NSD-NRC-96-4675, Submits AP600 Source Term for Design Basis LOCA Analysis

Submits AP600 Source Term for Design Basis LOCA Analysis

ML20101M869
Person / Time
Site:	05200003
Issue date:	04/01/1996
From:	Mcintyre B WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:	Quay T NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
DCP-NRC0484, DCP-NRC484, NSD-NRC-96-4675, NUDOCS 9604080125
Download: ML20101M869 (9)

Text

-

w\\

v,y

=

Westinghouse Energy Systems Ba 355 Electric Corporation Pittsburgh Pennsylvania 15230-0355 NSD-NRC-96-4675 DCP/NRC0484 Docket No.: STN-52-003 April 1,1996 Document Control Desk U.S. Nuclear Regulatory Commission

{

l Washington, D.C. 20555 ATTENTION:

T.R. QUAY

SUBJECT:

AP600 LOSS OF COOLANT ACCIDENT SOURCE TERM MODEL

Dear Mr. Quay:

The attachment to this letter describes the source term model which is being used for AP600. This model is based on the NRC source term model described in NUREG-1465 and supersedes the model currently in the AP600 SSAR (Revision 5). The use of this revised source term model has been discussed with NRC staff. The AP600 SSAR (subsection 15.6.5) will be updated to reflect the revised source term model in a revision, which will be issued July,1996.

The NRC technical staff should review this material in preparation for a working level meeting to discuss any concerns they might have. Mr. Butler will contact Mr. Kenyon to arrange this meeting.

Please contact John C. Butler on (412) 374-5268 if you have any questions concerning this transmittal.

D a j i:?r u Brian A. McIntyre, Manage Advanced Plant Safety and Licensing

/nja Enclosure cc:

T. Kenyon, NRC W. Huffman, NRC J. Lee, NRC R. Emch, NRC J. Kudrick, NRC M. Snodderly NRC N. Dudley, ACRS

~

N. J. Liparulo, Westinghouse (w/o Enclosure) mor 9604080125 960401 l

PDR ADOCK 05200003 7

A PDR C

AP600 Source Term for Design Basis LOCA Dose Analysis The source term methodology being used for the AP600 follows NUREG-1465 (Reference 1).

NUREG-1465 defines the NRC's revised source term model. As identified in NUREG-1465, an applicant may utilize justifiable deviations from the NUREG-1465 source term parameters. Specific features of the source term model being used for the AP600 are provided below and the departures from the NUREG-1465 model are identified.

Coolant Activity Release Consistent with NUREG-1465, the first phase of activity release is the release of reactor coolant activity during the first 30 seconds. (It is noted that NUREG-1465 specifies that the duration of the coolant activity release phase is 10 minutes for plants with leak-before-break approval but this is conservatively ignored.)

The noble gas concentrations in the primary coolant are consistent with operation at a design basis fuel defect level of 0.25%. The iodine concentration in the primary coolant is 24 pCi/g Dose Equivalent I-131 which reflects a pre-existing iodine spike. The coolant activity is relevant only in determining releases via the containment purge prior to its isolation. Once there is fuel damage, the activity from the primary coolant becomes insignificant.

Core Gap Activity Release Analysis of the LOCA core response for the AP600 shows that there may be early failure o a r

small fraction of the fuel rods (less than 5% of the rods). It is assumed that 5% of the rodt release their gap inventory instantaneously at the end of the reactor coolant activity release phase. Consistent with NUREG-1465 guidance for fuel that has continued cooling availat<e, the fission product gap inventory is 3% for noble gases, iodines, and cesiums. As described in Reference 2, core cooling will remain for a period of time. The release constitutes 0.15%

of the core fission product inventory.

As described in Reference 2, there is calculated to be a 53 minute delay from the initiation of the postulated large break LOCA to the onset of fuel clad failures due to fuel heatup after core uncovery. This time delay is rounded down to 50 minutes. This is a departure from NUREG-1465 which specifies that the gap release phase begins at 30 seconds (or at ten minutes if there is leak-before-break approval). In NUREG-1465, recognition is given that the timing of onset of core damage for a plant such as the AP600 may differ from the timing specified in the report. Reference 2 provides the technicaljustification, based on design specific features, to support the use of a 50 minute delay before core damage is initiated.

