ML20028B785
| ML20028B785 | |
| Person / Time | |
|---|---|
| Site: | Sequoyah  |
| Issue date: | 11/30/1982 |
| From: | Novak T Office of Nuclear Reactor Regulation |
| To: | Fraley R Advisory Committee on Reactor Safeguards |
| References | |
| RTR-NUREG-0011, RTR-NUREG-11 NUDOCS 8212060033 | |
| Download: ML20028B785 (59) | |
Text
}
}
l s
DISTRIBUTION
/!1 i
Docket Ilos. 50 4 E7/328' ND\\,,u) 0 Gp-L'.NRC PDR -
> ~ "
- " ~2
,PRC S Docket Ifos.: 50-327 E'. Adensam and 50-328 C. Stahle M. Duncan D. Eisenhut/R. Purple R. Mattson IEl'ORANDUM FOR:
R. Fraley, Advisory Comittee on R. Vollmer Reactor Safeguards g
FRON:
T. H. Novak, Assistant Director y.$ uTL6N for Licensing, Division of Licensing SUP, JECT:
DRAFT SER FOR SEQU0YAH UNIT I SUBC0ff1ITTEE ftEETING Attached is a copy of the draft SSER for Sequoyah Unit I addressing the staff's review of the permanent hydrogen nitigation systen for Unit 1.
As discussed with you on the telephone the week of Novenber 15, 1982, we have transmitted copies of this draft to the following subcomittee members and consultants:
C. Mark C. Seiss I. Caton W. Lipinski G. Schott M. Sickel Z. Zudans so that they may have an opportunity to review the draft prior to the subcomittee neeting on Decenber 7,1982. We currently plan to have a final draft available for the subcomittee by C.O.B. Decenber 6,1982.
Orig nD131gnadb/,:
i Themas L IIO'd '
Thomas M. Novak, Assistant Director for Licensing Division of Licensing cc: See next page F212060033 821130 PDR ADOCK 05000327 i
E PDR I
n
..$p.t..
. Sj..... AD. t..f e,,m,
suan=s >..EU.eIIs.4mi.e.t
.,.e..s.,.......
~..I
. k....
.11../..p..../. 8. 2..
1 1../....g..../. 8. 2.. H../../.f..../. 8. 2....
/'
. q
,re >
Nnc ronu sia nomi nacu ono OFFICIAL RECORD COPY usem i i_mm
b SEQUOYAH Mr. H. G. Parris Manager of Power Tennessee Valley Authority 500A Chestnut Street, Tower II Chattanooga, Tennessee 37401 cc: Herbert S. Sanger, Jr., Esq.
General Counsel Tennessee Valley Authority 400 Commerce Avenue E 11B 33 Knoxville, Tennessee 37902 Mr. H. N. Culver Tennessee Valley Authority 400 Commerce Avenue, 249A HBB Knoxville, Tennessee 37902 Mr. Bob Faas Westinghouse Electric Corp.
P.O. Box 355 Pittsburgh, Pennsylvania 15230 Mr. Jerry Wills Tennessee Valley Authority 400 Chastnut Street, Tower 11 Chattanooga, Tennessee 37401 Mr. Donald L. Williams, Jr.
Tennessee Valley Authority 400 West Summit Hill Drive, W10B85 Knoxville, Tennessee 37902 Resident Inspector /Sequoyah NPS c/o U.S. Nuclear Regulatory Commission 2600 Igou Ferry Road Soddy Daisy, Tennessee 37379 James P. O'Reilly, Regional Administrator U.S. Nuclear Regulatory Commission, Region II 101 Marietta Street, Suite 3100 r
Atlanta, Georgia 30303
y.
e-t NUREG-0011 as hhh-Supplement No. 6 Safety Evaluation Report related to the operation of Sequoyah Nuclear Plant, Units 1 and 2 Docket Nos. 50-327 and 50-328 Tennessee Valley Authority U.S. Nuclear Regulatory Commission Office of Nuclear Reactor Regulation November 1982
- psna, u
s
/
Dun
o ABSTRACT l
l l
l 11/23/82 iii SEQUOYAH SSER6 TC ABS
a o
TABLE OF CONTENTS E BS.*.
ABSTRACT..............................................................
iii 1
INTRODUCTION.....................................................
1-1 22 TMI REQUIREMENTS.................................................
22-1 II.B.7 Analysis of Hydrogen Control..................................
22-1 1
Background.................................................
22-1 2 System Descripti,on.........................................
22-2 3 Combustion / Igniter Testing.................................
22-6 3.1 The Whiteshell Test Program...........................
22-7 3.2 The Factory Mutual /Acurex Test Program................
22-10 3.3 Tayco Igniter Testing.................................
22-12 3.4 Staff Conclusions Regarding Testing...................
22-14 4 Hydrogen Mixing and Distribution...........................
