 remaining on clad/fuel surfaces in the gap would be likewise readily dissolve. Even if only water vapor penetrates into some parts of the fuel-clad gap a solution will form, the solid CsI and I2 would dissolve, and the iodine still would be expected to eventually end up in the pool if it does not decay first. Whether substantial iodine originally in the gap in solid form transfers to pool before decay or before the end of the calculation period depends on the rate of release.

The definition of gap inventory/gap fraction for use for FHA was based on the consideration of radionuclides that are available for immediate release and consists of radionuclides that have migrated to the fuel-clad gap. (USNRC, 1972). This Guide also considered subsequent release from evolution to be less significant. The immediate release of radionuclides was considered to be the largest contributor to potential dose at the time with evolution contributing 10-20% of overall release. It is clear that dissolution of iodine-containing species and subsequent re-evolution of iodine from the pool was not a major consideration at the time of the writing of this Guide.

Subsequent research has shown that re-evolution can be significant primarily because both the instantaneous I2 decontamination fraction has increased and because of the consideration that less of the iodine in the gap is available for immediate release since most of the iodine in the gap is expected to be solid (i.e., CsI).

Should a spent fuel rod break under water, not only would it be expected that the soluble contents in the gap be released to the pool, water could also leach inventory that is currently not considered part of the gap inventory.

It is unknown how much would be released by leaching during the time considered for the event.

It is expected that leaching of radionuclides from intact fuel pellets would be slow relative to releases from the gap. High burnup fuels fragment and these fragments can be expelled upon rupture. Experiments on leaching from fuel could be used to estimate the rate.

54

Although data may be available to assess these release rates the NRC has not currently collected the information necessary to estimate them for evaluating FHA conditions. Data of the following type may provide information in this area: The leaching of radionuclides from irradiated fuel has been studied for long-term spent fuel storage. These studies focused on longer-lived radionuclides. Interest into radionuclide leaching has increased as part of the cleanup effort for the Fukushima accidents resulting in new research programs. Previous experiments/accidents may also provide useful data on release rates and fractions and could give us an indication of whether significant release of I-131 to the pool from irradiated fuel is likely before decay.

The radioactive decay constant for iodine-131 was used: 1 x 10-6 s-1.

An organic iodine fraction of 0.0015 was used which is consistent with current guidance.

It is unclear what the basis was for changing this fraction from 0.0025 which was based on Westinghouse data on organic content remaining from the fuel manufacturing process.

The initial choice of 0.25% was estimated by Burley based on thermodynamic calculations and were checked against experiments. Early data showed considerable scatter on the organic iodide formation from fuel. The organic iodide fraction in the gap could not be measured directly so was measured in small vessels following release from fuel or simulants. The organic fraction increased over time in the vessels. Although the initial organic fraction was difficult to estimate the results indicated that the initial gap concentration was likely low and that most of the organic iodides were formed primarily in the vessels after release. The uncertainty due to data scatter led to different choices of what fraction to assume for analyses.

A pH of 4 was assumed.

The pH was assumed to be the lowest within an expected range to conservatively estimate the volatile iodine fraction.

An example of the evaluation of the iodine volatile fraction is provided in the section titled 4.3.5,Simplified Calculation with concentration-dependent speciation, below.

Evaluation of the Pool Mass Transfer Coefficient This section provides an estimate of the Mass Transfer Coefficient (MTC) based on the THAI experiments (Yuill, 1970). It involves a model-based MTC that accounts for iodine transport both in the liquid and gas phases assuming the gas-phase iodine concentration is negligible. This approximates the transfer rate that would result if the building has a high ventilation rate that clears the gas. The partition coefficient of I2, the ratio of the I2 concentration in the liquid to that in the gas, substantially affects the transfer rate.

The previous re-evolution work by Yuill et al., uses a correlation for the rate of evaporation of water and organic liquids into air from plane surfaces located parallel to the direction of flow. The mass transfer coefficient derived using this approach from pools differs from the overall MTC in pools, which is governed by the water-phase mass transport rate.

The THAI-23 experiment was conducted for the very purpose of evaluating the MTC of iodine from pools. The experiments are intended to bound the possible iodine evolution scenarios by considering both stagnant and recirculating conditions. The Yuill MTC correlation was based on 55

an assumed gas flow rate. This gas-phase-only approach, although it provides a reasonable first estimate in absence of more-directly applicable data and modeling, does not directly apply to the FHA scenario. The use of an MTC with this approach for the FHA scenario also requires a somewhat arbitrary assumption of gas velocity over the pool.

Fischer et al. (Fischer, 2012), evaluated the mass transfer coefficient for iodine in pools based on THAI experiment Iod-23. The authors provide two liquid-phase mass transfer coefficients for iodine: 3.5 x 10-7 m/s for a stagnant pool and 3.6 x 10-5 m/s for a convecting pool.

Initial calculations were run with both of these MTCs on the expectation that they should bound those of source-term or severe-accident conditions. The spent-fuel-pool-specific MTC was used after it was evaluated.

Spent fuel pool conditions under FHA design basis calculations are expected to be closer to the stagnant conditions than the recirculating conditions. Some pool recirculation is expected due to natural circulation resulting from the decay heat from the fuel.

The authors determined the liquid-phase mass transfer coefficients for the THAI test using a surface renewal model as follows:

= , (Equation 4-5) where kL is the liquid pass mass transfer coefficient, DIL the diffusivity of I2 in the liquid and tc the contact time between the liquid and the gas surface.

The overall mass transfer coefficient based on the liquid phase is then:

1 1 ,

= + (Equation 4-6) where KL (upper case K) is the overall mass transfer coefficient based on the liquid phase, kG is the gas phase mass transfer coefficient, and Pc is the partition coefficient which describes the ratio of the I2 concentration in the liquid to the I2 concentration in the gas at equilibrium conditions.