At the beginning of the gap release phase, there is assumed to be an instantaneous release of the gap inventory of the remaining 95% of the rods that did not suffer early failure. With the 3% gap fraction, this release of activity is 2.85% of the core inventory. An additional 2% of the ccre inventory is assumed to be released over the half hour duration of the gap release phase. This modeling of the gap release phase is consistent with NUREG-1465 with the exception of the assumption of the 5% of rods failing early and the assumption of the 50 minute delay in onset of the gap release phase.

2720A

It is noted that the instantaneous failure of 95% of the fuel rods is an extremely conservative approach to modeling of gap releases. A more appropriate model would be to assume that the.

2.85% of core inventory is released over the half hour long gap release phase. The use of the instantaneous release model is not intended to be an approval of this aspect of the model; it is being followed to minimize the number of differences between the source term used for the AP600 and the source term defined by NUREG-1465.

Core Melt Release l

After the gap release phase, there is an in-vessel release phase which lasts for 1.3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> and j

which releases activity to the containment due to core melting. The fractions of the core activity released to the containment atmosphere during this phase are compared below with the j

values from NUREG-1465:

AP600 Source Term NUREG-1465 '

j Noble gases 0.95 0.95 Jodines 0.35 0.35 Cs & Rb 0.25 0.25 Tellurium group 0.05 0.05 Noble metals (Ru group) 0.0025 0.0025 Ba & Sr 0.004 0.02 i

Cerium group 0.0001 0.0005 i

Lanthanide group 0.0001 0.0002 l

The releases are assumed to occur at a constant rate over the 1.3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> phase duration. The release duration and the release fractions for the first five nuclide groups are consistent with the guidance of NUREG-1465. Exception is taken to NUREG-1465 in regard to the release fractions for the last three nuclide groups. The use of the lower release fractions for these nuclide groups is recommended by the Department of Energy's Advanced Reactor Severe Accident Program (ARSAP) based on evaluation of measurements from the Severe Fuel Damage experiments in the Power Burst Facility and from the Three Mile Island Unit 2 -

accident'(Reference 3).

There is no additional release of activity to the containment atmosphere after the in-vessel j

release phase. This is consistent with the staff position as stated in SECY-94-302 (Reference 4).

Chemical Form of Non-Gaseous Releases Consistent with NUREG-1465, the iodines are assumed to be 95% particulates (cesium iodide) and 5% elemental prior to release from the fuel. After release to the containment, a portion of the elemental iodine is assumed to combine with organic materials to create organic iodine compounds. No time delay is assumed for the formation of organic iodine. As stated by NUREG-1465, the final species distribution is 95% Csl,4.85% elemental, and 0.15%

organic.

Consistem with NUREG-1465, all other non-gaseous activity releases to the containment atmosphere are assumed to be in particulate form.

2720A -

-.~

References

1. NUREG-1465, " Accident source Terms for Light-Water Nuclear Power Plants," February 1995, L. Soffer, et al.

2.

Letter NTD-NRC-94-4335,11/2/94,

Subject:

" Position Paper on AP600-Specific Time Delay in the Physically Based Source Term" 3.

" Releases to Containment of Low Volatility Fission Products from In-Vessel Processes,"

2/14/95, R. R. Hobbins and D. J. Osetek (attached)

4. SECY-94-302, " Source Term-Related Technical and Licensing Issues Pertaining to Evolutionary and Passive Light-Water-Reactor Designs," December 19, 1994.

4 l

1 2720A -

- - _ - - ~ _ - - - -..-.-

! '. *. APR 18 '95 01:30PM ARD/AL W (415)855-7945 P.2 Releases to Containment of Low Volatility Fission Products from In-Vessel Processes R. R. Hobbins D. J. Osetek

}

j February 14,1995 ii' 1

introduction j

in support of the design certification of the AP 600, the releases of the low volatile fission products Sr, Ba, Ce, and La to containment from early in-vessel processes have been reviewed. An updated recommendation and the technical basis for the j

recommendation are provided. The applicability of fission product release data from the TMI-2 accident is evaluated.

Backoround i

l Initial recommendations based on experimental measurements were provided in j

Reference 1. These values for fraction of core inventory released to containment, i

including an estimated fractional retention in the reactor coolant system of 0.6, were 1

0.004 for Sr and Ba and 0.00004 for Ce and La. The values. reported in Reference 1 bound measurements from Severe Fuel Damage (SFD) experiments in the Power Burst Facility (PBF) under realistic conditions and those from the Three Mile Island Unit 2 (TMI-2) accident.