22-15 5
Detonations................................................
22-17 6 Degraded Core Accidents and Hydrogen Generation............
22-21 7 Sequoyah Containment Structural Capacity...................
22-22 8 Containment Analysis.......................................
22-24 8.1 Containment Codes.....................................
22-24 8.2 Containment Pressure and Temperature Calculations.....
22-29 8.3 Confi rmatory Analysis and Conclusion..................
22-35 9 Survivability of Essential Equipment.......................
22-37 9.1 Essential Equipment...................................
22-38 9.2 Thermal Environment Response Analysis.................
22-39 9.3 Pressure Effects......................................
22-45 9.4 Staff Conclusions Regarding Equipment Survivability...
22-45 l
10 Overall Conclusions........................................
22-46 APPENDIX A CONTINUATION OF CHRONOLOGY OF NRC STAFF RADIOLOGICAL SAFETY REVIEW 0F SEQUOYAH STATION APPENDIX B BIBLIOGRAPHY 11/27/82 v
SEQUOYAH SSER6 TC ABS
+
P, age LIST OF TABLES 22.1 Internal static pressure capacity for hydrogen burning, psig.....
22-23 22.2 Containment sensitivity studies..................................
22-32 22.3 Essential equipment..............................................
22-39 22.4 Comparison of analytically calculated thermal responses during hydrogen burn with qualification temperatures.....................
22-43
]
l i
l l
I I
l l
i 11/23/82 vi SEQUOYAH SSER6 TC ABS
e e
1 INTRODUCTION The purpose of this Supplement 6 to the Sequoyah Safety Evaluation Report (SER) is to update the staff's evaluation of the issue related to the hydrogen mitigation system that is being installed in Unit 1 and is to be installed in Unit 2 during the first refueling outage of that unit.
Except where noted, the material herein supplements the information that has been reported previously.
The following sections of this supplement are numbered to correspond to that in the SER and earlier supplements.
This supplement provides the basis for the staff's concluding that sufficient information is available to permit the installation and operation of a modified hydrogen mitigation system for Sequoyah Unit 1.
11/27/82 1-1 SEQUOYAH SSER6 SEC 1
l 22 TMI-2 REQUIREMENTS II.B.7 Analysis of Hydrogen Control
1 Background
The staff's licensing requirements relative to the provisions for hydrogen con trol beyond those prescribed in 10 CFR 50.44 have evolved from numerous deliber-ations among the Nuclear Regulatory Commission (Commission), the Advisory Com-mittee on Reactor Safeguards (ACRS), the NRC staff, and applicants and licensees.
In summary, the staff's requirement for ice condenser containments is that a supplemental hydrogen control system be provided so that the consequences of the hydrogen release generated during the more probable degraded core accident sequences do not involve a breach of containment nor adversely affect the func-tioning of essent'ial equipment.
In Supplements 4 and 5 to the Sequoyah SER (NUREG-0011), the staff concluded that the interim distributed ignition system (IDIS) installed at Sequoyah Units 1 and 2 is acceptable as an interim hydrogen control measure for degraded core accidents.
However, the ' staff N ommended that the detailed review of the distributed ignition system continue, so that a number of issues related to degraded core hydrogen control could be more thoroughly investigated before it endorsed a long-term commitment to deliberate ignition.
These issues included items related to combustion phenomena as well as further consideration of a spectrum of degraded core accident sequences.
Based on these recommendations, the operating licenses of Sequoyah Units l'and 2 were conditioned to require that the licensee, the Tennessee Valley Authority (TVA), continue research programs on hydrogen control measures and the effects of hydrogen-burn safety functions during the interim period of operation.
The research program was to include: (1) improvement of calculational methods for containment temperature and ice condenser response to hydrogen combustion, (2) research to address the potential for local detonation, (3) confirmatory tests 11/27/82 22-1 SEQUOYAH SSER6 SEC 22
t D
O e
l on selected equipment exposed to hydrogen burns, (4) new calculations to predict differences between expected equipment temperature environments and containment temperatures, and (5) evaluation and resolution of any anomalous results occurring during the course of the test program.
The license condition required that TVA, by the end of the first refueling outage, provide the bases for a Commission determination that an adequate hydrogen control system for the plants is installed and will perform its intended function in a manner that provides adequate safety margins.
As part of its research activities, TVA in cooperation with Duke Power and American Electric Power (AEP) continued to investigate alternative measures of hydrogen control.
As a result of continued studies, TVA has concluded that a deliberate ignition system, similar to the IDIS, provides adequate safety margins in controlling the consequences of degraded core accidents.
The new system, designated the permanent hydrogen mitigation system (PHMS), has been installed in Sequoyah Units 1 and 2.
The PHMS is identical in concept to the interim system but provides system design improvements.
A detailed discussion of the PHMS is provided later in this supplement.