To evaluate the liquid phase mass transfer coefficient for the spent fuel pool in the same manner two things were needed, the I2 diffusivity in water, and the contact time. An approach based on natural convection using decay power and fuel dimensions was also considered but insufficient data or CFD calculations with surface velocity were found to check against.

The diffusivity used for I2 in water was 2.79E-9 m2/s, the same as that used by Fischer. Although CFD calculations showing surface velocities showing pool liquid surface velocities were not found, some calculations were found of other pool flows. Analyses showed localized maximum velocities on the order of 0.1 m/s. The velocities in most of the pool were substantially lower than this value.

This value was then taken as the representative surface velocity in the absence of information on surface flows from which contact time could be estimated. Considering symmetric recirculation normal to the short length of the pool, the water would rise in the center after being heated by the pool and then flow with the assumed velocity to the wall. The distance the water travels on the 56

surface is then 9 m / 2 = 4.5 m. At a constant rate of 0.1 m/s the liquid-gas contact time is 45 s.

The considered liquid phase mass transfer coefficient is then:

2.799

= = 4.44 6 / (Equation 4-7) 45 (7.87E-6 m/s if no pi which likely was included from the circular geometry in Fischer)

This approach to contact time was assumed in the absence of data or calculations on surface flow field. This value is likely conservative for a few reasons: (1) The surface velocities are likely lower than this value. The 0.1 m/s in the different analyses were localized and much greater than in the rest of the pool. (2) The velocity on the surface is not constant throughout travel. It likely accelerates sideways once it reaches the surface to its maximum then decelerates upon approaching the wall. (3) Recirculation would likely preferentially occur in the longer rather than the shorter pool dimension. All these effects increases the total surface liquid-gas contact time which decreases the mass transport coefficient. On the other hand, recirculation from turbulence could increase transfer to the surface. This is not expected to be a significant effect.

For the gas phase MTC the approach used by both Burley and Yuill was used to obtain a gas-phase mass transfer coefficient, kG, of 5.88E-4 m/s assuming a gas velocity of 0.1 m/s.

A partition coefficient of 28.5 (cI2L/cI2G at equilibrium) was used based on a relation from Fogg (Fogg, 2003). The Fogg model had a near-median value of different models evaluated at the pool temperature, 322.039 K (120 F). Values from the different models ranged from 25.8 to 31.0.

From these parameters, a combined liquid-phase-based overall mass transfer coefficient, KL, of 3.66E-6 m/s was evaluated.

It was shown that combining both mass transfer coefficients reduces the mass transfer coefficient (they combine in a manner similar to resistors in series) relative to either phase and that using the liquid phase mass transfer coefficient is a close but still conservative alternative to the overall mass transfer coefficient. Using the liquid only mass transfer coefficient bounds the mass transfer coefficients obtained when using the same gas velocity as used by Burley (50 fps, 15.24 m/s, 34 mph).

The recirculation system, if operating, also contributes to recirculation. Characteristics for the surface flow induced by the recirculating system were not evaluated.

Calculation with concentration-dependent volatile fraction The pool was assumed to have a length of 12.0 m (39.5 ft), a width of 9.0 m (29.5 ft), and a depth of 10.7 m (35.0 ft). This results in a pool surface area of 108 m2 (1162 ft2), a volume of 1152 m3 (40,683 ft3), and a pool-gas surface to volume ratio of 9.374 x 10-2 m-1 (2.857 x 10-2 ft-1).

The results of the calculation considering the change in the iodine volatile fraction are shown in Figure 10. The calculation was performed numerically (simple explicit) with 64 loguniform time steps from 10 to 1e7 s, or 116 days. The plot starts at 10,000 s, or roughly 2.8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />. The 116 57

days is sufficient time for essentially all the I-131 to decay. The results are not much changed from those at 30 days (~2.6E6 s).

The plot contains 4 curves. The blue curve labeled In pool, shows the fraction of I-131 remaining in the pool. The gray curve labeled Decayed, shows the fraction of I-131 iodine that decayed before evolving. The orange curve labeled Evo x 100, shows the evolved fraction multiplied by 100 so that it shows on the figure. This is also the evolution release fraction, RF. Finally the yellow curve shows the pool evolution DF (= 1/RF). Nearly all the iodine decays before leaving the pool. Only a small fraction, about 0.5%, is predicted to evolve before decay, resulting in a pool evolution DF of 202. The release occurs over a long duration.

Figure 11 shows a similar plot but applies a stagnant-pool liquid-phase mass transfer coefficient (1E-7 m/s). This plot looks somewhat similar but the evolution curve is multiplied by 1000 so the evaluation RF is 0.47% so the evolution DF is almost 2100.

Radioactive I disposition 1 1.0E+06 0.9 1.0E+05 0.8 0.7 1.0E+04 0.6 Fraction 0.5 1.0E+03 DF 0.4 1.0E+02 0.3 0.2 1.0E+01 0.1 0 1.0E+00 1.0E+04 1.0E+05 1.0E+06 1.0E+07 Time (s)

In pool Evo x 100 Decayed Revol DF Figure 10 Iodine disposition in pool for the numerical evolution calculation 58

Radioactive I disposition, stagnant pool 1 1.0E+06 0.9 1.0E+05 0.8 0.7 1.0E+04 0.6 Fraction 0.5 1.0E+03 DF 0.4 1.0E+02 0.3 0.2 1.0E+01 0.1 0 1.0E+00 1.0E+04 1.0E+05 1.0E+06 1.0E+07 Time (s)

In pool Vol x 1000 Decayed Revol DF Figure 11 Iodine disposition for a numerical evolution calculation assuming stagnant conditions.

Simplified Calculation with concentration-dependent speciation This simplified calculation is a similar to the above calculation except that it uses a constant evolution rate which makes it easy to solve analytically. The evolution rate depends on the total iodine (radioactive + non-radioactive) in the pool. Since the evolution rate is relatively low and most of the iodine in the pool is non-radioactive and thus does not decay, the overall change in evolution rate does not change significantly over the course of the scenario.