A later, more detailed evaluation (2) of measurements from the SFD 1-3 and SFD 1-4 i

experiments that utilized high burnup fuel and from the TMI-2 accident led to values of j

0.0014 for Sr and Ba and 0.00005 for Ce and La, in reasonable agreement with the values reported previously in Reference 1. The evaluation in Reference 2 was based on calculations of release rates (from measurements of release fractions divided by the time above 2200 K) for the heatp and initial melting phase using data from the PBF-SFD tests, and the crust-encased molten pool phase using TMI-2 data.

Recommendations The methodology for calculating the releases of low volatility fission products in Reference 2 nas been revised to (a) increase the time for release from 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> to 1.3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> (in agreement with NUREG-1465 [Ref. 3]) and to (b) increase the release values by a factor of 1.33 to account for uncertainty in the fraction of the core participating in the high temperature excursion. These revisions lead to values of 0.0024 for Sr and Ba i

and 0.0001 for Ce and La.

To provide margin and account for uncertainty in the application of experimental data to i

plant severe accident analyses, it is recommended to use the higher of the values from Reference 1 and the revised values from Reference 2. Thus, the recommended values j

are 0.004 for Sr and Ba and 0.0001 for Ce and La. These values are factors of 2 to 5 1

4 APR 10 '95 17:46 415 855 7945 PAGE.02

l APR 18 '95 01:38PM ARD/ALWR (415)855-7945 P.3

'~

lower than the values in NUREG-1465 [Rd. 3), which tre 0.02 for Sr and Ba, 0.0002 for La, and 0.0005 for Ce.

The recommended values are lower than the NUREG-1465 values because the experimental results on which the recommended values are based include prototypic effects such as ceramic melt pools having a low surface-to-volume ratio and a self-healing crust, whereas the code calculations and expert opinion on which the NUREG-1465 values are based do not include these important effects, which became apparent after detailed evaluation of recent research results.

l The release values from all three referenced sources, along with those currently recommended, are presented in Table 1, below.

Table 1. Low volatile release fractions to containment Element Recommended ARSAP-1991 ARSAP-1992 NUREG-1465 l

Revised Sr,Ba 0.004 0.004 0.0024 0.02 La 0.0001 0.00004 0.0001 0.0002 Ce 0.0001 0.00004 0.0001 0.0005 Applicability of TMI-2 fission product release data l

in the TMI-2 accident, the core was reflooded before melt relocation to the lower plenum and, therefore, the applicability of TMI-2 fission product release data to accident scenarios involving later reflooding needs to be established. The time above 2200 K used to compute fission product release rates for TMI-2 in Ref. 2 is 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />. The literature on the TMI-2 accident, including calculations of core temperatures, analyses of water level, reactor instrument data during the accident, and fission product release measurements has been carefully reviewed.

It is clear that temperatures in the core were above 2200 K for at least 60 minutes, and very likely considerably longer (i.e., up to 630 minutes). It is also clear that the hottest region in the core was uncovered or partially uncovered for about 60 minutes when maximum temperatures exceeded 2200 K. Furthermore, it has been shown that the temperature in the melt, the crust thickness surrounding the melt, and the crust surface j

temperature are similar whether water is present or not. The fission product release pathway in the TMI-2 accident was through water which was fully characterized after the accident, along with surfaces along the pathway and sediments, to arrive at total fractional releases from the core. Therefore, fission product release rates calculated from measured release fractions and a 60 minute time for release can be conservatively applied to accident analyses.

The following information is cited in support of the above conclusion. Eight calculations of core temperature vs. time were found in the " Summary of the Three Mile Island Unit l

2 Analysis Exercise" [4] that were in reasonable agreement with data from the accident.

2 APR 18 '95 17:46 415 855 7945 PAGE.03

APR 18 '95 01:39Pf1 ARD/ALWR (415)855-7945 P.4 Th$n time calcul ted to reach 2200 K rcngid from 134-153 minutts aftsr racctor scram.

The average of the eight calculations is 144 minutes.

Although it is not known exactly how much water was put into the core during the B-loop pump transient at 174 minutes, it has been estimated that approximately 28 m3 may have been added [5]. This is just enough to cover the core, neglecting vaporization. Kuan, et al. [6] calculated from the 3 MPa pressure rise in the 200 seconds following initiation of the 2-B pump transient that 4,000 kg of water was vaporized. This corresponds to a volume of 4 m3 of water. It is known from measurements of reheating of in-core thermocouples in peripheral assemblies and increasing source range monitor response to levels greater than those recorded at 174 minutes that core heatup continued and water level in the core decreased after the B-loop pump transient (5). It appears possible that the hottest zone in the core may have been covered briefly as a resu!t of the influx of water during the 2-8 pump transient, but the water level resumed its downward trend.