The approach taken by TVA for establishing that the PHMS provides adequate safety margins relies on analytical modeling of the containment and equipment response to the degraded-core event.
Because the models involve simplifying assumptions and input parameters describing such complex phenomena as contain-ment mixing, flame speeds, and equipment heatup, the utility research program serves to verify key assumptions in the analyses.
2 System Description
The PHMS is a system of igniters and ancillary equipment TVA has installed within the containment of Sequoyah Units 1 and 2.
The igniters are designed to ensure a controlled burning of hydrogen in the unlikely event that excessive quantities of hydrogen, well beyond the design bases required by 10 CFR 50.44, are generated as a result of a postulated degraded core accident.
The PHMS is designed to promote the combustion of hydrogen in a manner such that containment overpressure failure is prevented.
11/27/82 22-2 SEQUOYAH SSER6 SEC 22
.D TVA has selected and tested a 120-V ac hermetically sealed thermal igniter manufactured by Tayco Engineering as the igniter to be installed in the PHMS.
The heating element is formed into a cylindrical coil approximately 1.75-in.
long and 0.75-in. in diameter.
Power is supplied directly to the igniter at 120 V ac. The igniter is mounted in a National Electrical Manufacturers Association (NEMA) Type 4 enclosure with the heating element protruding.
This enclosure is designed to remain watertight under various environmental conditions, including exposure to water jets. A spray shield is provided above the igniter to protect it from a dirstet spray.
The igniters in the PHMS are equally divided into two redundant groups, with 16 separate circuits per group, each with an independent circuit breaker and two igniters per circuit.
Each group has independent and separate control, power, and igniter locations that ensure adequate coverage even in the event of a single failure.
Manual actuation capability for each group is provided in the main control room (one switch per group), along with the status (on-off) of each group.
The igniters are powered from Class 1E power panels that have normal and alternate power supply from offsite sources.
In the event of a loss of offsite power, the igniters would be powered from the emergency diesel generators.
Group A igniters receive power from the train A diesels, and group B igniters from the train B diesels.
In addition, the igniters will be seismically supported.
i The permanent hydrogen mitigation system installed in Sequoyah Units 1 and 2 l
consists of 64 igniter assemblies distributed throughout the upper, lower, and ice condenser compartments.
i Following a degraded core accident, any hydrogen that is produced would be re-leased into the lower compartment.
To cover this region, 22 igniters (equally divided between trains) will be provided.
Eight of these will be distributed on the reactor cavity wall exterior and crane wall interior at an intermediate elevation.
Two igniters will be located at the lower edge of each of the five steam generator and pressurizer enclosures, two in the top of the pressurizer enclosure, and another pair above the reactor vessel in the cavity.
11/27/82 22-3 SEQUOYAH SSER6 SEC 22
~
o Any hydrogen not burned in the lower compartment would be carried up through the ice condenser and into its upper plenum.
Because steam would be removed from the mixture as it passes through the ice bed, thus concentrating the hydragen, mixtures that were nonflammable in the lower compartment would tend to become flammable in the ice condenser upper plenum.
This phenomenon is supported by the CLASIX containment analysis code, discussed later in this SSER, which predicts that more sequential burns will occur in the upper plenum than in any other region.
Controlled burning in the upper plenum is prc.ferable because upper plenum burns involve smaller quantities of hydrogen and allow for the expansion of the hot gases into the upper compartment, thereby reducing the peak pressure.
TVA has chosen to take advantage of the beneficial characteristics of combustion in the upper plenum by distributing 16 igniters around it.
The igniters are located on the containment shell side of the upper plenum at 16 equally spaced azimuthal locations. To handle any accumulation of hydrogen in the upper com-partment, four igniters will be located in the upper compartment dome. Addi-tional igniters are located at lower elevations in the upper compartment to take advantage of upward flame propagation at lower hydrogen concentrations; specifically, four igniters are located near the top and inside the crane wall, and one is located above each of the two air return fans.
The air return fans provide recirculation flow from the upper compartment through the dead-ended volume and back into the lower compartment.
To cover the deadended region, there will be a pair of igniters in each of the eight rooms through which the recirculation flow passes.
The staff has reviewed the number and locations of igniters provided in the PHMS and finds the system layout acceptable.
The staff notes, however, that the PHMS system would be improved by locating the upper plenum igniters alternately between the containment shell side and the crane wall side of the upper plenum in a staggered fashion, and locating additional igniters at lower elevations in the upper compartment.
Installation of upper plenum igniters in a staggered arrangement will further reduce the likelihood of flammable mixtures bypassing the igniters, while additional upper compartment igniters would provide added assurance that any burning in the upper compartment will occur as discrete burns at low hydrogen concentrations characterized by upward flame propagation, 11/27/82 22-4 SEQUOYAH SSER6 SEC 22
~
rather than as a global burn. TVA is unable to relocate upper plenum igniters or add more upper compartment igniters during this refueling of the Sequoyah Unit 1.