The calculation process is shown in detail.

This calculation also evaluates the possibility of filtration by a recirculation system. A similar approach can be used to evaluate removal by other mechanisms.

This method uses the exact solution to the equations rather than a numerical approach.

This simplified approach is the recommended calculation approach for iodine re-evolution.

The following information is needed:

• Vpool - spent fuel pool volume

• Spool - spent fuel pool surface area

• Qrecirc - volumetric flow of recirculation system (to evaluate effects of filtration) 59

• F - Overall recirculation filter efficiency for iodine - Expected to be ~ 1 (to evaluate the effects of filtration)

• NI131gap - bundle radioactive iodine in gap (moles)

• NI129gap - bundle non-radioactive iodine in gap (moles)

• KL = mass transfer coefficient - 3.66E-6 m/s

• pH - acidity of pool Note: Vpool, Spool, KL, and Qrecirc must use consistent units. (for the purpose of calculating concentrations in M (moles/liter) Vpool must be converted to liters)

Calculation Sequence:

The Calculation sequence is as follows:

1 - Calculate amount of iodine (radioactive and non-radioactive) in gap 2 - Calculate volatile iodine fraction in pool 3 - Calculate removal coefficients 4 - Evaluate release, either

• overall release (neglecting time), or

• time-dependent release 4.3.3.1 Amount of iodine in pool Both the radioactive and non-radioactive iodine in the pool affect the radioactive iodine evolution.

The calculations operate on moles so iodine isotopes quantities must be converted to moles.

For a given mass of iodine the number of moles of iodine can be calculated from the mass, m, in grams and its atomic weight:

Alternatively, for radioactive materials the number of moles can be calculated from the activity in Becquerels (Bq):

131 131 = (Equation 4-10) 131 .

Activities in Curies must be converted to Becquerel.

1 Ci = 3.7 x 1010 Bq The radioactive iodine concentration can be decayed accounting for time before fuel movement.

If this is done, the activity of other iodine isotopes at this time should be added to the I-131 activity.

In the calculation above (assumed 100 hr decay before fuel movement) the other isotopes contributed an additional 4 percent.

60

4.3.3.2 Evaluation of the volatile iodine fraction Determine the fraction of I atoms in the pool that are in I2 (volatile) form.

• Calculate radioactive and total concentrations in pool by:

o Cr = concentration (M) (moles I atoms /L) of radioactive I atoms = NI-131gap / Vpool o Ct = total I concentration (M) (moles I atoms /L) = ( NI129gap + NI-131gap) / Vpool Note: Vpool must be converted to liters to calculate concentrations in moles / liter.

• Calculate the H+ concentration:

o Ch = [H+] = 10-pH

• Calculate the [I2] / [I-]2 concentration ratio, Ri 9:

o Ri = [I2] / [I-]2 = Ch2 / ( 6.0603E-14 + 1.4708E-09 Ch)

• Calculate the fraction of I atoms in I2 form:

o First evaluate Bm (Negative B for quadratic equation below)

Bm = 4 Ct + 1 / Ri o Then evaluate the volatile fraction, Xe (fraction of I atoms in I2 form):

Xe = ( Bm - 2 16 2 ) / ( 4 C )

t 4.3.3.3 Determination of removal coefficients.

Three removal coefficients pertain to 3 removal mechanisms: radioactive decay, evolution, and filtration in a recirculation system.

The radioactive decay removal coefficient, r, is the common one used in the radioactive decay equation:

r = I-131 = 1E-6 s-1 The evolution removal coefficient, e, is calculated using the mass transfer coefficient, the pool surface-to-volume ratio, and the fraction of I that is in I2 form.

e = KL Xe Spool / Vpool 9 Combined Speciation Rate from Beahm, et. al. Iodine Evolution and pH Control (NUREG/CR-5950) 61

The removal rate is reduced to account for the fraction of iodine that is volatile and thus available to evolve to the gas space.

This evolution rate applies to both non-radioactive and radioactive iodine.

In the numerical FHA evolution calculation above, e was determined to be 5E-9 s-1 (0.000432 /

day). For the filtration scenarios the non-radioactive iodine will also be rapidly depleting lowering the volatile fraction (the assumption of constant evolution removal rate does not hold true). These cases involve relatively rapid removal of iodine relative to others so it is not considered necessary to evaluate the decrease in evolution rate throughout the scenario.

Note: The daily fractional release rates are e converted from s-1 to day-1 by multiplying by 86400 s / day.

Initial calculations showed that depletion of non-radioactive iodine is slow so that the volatile fraction and evolution removal rate does not change significantly. As an indication of the effect of not accounting for the change in removal coefficient: by the end of a 30-day period the pool removal coefficient had dropped to 4.9E-8 s-1 from an original 5.0E-8 s-1. After the full simulation time (116 days) the removal coefficient had dropped to 4.7E-8 s-1. Almost all of the decay is complete by 30 days. The small decrease over the radioactive decay period is not significant.

For this calculation Spool and Vpool units must be consistent.

The filtration removal coefficient, f, is calculated using the recirculation system volumetric flow, Qrecirc, the volume of the pool, and filtration efficiency, F:

f = F Qrecirc / Vpool If no recirculation is considered, f = 0 (or simply not included in the calculation).

For this calculation Qrecirc and Vpool units must be consistent.

If fractional recirculation (Qrecirc/Vpool) is in the range of 2day-1 (2.3E-5 s-1) and filter efficiency, F, is approximately 1 the equation results in the following filtration removal coefficient:

f = 2.3E-5 s-1 The time units for all removal coefficients must be consistent.

4.3.3.4 Evaluate release In evaluating releases, only the radioactive iodine concentration is considered.

Two methods of evaluating release are shown, a transient release and an overall release. Both the transient and overall releases are equivalent if the transient is evaluated to full radioactive iodine depletion. The only reason to use the overall release over the transient release is to simplify the calculation by assuming it all gets released at once. This may be desired for simplicity if recirculation filtration is credited.