The next water input to the core was the result of high pressure injection system operation from 200-217 minutes. Based on cessation of both reactor coolant system depressurization and pressurizer level decrease at 207 minutes, the core was refilled at this time [7). Analysis indicates that at 200 minutes the water level was at 2 m in the core [7]. The core former walls were penetrated by molten core debris at elevations extending to about 1.8 m above the bottom of the core [8). Prior to melting through the core former wall at 224 minutes the molten pool must have extended somewhat above l

1.8 m, but just how much is unknown. Therefore, portions of the core region at temperatures above 2200 K were likely covered at 200 minutes.

Given the above information, it is likely that the hottest region in the core (with temperatures above 2200 K) was uncovered in the period 144-174 minutes, partially covered from 174-207 minutes, and fully covered beyond 207 minutes. This information indicates that the hottest region of the core (> 2200 K) was uncovered for 30 minutes and partially uncovered for portions of an additional 33 minutes.

It has been calculated that the molten pool in the core did not cool to below 2800 K for some 550 minutes following the relocation of a portion of the melt to the lower plenum in the period 224-226 minutes (9). Therefore, temperatures were likely above 2200 K in the core for at least 630 minutes.

Calculations of heat transfer from molten corium pools [9-11] indicate that there is little difference in the fraction molten, the crust thickness, or the crust surface temperature with or without water present. This is because the heat transfer over much of the pool crust surface tends to be dominated by radiation (film boiling regime) to water or nearby structures. For a decay heat generation rate within the corium of 1.5 x 10 W/m3 (~

6 the TMI-2 value), the crust thickness and surface temperature along the lower boundary of the pool range from approximately 5.5-0.7 cm and 1300-1750 K. For these same conditions, it is calculated that the crust thickness on the top surface of the pool s approximately 0.3 cm and the surface temperature is about 2100 K. Only in the l

l 17:47 415 855 7945 PAGE.04

l

~

^ M i8 '95 01:39PM ARD/RM (415)855-7945 P.4 l

Tha time calculittd to rroch 2200 K rcnged from 134-153 minutes cft:r rsector scram.

j The average of the eight calculations is 144 minutes.

Although it is not known exactly how much water was put into the core during the B-loop pump transient at 174 minutes, it has been estimated that approximtely 28 m3 l

may have been added [5]. This is just enough to cover the core, neglecting vaporization. Kuan, et al. [6] calculated from the 3 MPa pressure rise in the 200 l

seconds following initiation of the 2-8 pump transient that 4,000 kg of water c.ss vaporized. This corresponds to a volume of 4 m3 of water. It is known from i

measurements of reheating of in-core thermocouples in peripheral assemblies and j

l increasing source range monitor response to levels greater than those recorded at 174 j

minutes that core heatup continued and water level in the core decreased after the B-loop pump transient [5). It appears possible that the hottest zone in the core may have l

been covered briefly as a result of the influx of water during the 2-B pump transient, but i

the water level resumed its downward trend.

i The next water input to the core was the result of high pressure injection system l

l operation from 200-217 minutes. Based on cessation of both reactor coolant system i

depressurization and pressurizer level decrease at 207 minutes, the core was refilled at this time [7]. Analysis indicates that at 200 minutes the water level was at 2 m in the core [7]. The core former walls.were penetrated by molten core debris at elevations i

extending to about 1.8 m above the bottom of the core [8). Prior to molting through the core former well at 224 minutes the molten pool must have extended somewhat above 1.8 m, but just how much is unknown. Therefore, portions of the core region at temperatures above 2200 K were likely covered at 200 minutes.

j Given the above information, it is likely that the hottest region in the core (with j

temperatures above 2200 K) was uncovered in the period 144-174 minutes, partially l

covered from 174-207 minutes, and fully covered beyond 207 minutes. This information indicates that the hottest region of the core (> 2200 K) was uncovered for 30 minutes and partially uncovered for portions of an additional 33 minutes.

It has been calculated that the molten pool in the core did not cool to below 2800 K for some 550 minutes following the relocation of a portion of the melt to the lower plenum in the period 224-226 minutes [9]. Therefore, temperatures were likely above 2200 K in the core for at least 630 minutes.