The staff may require TVA to relocate the upper plenum igniters in a staggered arrangement before restart following the next refueling for Sequoyah Unit 1 depending on the outcome of certain confirmatory testing as detailed in Section 10.
The staff considers the present igniter locations to be acceptable for operation during the interim period.
The staff also finds that the adequacy of the number and locations of the upper compartment igniters should be con-firmed on the basis of certain large-scale confirmatory tests to be conducted at the Nevada Test Site in early 1983 as part of a joint Electric Power Research Institute (EPRI)/NRC hydrogen research program.
These tests will include dynamic simulations of degraded core accidents at a scale comparable to the actual con-tainment building, and will serve to identify scale effects on combustion phenomena.
Upon completion of those tests, the staff will provide recommenda-tions regarding the adequacy of the upper compartment igniter coverage and any required design enhancements.
With respect to operating procedures, the TVA emergency operating instructions direct the operator to actuate the PHMS following any reactor trip or safety injection initiation.
These directions are included in the immediate actions of the diagnostic procedure used following reactor trip or safety injection, and actuation of the PHMS 15 verified in the procedure for responding to a l
loss-of-coolant accident (LOCA).
Thus, the operator will have sufficient time to manually actuate the PHMS for any event in which it would be required.
As recommended in SSER 5, the air handling units used for normal refrigeration in the ice condenser will be tripped for both units for accidents in which the PHMS is actuated.
The procedures call for PHMS to remain actuated until the unit reaches safe cold shutdown, and any threats as a result of hydrogen release are eliminated.
The staff concludes that these procedural instructions are adequate for actuation and termination of the PHMS.
In addition, the emergency operating l
instructions will be upgraded in response to TMI Action Plan Item I.C.1 and Commission Action on SECY-82-111.
The upgraded instructions will address operation of hydrogen integrity systems based on inadequate core cooling symp-toms and containment pressure and hydrogen concentrations.
The Tayco igniters have been subjected to endurance testing for a period of approximately 2 weeks.
l l
l 11/27/82 22-5 SEQUOYAH SSER6 SEC 22
l To ensure that the permanent hydrogen mitigation system will,"unctinn as intended, TVA has proposed a preoperational and surveillance testing program similar to that performed for the IDIS.
Preoperational testing, to be performed before restart after refueling, will verify that the current drawn by each group of igniters is within tolerance, and that the temperature of the igniter is at least 1800*F.
During the preoperational tests the current in each circuit will be measured and the results used as the baseline for future surveillance tests. The igniter system will be subjected to periodic surveil-lance testing; this testing will consist of energizing the PHMS in the main control room and taking current readings of the circuits.
If the current read-ings do not compare favorably with current measurements taken during preopera-tional testing, all igniters will be individually inspected to ensure their operability.
The staff will also require that igniter temperatures be measured at specified intervals.
The operability of at least 31 of the 32 igniters per train will maintain an effective coverage throughout the containment, if there are no inoperable igniters on corresponding redundant circuits that provide coverage for the same region.
The two trains of igniters should be operable during operational modes 1 and 2.
3 Combustion / Igniter Testing In support of the IDIS, TVA, Duke Power, and AEP conducted two testing programs to obtain information pertinent to the performance characteristics of the glow plug igniters.
Preliminary screening and qualification testing was performed at TVA's Singleton Laboratory.
Combustion tests using the glow plugs were performed by Fenwal, Inc. to study igniter performance under various environ-mental conditions (Cross, 1980; Mills, 1981).
Based on the results of these programs, the staff concluded in Sequoyah SSER 4 that the glow plug igniter would perform its intended function under various conditions.
During the past 2 years, to further evaluate the efficacy of ignition systems and to investigate possible enhancements to proposed deliberate ignition systems, the ice condenser utility owners and the Electric Power Research Institute (EPRI) have sponsored several test programs.
This work is summarized 11/27/82 22-6 SEQUOYAH SSER6 SEC 22
in the TVA Executive Summary Report dated September 27, 1982.
Basic combustion and igniter studies were conducted in a test program conducted at the Whiteshell Nuclear Research Establishment to evaluate the glow plug and Tayco igniters, along with testing to investigate the following items:
lean mixture combustion, rich mixture deflagrations, fan-and obstacle-induced turbulence, and the effects of a compartmentalized geometry.
To determine if a water spray / fog consisting of smaller water droplets than conventional containment spray systems would improve the overall performance of deliberate ignition systems, the utilities sponsored testing with the Factory Mutual Corporation and Acurex Corporation.
Factory Mutual investigated in a small-scale facility the pressure suppression effects of a small droplet spray / fog. Acurex addressed the same phenomenon, as well as the effects of igniter location in a larger scale vessel.