Overall (asymptotic) release 62

This evaluates the overall release fractions, RFs, of radioactive iodine after all the radioactive iodine has been depleted from the pool by either radioactive decay or recirculation system filtration.

The evolution RF (of the amount of iodine that makes it to the pool) is simply the ratio of evolution to total lambdas:

RFe = e / (e + r + f) = e / (e + 16 + 2.35)

For initial FHA example case this becomes:

The example case is evaluated for both not-crediting and crediting recirculation system filtration using both the stagnant and convective pool mass transfer coefficients.

Not crediting recirculation and filtration:

RFe,calculated = 5.00E-9 s-1 / ( 5.00E-9 s-1 + 1E-6 s-1) = 4.98E-3 s-1 (an evolution DF of 201)

Considering a variable speciation and depletion rate results in a DF of 42 for this case. The similarity of the numbers confirms the reasonability of the constant release rate assumption since there should be less deviation for a lower MTC. There may be a larger variation for the filter cases but I-131 releases for those scenarios are low.

RFe,stagnant = 4.78E-10 s-1 / ( 4.78E-10 s-1 + 1E-6 s-1) = 4.78E-4 s-1 (an evolution DF of 2,092)

Crediting recirculation and filtration:

RFe,calculated,recirc = 5.00E-9 s-1 / ( 5.00E-9 s-1 + 1E-6 s-1 + 2.3E-5 s-1) = 2.08E-4 s-1 (an evolution DF of about 4,800)

RFe,stagnant,recirc = 4.78E-10 s-1 / ( 4.78E-10 s-1 + 1E-6 s-1 + 2.3E-5 s-1) = 1.99E-5 s-1 (an evolution DF of 50,000)

Transient release The removal coefficients can be used in RADTRAD by one of two ways:

1. Using a volume in RADTRAD that represents the pool and flow rate that represents evolution to the refuel floor with a volumetric evolution rate such that flow to the refuel floor is as follows:

a. Qe = e Vpool

b. f is used if recirculation filtration is credited

i. Alternatively, a loop and filter can be modeled in RADTRAD instead of using f. It will return the same result either way.

c. r = 0 since RADTRAD already calculates radioactive decay.

d. In this calculation RADTRAD calculates the transient depletion of iodine in the pool.

63

2. Using the exact release equation to evaluate releases to the RADTRAD refuel floor or environment volume:

a. In this case Qe and f are used as in the pool-volume approach.

b. r = 1E-6 s^-1 since RADTRAD is not calculating radioactive decay in the pool with this approach.

The depletion equation being solved is:

dN/dt = - = - N (e + r + f) the solution to which is:

N/No = e-t = e-(e + r + f) t represents the total removal coefficient for radioactive iodine in the pool. Using the examples above:

= (e + r + f) = (e + 16 + 2.35) = 2.4E-5 s-1 Both the stagnant-pool and convecting-pool evolution coefficients are small relative to the decay and filtration removal coefficients so that the total removal coefficient is effectively independent of the evolution removal coefficient.

Since radioactive activity is proportional to radioactive decay constant this solution also applies to activity:

A/Ao = N r / (No r) = e- t = e-(e + r + f) t Similarly one can also divide the numerator and denominator by Vpool (essentially multiplying by

1) so that equation is in terms of either [molar] concentration or activity concentration.

The equation for the release rate to the gas is:

dNe/dt = N e Ne represents the number of moles evolved to gas Ne = Integral (N e) dt l from 0 to t = Integral (e e- t) dt l from 0 to t

= No (1 - (e/ ) e- t)

The evolution release fraction then is:

RFe = Ae/Ao = Ne/No = 1 - (e/ ) e- t = 1 - (e/ (e + r + f)) e- (e + r + f) t A similar approach can be taken to determine the amount of radioactive iodine that is lost from the pool by decay or collected in the filtration system.

Figure 12 shows the radioactive iodine disposition results using the simplified transient equations and the evaluated MTC with the values determined for the example FHA scenario. It compares the filter results to the no-filter results. A log concentration scale is used to show highly differing 64

concentrations. If plotted on a linear concentration scale, the no-filter curves would look almost identical to those for the numerical calculation that considers the change in iodine volatile fraction shown in Figure 10.

PF refers to I-131 fraction remaining in the pool, RF to the release fraction to gas, FF to the fraction deposited on the recirculation filter, and Rad to the fraction that has decayed in the pool.

Previous calculations using a recirculating MTC from Fischer indicated that, if filters are credited, that evolution release is low even if pool is assumed to be quite convecting (MTC of ~ 3.6E-5 m/s). Efficient iodine filters in a recirculation system that moves 2 pool volumes in a day depletes the iodine from the pool in about 3 days.

65

Comparison between filter and no filter, convective pool MTC 1.E+00 1.E-01 Fraction 1.E-02 1.E-03 1.E-04 1.E+04 1.E+05 1.E+06 1.E+07 Time (s)

PF, no filt PF, filter RF, no filt RF, filter FF, filter Rad, filter Rad, no filt Figure 12 Comparison of iodine evolution between filter and no-filter scenarios using the calculated pool MTC, 3.66E-6 m/s Evolution model limitations There are a few limitations to the evolution analysis. Some of these include using nominal values for the CsI, I2, and methyl iodide quantities within the gap, not evaluating iodine deposition on surfaces or organic iodide formation, and not having attempted to assess the uncertainty of the results.

Activity due to shorter-lived iodine isotopes was considered to come from I-131. For this example, simple decay from an assumed isotopic inventory at shutdown was used to evaluate activity at the time of the postulated FHA. Ingrowth was not considered but could be significant.