Calculations of heat transfer from molten corium pools [9-11] indicate that there is little difference in the fraction molten, the crust thickness,,or the crust surface temperature with or without water present. This is because the heat transfer over much of the pool I

crust surface tends to be dcminated by radiation (film boiling regime) to water or nearby structures. For a decay heat generation rate within the corium of 1.5 x 10 W/m3 (~

6 the TMI-2 value), the crust thickness and surface temperature along the lower boundary of the pool range from approximately 5.5-0.7 cm and 1300-1750 K. For these same conditions, it is calculated that the crust thickness on the top surface of the pool is approximately 0.3 cm and the surface temperature is about 2100 K. Only in the 3

APR 10 '95 17:47 415 855 7945 PAGE.04

. ~ APR '18 '95' ~ Ob 40PM ARD/ALJR (415)855-7945

~

^

P.5 1

special case of nuclerta boiling on th3 upper surface was a v:ry different result

?

obtained (crust thickness 0.8 cm, surface temperature 584 K [ saturation temperature)).

Calculations of mass transfer mechanisms within the pool (12) indicate that low volatility fission products are transported to the pool surface very slowly by diffusion (time constant ~220-390 days), more rapidly by natural convection (time constant ~7-12 i

hours), and most rapidly by diffusion to bubbles saturated with high volatility fission product vapors (time constant 3 seconds) whose rise time in the pool is of the order of 50 seconds. The bubble release mechanism is limited by the very low partial pressure of the low volatility fission products which tends to be further reduced by the very small mole fraction of these species in the melt. Even greater barriers to low volatile fission product release may be diffusion through the self-healing crust and j

vaporization from the relatively cool crust surface (1300-2100 K).

I Conclusion i

Analyses of the TMI-2 accident suggest that the core damage phenomena and j

associated in-vessel fission product release behavior are applicable to the severe i

accident scenarios used as the NRC licensing basis for advanced reactors. The

{

recommended values for low volatility fission products (Sr, Ba, Ce and La) have been j

developed from the TMl-2 accident results and recent in-pile tests. These values i

include some conservative margin to allow for uncertainties in core temperature history i

and release phenomena, and are deemed appropriate for the AP-600 and other advanced PWR accident analyses.

4 APR 18 '95 17:48 415 855 7945 PAGE.05

APR 18.'95 01:41PM ARD/ALWR (415)855-7945 k6

[

l.;.

.s

~

References

1. D. E. Leaver, et al., " Passive ALWR Source Term," DOE /ID-10321, February 1991.

2. D. J. Osetek, " Low Volatile Fission Product Release During Severe Reactor Accidents," DOEllD-13177-2, October 1992.

3. USNRC, " Accident Source Terms for Light Water Nuclear Power Plants," Final NUREG-1465, February 1995.

4. D. W. Golden, et al., " Summary of the Three Mile Island Unit 2 Analysis Exercise,"

Nucl. Technol. 87. 328 (1989).

5. J. M. Broughton, et al., "A Scenario of the Three Mile Island Unit 2 Accident," Nucl.

Technol. E,34 (1989).

8. P. Kuan et al., " Thermal Interactions During the Three Mile Island Unit 2 2-B Coolant Pump Transient," Nucl. Technol. E, 977 (1989).

7. Y. Nomura and J. L Anderson, " Analysis of the Refill Phenomena During the Three Mile Island Unit 2 Accident," Nucl. Technol. E, 912 (1989).

8. J. L. Anderson and J. J. Sienicki, ' Thermal Behavior of Molten Corium During the Three Mile Island Unit 2 Core Relocation Event," Nucl. Technol. E,283 (1989).

1

9. R. L. Moore, et al., "Three Mile Island Unit 2 Degraded Core Heatup and Cooldown Analysis," Nucl. Technol. E, 990 (1989).

l

10. M. Epstein and H. K. Fauske, "The Three Mile Island Unit 2 Core Relocation - Heat Transfer and Mechanism," Nucl. Technol. 87,1021 (1989).

11. J. E. O'Brien and G. L. Hawkes, " Thermal Analysis of a Reactor Lower Head with Core Relocation and External Boiling Heat Transfer," Heat Transfer - Minneapolis 1991, AIChE Symposium Series 283, Vol. 87, S. B. Yilmaz and B. G. Volintine..Eds.

(1991).

l l

12. D. A. Petti, et al., " Analysis of Fission Product Release Behavior from the Three Mile Island Unit 2 Core," Nucl. Technol. E, 243 (1989) i T

5 APR 10 '95 17:49 415 855 7945 PAGE.06