3.1 The Whiteshell Test Program The experimental program carried out at Whiteshell consisted of small-scale igniter testing and large-scale combustion testing (Mills, 1982a, b, c; Kammer, 1982).
Small-scale tests were performed in a 17-liter vessel to investigate the effect of igniter surface temperature and type on the lower flammability limits of lean hydrogen-air-steam mixtures. The small-scale test program consisted of three phases. Data en the lower flammability limit were obtained in Phases 1 and 2 using a GMAC-7G glow plug operating at 14 and 12 V, respectively.
In Phase 3 the tests were repeated using the Tayco igniter.
Hydrogen concentra-tions were varied between 4 and 15%, and steam concentrations varied between 0 and 60% in all three phases.
Evaluation of the experimental data indicates that for quiescent mixtures, igni-tion occurs below hydrogen concentrations of 8.0% for steam concentrations of up to 30%. Consistent with other test data for steam concentrations above 30%,
the flammability limit was shifted upward to higher hydrogen concentrations.
The igniter consistently initiated combustion of mixtures with steam concentra-tions up to approximately 60%.
Experimental results showed reliable ignition 11/27/82 22-7 SEQUOYAH SSER6 SEC 22
l of turbulent mixtures with hydrogen concentrations of 5%, even for steam con-centrations up to 40%.
The surface temperature of the igniter at the time of ignition ' as measured for each test.
For dry mixtures, the Tayco igniter surface temperature at ignition was approximately 1200*F.
Test data show that the igniter surface temperature at the time of ignition increases with steam concentration.
This is consistent with the trend observed for GM glow plug igniters.
The large-scale tests were performed at Whiteshell using a 7.5-ft diameter sphere. The purpose of the tests was to investigate four different items: lean mixture combustion, rich mixture deflagrations, fan-and obstacle-induced tur-bulence, and compartmentalized geometry effects.
Spark ignition was used in these tests.
In Phase 1 of the program, lean mixture tests were performed in the sphere to investigate the combustion phenomena under various conditions of steam and fan-induced turbulence.
Hydrogen concentrations were varied from approximately 5 to 10 volume percent, and steam from 0 to 30%.
Fans were activated in several of the tests.
I Test results for quiescent mixtures with bottom ignition indicate that l
combustion was initiated at a 5 volume percent hydrogen concentration.
Only i
(
l about 20% of the hydrogen was burned at this concentration.
For an 8% hydrogen concentration, virtually complete combustion was observed.
These re'sults are in general agreement with previously published data on the flammability of lean mixtures.
Tests with steam present show that the addition of 15% steam does not have a significant effect on the completeness of burn.
Results obtained with fans activated confirm that turbulence enhances the rate and completeness of combustion.
An increase in peak pressure to the corresponding adiabatic value was also observed.
These findings corroborate the results of tests at Fenwal (Cross, 1980; Mills, 1981), Lawrence Livermore s
National Laboratory (LLNL) (NUREG/CR-2486), and Sandia National Laboratory (Roller and Falacy,1982), but more importantly they indicate that turbulent plant conditions will promote burning at relatively lean concentrations.
11/27/82 22-3 SEQUOYAH SSER6 SEC 22 e
Ouring Phase 2 of the Whiteshell program, a series of rich mixture deflagration tests was performed to supplement existing knowledge of combustion of hydrogen-steam-air mixtures at high hydrogen concentrations and to confirm that detona-tions would not result.
For these tests, hydrogen concentrations were varied from 10 to 42 volume percent, and steam from 0 to 40 volume percent.
Fans were activated in several tests.
Complete combustion was achieved in nearly all tests, including those with a quiescent mixture of 10% hydrogen and 40% steam.
For both dry mixtures and mixtures with steam present, the measured pressure was always less than the theoretical adiabatic pressure.
rhis same result was observed in Sandia com-bustion tests conducted as part of the NRC research program.
Furthermore, no detonations were observed even at stoichiometric and higher concentrations of hydrogen which are classically considered to be detonable.
The absence of detonations is attributed to the fact that the energy release rate of the igniter is significantly less than that required to initiate a detonation.
In Phase 3 of the Whiteshell test program, the effects of turbulence induced by fans and gratings on the extent and rate of combustion were investigated.
Hydrogen concentrations ranged from 6 to 27 volume percent in these tests.
Results show that for rich mixtures, forced turbulence increases the rate of pressure rise but does not increase the peak pressure. With regard to the effect of gratings, the test results indicated that in lean mixtures without fans, the presence of gratings tended to increase the magnitude and rate of pressure rise.
At high concentrations or with fans, the gratings reduced both the magnitude and rate of pressure rise by acting as heat sinks.
In summary, the Phase 3 results indicate that no unanticipated pressure effects result from forced turbulence, even at high concentrations of hydrogen.
In the fourth and final phase of the Whiteshell program, compartmentalized geometry effects were investigated.