Calculations that involve the evaluation of the iodine decay chain should be used to determine inventories at the time of the FHA event. The activities from I-131 and shorter-lived iodine isotopes were summed and assumed to be part of the I-131 inventory. This was done to simplify the evolution calculation to a single RF equation rather than a sum of RF equations. Since the other isotopes are shorter-lived this simplification is conservative.

Temperature can affect iodine evolution in a few ways: The initial gap temperature affects the fraction of molecular iodine, I2, that would be in gaseous form in the gap. The initial gap temperature also affects gap pressure. The reactions that determine iodine speciation (iodine 66

volatile fraction) and the iodine partition coefficient are both temperature dependent. Temperature affects water properties and the diffusivity of iodine in water.

Simplifications were made for the calculation due to the limited time involved in the analysis: All molecular iodine in the gap is assumed to be gaseous both for simplicity and conservatism with the large impact of rather uncertain gap temperature on volatility. Although partition coefficients were evaluated as temperature dependent, nominal temperatures are used elsewhere in the evolution mode including water properties, iodine diffusivities in water, and in the chemical reaction rates that determine speciation.

Temperature-dependent iodine transport properties were not available. A temperature-dependent iodine diffusivity model was found during the literature review but the document containing the model was not obtained until after the analyses was conducted. Insufficient time was available to go back and redo the analysis.

The evaluation of the volatile fraction uses reaction rates valid at 25oC rather than actual temperature and does not account for the potential for iodate (IO3-) on the assumption that the concentration is negligible. Beahm, et al. (ORNL, 1992) indicated that temperature dependent reaction rates were not available for some of the reactions considered in the model but that experiments indicate that the volatile fraction of iodine decreases with increasing temperatures.

This effect was confirmed by other experiments and analyses (NUREG/CR-2900, NUREG/CR-3514, NUREG/CR-2493). These included the analyses of iodate and IO3-. The combined effect of simplifications (temperatures, surface adsorption, additional species, organic iodide production) seems conservative. Future calculations could account for temperature differences and the presence IO3- (if not also other iodine-containing species) and ensure that radiolytic I2 formation are not underpredicted). Considering surface deposition in the calculation while accounting for the possibility that radiation can reverse deposition, accounting for the temperature dependence of reactions, and accounting for organic iodide production, would be something that should be performed in a best-estimate calculation. It is not certain that the additional complexity is worth the expense for a design basis calculation. 10 4.4 General Description of the Alternative DBA FHA Method The fission product release from the breached fuel is based on the environmental conditions in which fuel handling operations are taking place. Under these conditions, a time period is considered between power operation and the movement of recently irradiated fuel to account for both radioactive decay, less decay power, and the use of pool water temperature to determine 10 A better estimate can be made of the pool surface velocity and the liquid-gas contact time used for the determination of the liquid-phase mass transfer coefficient if a computational fluid dynamics analysis were available. The results of such an analysis could be used to calibrate a natural-convection model that uses decay power and pool dimensions as input.

67

internal gas temperature and thus pin pressure. This time period may be controlled by the facilities Technical Specifications.

The pool water temperature is normally maintained between 100 to 125°F, depending on the facility design. When a full-core unload of fuel is required, this temperature will rise due to the higher decay heat generation of the freshly discharged fuel. Depending on the amount of fuel discharged and the time since shutdown, the spent fuel pool temperature may rise to as much as 150°F. The NRC FRAPCON computer code was used to compute a bounding end-of-life rod internal pressure for fuel being moved in the reactor cavity or spent fuel pool. The code calculates the steady-state response of light-water reactor fuel rods during long-term burnup.

The code can compute the temperature, pressure as functions of time-dependent fuel rod power and coolant boundary conditions. The phenomena modeled, in part, include: 1) heat conduction through the fuel and cladding to the coolant (pool water in this case) and the fission gas release from the fuel and rod internal pressure. Following a 24-hour time period and a pool water temperature of 150°F, a bounding end-of-life rod internal pressure of 760 psig is estimated.

The estimated activity in the fuel pin gap applies the non-LOCA gap inventory fractions presented in Table 3 of RG 1.183. To determine the fission product inventory in one damaged fuel assembly, the number of fuel pins in the dropped assembly were divided by the number of fuel pins in the core. It is assumed 100% the fuel pins rupture in the dropped assembly. To account for differences in power level across the core, applicable maximum core radial peaking factors were applied in determining the inventory of the worst-case damaged rods.

The iodine gap activity in the damaged rods is assumed to be released in two stages. Following the guidance in RG 1.183, Regulatory Position 3.5 and Appendix B, Section 1.3, the chemical form of radioiodine released from the fuel to the pool water should be assumed to be 95 percent CsI, 4.85 percent I2, and 0.15 percent organic iodide. The first stage is the instantaneous gaseous release from the fuel gap in rising bubbles where I2 and organic iodine are conservatively assumed to be in vapor form and subsequently decontaminated by passage through the overlying pool of water into the building atmosphere. This activity is then vented to the environment over a 2-hour period. The second stage is the protracted release initiated 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> (following the initial gaseous release) following the fuel bundle drop. The CsI in the fuel gap of the damaged assembly is conservatively assumed to completely dissociate into the pool water then slowly re-evolve into the building atmosphere as I2 due to the low pool water pH. This activity is vented directly to the environment for a period of thirty days.

The other radionuclides considered include xenons, kryptons, halogens, cesiums, and rubidiums.

First Stage - Instantaneous Release An overall iodine DF is a function of bubble size and rise time through the water column, both of which are functions of fuel pin pressure. If the water depth is 19 feet or greater, an overall effective iodine DF for I2 and organic iodine can be computed based on a best-estimate rod pin pressure for the limiting fuel rods in the reactor core at the most limiting time in life. The time period between reactor shutdown and the movement of fuel may be used to compute radioactive decay and less decay power. The use of pool water temperature based on a full-core offload may be used to determine internal gas temperature and thus pin pressure.

68

To compute rod bounding internal pressures, the staff used the NRC maintained FRAPCON fuel rod thermal-mechanical fuel performance code by modeling an assembly with an aggressive power history that maximizes at-power rod internal pressure. The initial conditions of the assembly are based on a 24-hour time period following and assuming a pool water temperature of 150°F, simulations predicted a bounding end-of-life rod internal pressure of 760 psig.

The first stage assumes I2 and organic iodine are in vapor form and are decontaminated by passage through the overlaying pool of water. An overall iodine DF based on a pin pressure is computed as follows:

= 81.046 0.305 (Equation 4-11) where:

t = bubble rise time (sec), computed as a function of pin pressure, x (psig), as:

() = 9.2261 64 (Equation 4-12) d = bubble diameter (cm), computed as a function of pin pressure, x (psig), as:

() = 0.0002 x + 1.0009 (Equation 4-13)

Equation 4-11 computes an iodine DF of 662 based on a pin pressure of 760 psig. The updated Burley model computes a limiting iodine DF of 667 based on the assumption that 0.15% is organic iodine. For the purposes of this analysis, both are considered numerically equivalent.

Therefore, a slightly lower iodine DF of 650 is applied based on the analyses and various discussions above. The retention of noble gases in the pool water is negligible and a DF of 1 is assumed. Particulate radionuclides are assumed to be retained by the water in the pool water and an infinite DF is assumed.

2.5 Second Stage - Protected Release Iodine present in the gap is expected to end up in the pool. Although it is assumed to occur instantaneously, there is likely some delay. Cesium iodide, the predominant expected form of iodine in the gap. Due to the low pH of the pool water, the CsI then re-evolves into the building atmosphere as elemental iodine.

The mass-transfer coefficient developed for iodine in a pool of water applies the surface-film-renewal model based on the THAI experiments. 11 The overall mass transfer coefficient based on 11 Fischer, K. W. (2012). Experimental Determination and Analysis of Iodine Mass Transfer Coefficients from THAI Test Iod-23. Cologne (Germany): 5th European Review meeting on Severe Accident Research (ERMSAR-2012).

69

the liquid phase, KL, was evaluated to be 3.66E-6 m/s, the pool surface-to-volume ratio, Spool /

Vpool, of 9.374E-2 m-1, and the fraction of iodine that is in I2 form, Xe of 1.46E-2. With the overall mass transfer coefficient known, in conjunction with facility-specific parameters describing the spent fuel pool, the simplified approach to compute the iodine re-evolution coefficient from the pool, e, is calculated at a rate of 5.009E-9 per second. Finally, the re-evolution from the pool to the surrounding building is modeled as a volumetric flow rate, Qe, and calculated at 5.78E-6 (m3/s) or 3.47E-4 m3/min. This flow rate is modeled using RADTRAD by using a volume that represents the pool and the volumetric flow rate that represents re-evolution to the refuel floor.

4.5 RadTrad Models Four cases were modeled as follows:

1. Base-case staff model of facilities current licensing basis.

Note: Modeled facility FHA DBA as described in the UFSAR with NUREG-1465 Iodine speciation.

2. Base-case with IDF = 650.

Note: Modeling with RadTrad Model Editor Version 2.5.8, the Edit Inventory Scenario, Plant Parameters, the pool iodine DF input is hard wired for inputs between 1-200 and will not accept a value larger than 200. The work-around for this issue is to adjust the gap fractions for iodine inputs. This is done by applying a golden ratio between the iodine fractions to an equivalent pool iodine DF as follows:

2.5.8 The RadTrad:

Gap Fraction I-131 input = (1

• 0.08) / 650 = 1.23E-4.

Gap Fraction Other Iodine input = (1

• 0.05) / 650 = 7.69E-5.

The RadTrad Pool Iodine DF input = 1.0.

3. Base-case with IDF = 650 and re-evolution.

Note: Adjusted Case 1 to model an iodine DF = 650. To do so, adjust Source 1, Pool Iodine DF = 1.0, Gap Fraction I-131 = 1.23E-4, and Gap Fraction Other Iodine = 7.69E-

5. Change Source 2, Source Term Fraction = 0.998 to account for remaining iodine.

Change Source 2, Elemental Faction = 1.0 to account for re-evolution of CSI to I2. Turn set radionuclides other than iodine to zero.

4. Base-case with only re-evolution pathway modeled.

70

Note: Adjusted Case 1 to only re-evolution. To do so, adjust Source 1, Source Term Faction = 0.0. Change Source 2, Source Term Fraction = 0.998 to account for remaining iodine. Change Source 2, Elemental Faction = 1.0 to account for re-evolution of CSI to I2. Turn set radionuclides other than iodine to zero. In Edit Inventory Scenarios, change Gap Fractions to RG 1.183. Resets by hitting the button RG 1.183. In Edit Inventory Scenarios, set Pool Iodine DF to 1.0.

Figure 13 provide a visual of the revised FHA model with the pool re-evolution modeled.

Figure 13: Visual of FHA Model with new Iodine DF model and Re-evolution Pilot Plant Control Room Immediately after the fuel bundle drop, radionuclides are assumed to be released from the pool to the refueling floor in sufficient quantities to initiate control room emergency filtration system due to high radiation.

Within 60 seconds of the isolation signal, the control room emergency filtration system initiates, which isolates the normal unfiltered control room air conditioning system supply. Prior to isolation, the total air intake rate is 3635 cfm (which includes normal air intake flow, infiltration leakage, and inleakage through opening and closing of doors). No credit is taken for filtration in the first 60 seconds. After isolation, the total air intake rate is 1210 cfm, which includes control room emergency filtration system intake flow, ingress/egress inleakage, and unfiltered inleakage (400 cfm unfiltered inleakage is assumed even when the isolated control room is at positive pressure). Control room emergency filtration system filter efficiency is specified as 90 percent for all iodine species. This is reduced by 1% to account for bypass. The resulting radionuclide concentration within the control room envelope is diluted by the air space volume.