Two connected compartments were simulated by attaching a 20-ft long, 1-ft-diameter pipe to the 7.5-ft-diameter sphere.
The effects of igniter location and unequal concentrations in each vessel were investigated for hydrogen concentrations ranging from 6 to 25 volume percent.
Two igniter locations were used, one at the end of the pipe and the other at 11/27/82 22-9 SEQUOYAH SSER6 SEC 22
i s
J s
j the center of the sphere..
For all tests, no detonations occurred, and the i
obserded peak pressures were less than the calculated. adiabatic values. With I
regard to, tests 'with unequal concentrations, no signific.ent effects of
/
pronagating ' flames.between two connected vessels were observed.
(.
3.2 The Factory Mutual /Acurex Test Program To determine whether a water spray consisting of droplets smaller than conven-tional spray systems would improve the overall performance of deliberate ign'i-tion system, a two part experimental program was carried out under the spoksbr-I ship of EPRI.,The Factory Mutual Corporation (FM) project was the first of'the two part program'(Mills,,1982a).
The purpose of the FM project was to evaluato-the effects of water fog density, droplet diameter, and teinoerature on th{
lower filammability limit of hydrogen-air-steam mixtures.
The FM work &lso served to identify a set of nominal conditions for the intermediate-scale' hydrogen, combustion studies dealing with +.he pressure-suppressant ef fects of -
fog. 'IM intermediate scale studies were conducted by. the Acur' x Corporation e
(ibid) an1 were the second part of the two part program.
, /
7 The FM tests were conducted in a plexiglass tube approximately 3.5-ft long and 6-in. ira diarweter.
A 2.8-Joule spark served'as the ignition source.
ber tl
- tests were alss. conducted with a GMAC-7G glow plugpas the ignition source to s
verify the applicability of these tests tolth0 installdd distributed ignition I -
p systems.
Five different spray nozzles were Ged to outair different fog con-ditions (i.e., different characteristid droplet sizes and densities). Mean droplet sizes from/approximately 10 ~to 160 micror.s were investigated at fog concentrations up to 0.1 volume pe'rcent.
Tests were conducted at water temperatures of 20 C (69.8*F), 50*C '(122 F), and 70 C (158 F), and hydrogen concentrations ranging from approximately 4 to 12 golume percent.
Results df the FM tests confirmed the analyticaf prediction that increased fog densities are required to achieve a given level of fog inerting when the characteristic droplet size is increased.
Test results showed that at ambient temperature, visually dense water fogs had only a marginal effect on the hydroger, lower flammability limit. At higher fog temperatures, somedhat larger 11/27/82 22-10 SEQUOYAH SSERC SEC 22
{
\\
increases in the flammability limit were observed.
TVA reported favorable agreement between the FM experimental data and theoretical models used to describe the effect of fog on hydrogen combustion.
As a follow-on to the small-scale FM tests, the effects of fogs and sprays on the characteristics of deflagration were investigated in larger scale tests conducted by Acurex.
A 630-ft3 vessel approximately 17 ft high and 7 ft in diameter was used for all tests.
The tests were carried out in two phases.
All Phase 1 tests were dynamic tests with the glow plug preenergized.
Tests were conducted with hydrogen injection, hydrogen / steam injection, and hydrogen /
steam injection with water spray present.
These tests investigated the effect of igniter location with igniter assembli E iocated near the top, at the center, or near the bottom of the test vessel.
The results of the Phase 1 Acurex tests suggest that lowering the igniter location produces milder pressures during hydrogen combustion.
This appears to be a result of increasing the fraction of the vessel volume exposed to upward propagating flames in lean hydrogen concentrations.
For these dynamic tests, repeated burns were produced with pressure increases of 1 to 6 psi; for several tests without sprays, the pressure rises were higher, with a maximum increase of 28 psi. Because the Phase 1 tests were transient in nature, combustion parameters such as hydrogen concentration at ignition and completeness of burn were not conclusively determined.
During Phase 2 of the project, Acurex investigated the effects of a water fog l
on the pressure rise that accompanies a deflagration.
Quiescent tests were l
conducted without water fog and with water fog at two different droplet sizes l
and concentrations.
Dynamic tests were conducted with hydrogen injection and with hydrogen / steam injection.
The igniter assembly was located near the bottom of the vessel for all tests.
In the Phase 2 tests, fogs were expected to reduce the pressure rise resulting l
from hydrogen combustion.
This was confirmed in the dynamic tests, but not in the quiescent tests.
The explanation offered by TVA for the lack of pressure 11/27/82 22-11 SEQUOYAH SSER6 SEC 22
l suppression in the quiescent tests was that the water fogs enhanced the rate of combustion, causing the deflagration to be more like an adiabatic burn.
For the transient tests conducted in Phase 2, the pressure increases from the repeated burns varied from 1 to 5 psi.
The small pressure rises are attributed to ignition occurring at lower hydrogen concentrations.
This conclusion is consistent with the Whiteshell findings that increased turbulence promotes ignition at lower hydrogen concentrations.
Because the containment post-accident environment would resemble the transient test conditions, the pressure rises associated with hydrogen combustion in containment are expected to be relatively benign, as observed in the transient tests.
3.3 Tayco Igniter Testing As discussed in previous supplements, the effectiveness and durability of the GM glow plug under endurance, cycling, and hydrogen combustion conditions has been demonstrated in testing conducted at Whiteshell and Singleton.
To show that the Tayco model igniter is comparable to the GM glow plug, equivalent tests have been performed on the Tayco igniter at Whiteshell and Singleton.
Tests of the Tayco igniter conducted as Phase 3 of the small-scale igniter test program at Whiteshell show that the Tayco igniter was as effective at igniting lean mixtures as the GM glow plug.
The results of the igniter surface tempera-ture tests suggest that the Tayco igniter is capable of igniting mixtures at surface temperatures 125F to 200F* less than the GM glow plug.
This could attributed to the helical geometry of the Tayco igniter, which may promote higher local gas temperatures within the coil.
Tests similar to those for the GM glow plug were conducted at Singleton to assess the durability of the Tayco igniters when they are subjected to endurance and cycling operations at minimum and maximum voltages and to hydrogen combustion.
To summarize, four Tayco igniters were subjected to a series of five tests.
These tests consisted of a 24-hour break-in at 120 V, continuous operation for 7 days at 120 and 135 V, and on/off cyclic operation at 120, 125, 130, and 135 V. Igniter surface temperature was monitored for the duration of the tests.
The stcady-state surface temperature remained above 1700 F through out the 11/27/82 22-12 SEQUOYAH SSER6 SEC 22
l l
test series.
The igniters were energized for a total of approximately 370 hours0.00428 days <br />0.103 hours <br />6.117725e-4 weeks <br />1.40785e-4 months <br /> each. All Tayco igniters performed successfully during the tests except for one which failed after 340 hours0.00394 days <br />0.0944 hours <br />5.621693e-4 weeks <br />1.2937e-4 months <br /> of operation.
Operation for 340 hours0.00394 days <br />0.0944 hours <br />5.621693e-4 weeks <br />1.2937e-4 months <br /> is considered acceptable because it is in excess of the expected time when igniter performance will be required.
Tests were also conducted in which the igniter was exposed to hydrogen combus-tion in flowing mixtures with entrance conditions ranging from 4 to 12 volume percent. The Tayco igniter initiated combustion and survived the burn environ-ment in all cases.
At the staff's request, additional tests were conducted at TVA's Singleton Laboratory to ensure that the Tayco igniter would operate as intended in a spray environment such as that in the upper compartment.
Tests were conducted using a single hollow cone spray nozzle of the san;e type used in Sequoyah and in the Fenwal spray tests for the glow plug igniter.
The nozzle was oriented vertically downward and was located 3 ft directly above the igniter.
The igniter was oriented horizontally and was mounted under a horizontal spray shield of the same configuration as those on the igniter assemblies to be installed in Sequoyah.
Igniter performance was assessed on the basis of measured surface temperatures for four different environmental conditions:
natural and fan-induced circula-tion, with and without spray.
In tests without sprays, the igniter surface temperature remained above 1700 F at all times. When the spray nozzle was activated, the igniter temperature dropped to 1650 F and 1600 F with the fan off and on, respectively.
Although the 1600 F surface temperature is above the maximum surface tempera-ture required for ignition as determined by Whiteshell, the staff considered the (70p in surface temperature significant, and requested that TVA provide additional assurance that the Singleton test conditions were representative of those expected in the plant.
Further TVA analysis of the Singleton horizontal plane at the igniter elevation was approximately equivalent to that which would l
l 11/27/82 22-13 SEQUOYAH SSER6 SEC 22
l be providad by operation of one of the two spray trains in the Sequoyah plant.
Moreover, because the majority of the spray flow with the hollow cone nozzle is concentrated at the periphery of the cone, the spray density directly below the test nozzle (i.e., at the location of the igniter assembly), would be even less than expected in Sequoyah with one spray operating. Therefore, in the view of the staff, these Singleton spray tests therefore did not adequately represent the containment spray environment.
When informed of the staff concerns, TVA suggested that earlier spray tests conducted at Singleton with the Tayco igniter spray shield removed also support the use of the igniter in a spray environment.
In these tests igniter surface temperatures of approximately 1500 F were observed. However, the staff con-siders these results inadequate for the final evaluation of the PHMS because (1) these tests were also conducted using the hollow cone nozzle, which obfuscates computation of spray density at the igniter with reasonable con-fidence, and (2) they did not address the synergistic effects of sprays and
" turbulence (fans).