RadTrad Notes and Input Parameters 71

Table 11: Pilot Plant DBA "Other Load" Drop Pilot Plant DBA "Other Load" Drop Parameters Analysis Methodology: RG 1.183 Reactor Core:

Reactor Power (MWt) 2381 Decay Time after Shutdown (hrs) 24 Fuel Bundle Type GE14 10x10 Number of Rods in the Core 47859 Number of Fuel Rods per Bundle 151 Number of Fuel Assemblies Damaged ---

Percent of Fuel Rods Failed ---

Number of Failed Fuel Pins (FHA) 151 Number of Failed Fuel Pins (Other Load) ---

Radial Peaking Factor 2 Water level above damaged fuel (ft) 23 Chemical Speciation:

Fuel Rod Plenum (Gap) Fraction Group Fraction I-131 0.08 Kr-85 0.1 Other Noble Gas 0.05 Other Halogens 0.05 Alkali Metals 0.12 Release to Environment:

Initial Release Pool Release (hrs) 0-2 hrs Re-evolution Release 2hrs-30days Pool Decontamination Factors Effective IDF: 650 Percent Releases:

Noble Gas (%) 100 Mitigation and Filtration:

SFP Volume (ft3) 40,682 (1152 m3)

SFP surface area (ft2) 1163 (108 m2)

Pool Re-evolution rate (ft3/min): 1.22E-2 (3.47E-4 m3/min)

Reactor Building Volume (ft3) 7.95x105 Normal Operations: N/A Emergency Operations: N/A Control Room Volume (ft3) 4.576x104 Isolation (sec) 60 Normal Operations:

Filtered Make-up Flow Rate (cfm)

Filtered Recirculation Flow Rate (cfm) N/A Unfiltered Make-up Flow Rate (cfm) 3235 Unfiltered Inleakage (cfm) 400 Exhaust (cfm) 3635 Emergency Operations:

Filtered Make-up Flow Rate (cfm) 810 Filtered Recirculation Flow Rate (cfm) N/A Unfiltered Make-up Flow Rate (cfm) N/A Unfiltered Inleakage (cfm) 400 Exhaust (cfm) 1210 Filter Efficiencies:

72

Elemental 0 (89%)

Organic 0 (89%)

Particulate 0 (89%)

Occupancy 0-24 hrs 1 1-4 days 0.6 4-30 days 0.4 Breathing Rate (m3/sec) 3.5x10-4 Table 12: Pilot Plant /Q Value for the Exclusion Area Boundary Time Period /Q Value (sec/m3) Comments 1 to 30 days 5.20E-04 Reactor Building Vent (Ground level release)

Table 13: Pilot Plant /Q Values for the Low Population Zone Time Period /Q Value (sec/m3) Comments 0 to 0.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> 2.90E-04 Reactor Building Vent (Ground level release) 0.5 - 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> 2.90E-04 Reactor Building Vent (Ground level release) 8 to 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> 7.30E-05 Reactor Building Vent (Ground level release) 1 to 4 days 2.50E-05 Reactor Building Vent (Ground level release) 4 to 30 days 5.20E-06 Reactor Building Vent (Ground level release)

Table 14: Pilot Plant /Q Value for the Control Room Time Period /Q Value (sec/m3) Comments 0 to 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> 4.15E-03 Reactor Building Vent (Ground level release) 2 to 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> 3.24E-03 Reactor Building Vent (Ground level release) 8 to 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> 1.32E-03 Reactor Building Vent (Ground level release) 1 to 4 days 9.01E-04 Reactor Building Vent (Ground level release) 4 to 30 days 7.22E-04 Reactor Building Vent (Ground level release) 4.6 Results Results are calculated in terms of total effective dose equivalent for the hypothetical maximum exposed individual. The total effective dose equivalent is the sum of the committed effective dose equivalent from inhalation and the deep dose equivalent from external exposure. The maximum exclusion area boundary total effective dose equivalent is determined within a two-hour period following the start of the radioactivity release. The low population zone total effective dose equivalent is determined for the most limiting receptor at the outer boundary of the low population zone for the duration of the accident which is modeled as 30 days. The control room total effective dose equivalent is determined for a period of 30 days without any operator action. These results represent realistic operating plant conditions and are considered conservative. Calculated radiological consequences for each case can be found in Table 15. Results from the case study show all scenarios are within the applicable total effective dose equivalent related radiological accident dose criteria of 10 CFR 50.67.

Generally speaking, the control room doses are most limiting. The estimated doses at all three receptors are lower than the licensees current licensing basis by making use of the alternative source term to more realistically model the DBA FHA based on the environmental conditions in 73

which fuel handling operations are taking place. Case 2 simply adjusts the pool iodine DF from a value of 200 to 650 where a near linear decrease is doses is observed. Cases 3 and 4 are based on the proposed alternative FHA model, modeling each release phase separately. Case 5 is based on the proposed alternative FHA model, modeling both releases phased in one RadTrad deck. The proposed FHA model would result in decreased computed doses of 98% at the exclusion area boundary (EAB), 94% at the low population zone (LPZ), and 91% at the control room (CR).

Table 15: Case study results for Pilot Plant Licensee CLB EAB (rem) LPZ (rem) CR (rem) (%-diff) 1.45 0.82 4.51 Case 1: Base-case staff model of facilities current licensing basis 1.20 0.67 4.46 (base)

Case 2: Base-Case staff model with IDF = 650 0.49 0.27 1.40 (-69%)

Case 3: Proposed DBA FHA, Gap Release only 0.024 0.014 0.07 Case 4: Proposed DBA FHA, Re-evolution only 0.01 0.03 0.32 Case 5: Proposed DBA FHA, Gap Release and Re-evolution 0.02 (-98%) 0.04 (-94%) 0.39 (-91%)

4.7 Recommendation and Guidance It is recommended the staff include this alternative FHA model as an acceptable alternative method Appendix B to RG 1.183. By doing so, the staff and licensees would be able to more readily enable the use of a more realistic DBA FHA source term.

It is recommended the alternative FHA model include the following:

• Account for the reduction of amount of radioactive iodine available to be released as gas due to limited I2 vapor pressure since I2 is solid at FHA temperatures.

• Assume initial gas pressure under FHA conditions:

• It would be more realistic to use the upper limit of gas generation models rather than assuming additional gas to reach limit. There is a sufficient understanding of fission gas generation.

• Consider a wait time (24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />) for less decay power and use spent fuel pool temperatures to determine gas temperature and thus pressure. In assuming higher pressures initially results in greater cooldown (decontamination) effects upon depressurization.

• Re-evolution of iodine from pool.

74

5. REFERENCES A.K. Postma et. al. (1970). Iodine Removal from Containment Atmospheres by Boric Acid Sprays Report BNP-100. (ADAMS Accession No. ML12038A261). Richland: Battelle Northwest Pacific Northwest Labratories .

Burley, G. (1971). Evaluation of Fission Product Release and Transport (Accession ML16357A003). Washington: Atomic Energy Commission.

Cole, R. H. (1948). Underwater Explosions. Princeton: Princeton University Press.

Collins, J. L. (1987). Fission Product Iodine and Cesium Release Behavior Under Light. Nuclear Rechnology Vol. 81 78.

Datta, R., Napier, D., & Newwitt, D. (1950). Properties and Behavior of Gas Bubbles Formed at a Circular Orifice. Trans. Inst. Chem Engin. (London), 15-26.

E. C. Beahm et al. (1992). Iodine Chemical Forms in LWR Severe Accidents, NUREGICR-5732 (ONRL/TM-11861). Oak Ridge: Martin Marietta Energy Systems, Inc, Oak Ridge National Laboratory.

Fischer, K. W. (2012). Experimental Determination and Analysis of Iodine Mass Transfer Coefficients from THAI Test Iod-23. Cologne (Germany): 5th European Review meeting on Severe Accident Research (ERMSAR-2012) .

Fogg, P. S. (2003). Chemicals in the Atmosphere: Solubility, Sources and Reactivity. New York:

John Wiley & Sons.

G. I. Taylor, R. M. (1950). The Rate of Rise of Large Volumes of Gas in Water. Contribution in:

Underwater Explosion Research Vol. II. Washington: Office of Naval Research, Dept. of the Navy.

Nate, T., & Himmelblau, D. (1967). Mass transfer from large single bubbles at high reynolds numbers. AIChE, 13(6), 697-702.

ORNL. (1992). NUREG/CR-5950 "Iodine Evolution and pH Control". Oak Ridge: Oak Ridge National Laboratory.

Parker, G. M. (1967). Simulated Transient Accidents in TREAT, ORNL-4288, Annual Report, Nuclear Safety Program 1967, pp.44-54. Oak Ridge: Oak Ridge National Laboratory.

PNNL. (1979). Topical Report PNL-2575, LWR Nuclear Fuel Bundle Daa for use in Fuel Bundle Handling. Richland: Pacific Northwest Laboratory.

R. L. Datta, D. H. (1950). Properties and Behavior of Gas Bubbles Formed at a Circular Orifice.

Trans. Inst. Chem. Engin., 15-26.

USNRC. (1972). Assumptions Used for Evaluating the Potential Radiological Consequences of a Fuel Handling Accident in the Fuel Handling and Storage Facility for Boiling and Pressurized Water Reactors, Safety Guide 25. Washington: US Nuclear Regulatory Commission.

75

USNRC. (1972). Regulatory Guide 1.25, "Assumptions Used for Evaluating the Potential Radiological Consequences of a Fuel Handling Accident in the Fuel Handling and Storage Facility for Boiling and Pressurized Water Reactors," Revision 0. Washington: US Nuclear Regulatory Commission.

USNRC. (1983). Taylor, J. H., Babcock & Wilcox, letter to D. H. Moran. Washington: US Nuclear Regulatory Commission.

USNRC. (1999). Black, S.C., letter dated November 24, 1999, Task Interface Agreement 99-03

- Potential Non-Conservative Assumptions for Fuel Handling Accidents (ADAMS Accession No. ML993340497). Washington: US NUclear Regulatory Commission.

USNRC. (2000). Regulatory Guide 1.183, "Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors". Washingtion: US Nuclear Regulatory Commission.

USNRC. (2003). Regulatory Guide 1.195, "Methods and Assumptions for Evaluating Radiological Consequences of Design Basis Accidents at Light-Water NUclear Power Reactors".

Washington: US NUclear Regulatory Commission.

USNRC. (2004). Palo Verde Nuclear Generating Station, Units 1, 2, and 3 - Issuance of Amendments RE: Internal Fuel Pin Pressure (TAC Nos. MC0620 MC0621, and MC0622)

(ADAMS Accession No. ML0423720349). Washington: US Nuclear Regulatory Commission.

USNRC. (2010). NUREG-1801, "Generic Aging Lessons Learned (GALL) Report Final Report," Revision 2. Washington: U.S. Nuclear Regulatory Commission.

USNRC. (2017). NUREG 0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants. Washington: US Nuclear Regulatory Commission.

Westinghouse. (1970). WCAP-7518-L - Topical Report: Radiological Consequences of a Fuel Handling Accident, (attachment to accession 9804290400). Monroeville: Westinghouse Electric Corporation.

Westinghouse. (1971). WCAP-7828 - Topical Report: Radiological Consequences of a Fuel Handling Accident, (ADAMS Accession Number ML18025B589). Monroeville:

Westinghouse Electric Corperation.

Yuill, W. B. (1970). TID-4500, "Release of Radioiodine from Open Pools". Idaho Nuclear Corporation.

76