The staff has indicated to TVA that additional spray tests are needed to confirm satisfactory operation of the Tayco igniter in a spray environment.
These tests must ensure that in a spray environment similar to that expected in containment, the igniter will sustain a surface temperature sufficient to initiate combustion in lean mixtures. The staff will require that such tests be completed to its satisfaction before it grants final approval of the Tayco igniter.
3.4 Staff Conclusions Regarding Testing The staff has reviewed the combustion testing programs conducted as part of the TVA research effort and concludes that the results support the use of a distri-buted ignition system for post-accident hydrogen control.
Specifically, the results of tests conducted at Whiteshell show that thermal igniters will reli-ably initiate combustion for a wide range of hydrogen-steam-air mixtures.
Tests conducted at higher hydrogen concentrations illustrate the difficulty in initiating detonations, even at stoichiometric and higher concentrations.
11/27/82 22-14 SEQUOYAH SSER6 SEC 22
9 l
l Also, the observed effects of steam, induced turbulence, connected geometries, and unequal concentrations on the nature of hydrogen combustion confirm the staff's previous understanding.
Tests conJacted at Factory Mutual and Acurex provide additional information on the pressure-suppression and inerting effects of sprays, and fogs.
Similarly, none of the results obtained in these studies would support a negative finding relative to the use of deliberate ignition system. With regard to the Tayco igniter, a number of tests remain to be completed to provide further confirmation that the igniter will operate as intended in a spray environment.
However, igniter tests conducted to date provide a basis for concluding that the GM and Tayco igniters are equivalent.
4 Hydrogen Mixing and Distribution Analyses discussed in SSER 3 have indicated that hydrogen released during a postulated degraded core accident could be expected to be reasonably well mixed by the time it leaves the lower compartment.
Adequate mixing, in conjunction with ignition of lean mixtures, would effectively preclude the formation of detonable concentrations.
However, previous containment mixing analyses were cursory in nature, and did not attempt to quantitatively characterize hydrogen mixing and distribution within the ice condenser containment. A series of large-scale tests were, therefore, conducted by the Hanford Engineering Development Laboratory (HEDL) as part of the EPRI research program to provide additional assurance that large hydrogen concentration gradients will not occur (Mills, 1982a).
The mixing tests were conducted at HFDL's Containment Systems Test Facility (CSTF). This facility has a vessel that is 67 ft tall, with a diameter of 25 ft.
Because the upper compartment of the ice condenser containment will be well mixed by the sprays, the lower compartment region was chosen for modeling emphasis in the facility.
The interior of the CSTF was modified to represent a divider deck, reactor cavity, refueling canal, the air return fans, and ice condenser lower inlet doors.
For the purposes of these tests, geometric similarity was retained between the tests compartment and the lower compartment of an ice condenser containment.
Hydrogen ar.1 steam release rates used in t
11/27/82 22-15 SEQUOYAH SSER6 SEC 22
o l
tests were scaled to model the base case S 0 loss-of-coolant accident (LOCA)*.
2 Helium was used to simulate hydrogen in most of the tests because of site safety considerations.
Atmospheric temperatures, velocities, and gas concentrations were measured at several distributed points during the tests.
The test matrix for the HEDL program was designed to characterize hydrogen distribution for two release scenarios: (1) a 2-in. pipe break with a hori-zontal orientation and (2) a 10-in. pressurizer-relief tank rupture disc opening with a vertically upward orientation.
Two different release rates were investigated. The test program included tests with and without air return fans.
The results of the HEDL tests show that good mixing in the lower compartment can be expected if the air return fans remain operational throughout the accident. The air recirculation fans minimize both the peak helium concentra-tion and the maximum helium concentration difference between points in the test compartment.
In all cases with forced air recirculation, which included the two jet orientations and two different release rates, the maximum helium concentration difference between all points in the test compartment was less than 3 volume percent at all times and was generally on the order of 2%.
These concentration differences had stopped increasing even before the release period was over and were less than 1 volume percent within 5 minutes after stopping the source gas.
The HEDL test with no forced recirculation (air return fans inoperative) were inconclusive.
During the helium-steam release for these tests, the maximum concentration difference between all measurement points in the test compartment was 2 volume percent.
Following the helium-steam release, however, the test compartment developed a vacuum as the steam in the compartment condensed.
This reverse migration coupled with the lack of a mi".ing mechanism from either the fans or the jet itself created a concentration difference of as much as 7
- A single degraded core accident designated as S D in WASH-1400 (NUREG-75/014);
2 it is a small-break LOCA accompanied by the failure of emergency core cooling injection.
11/27/82 22-16 SEQUOYAH SSER6 SEC 22
volume percent helium.
Although the later portion of the test may in no way be prototypical of the plant, as TVA contends, neither does it support a conclu-sion that adequate mixing will occur without forced circulation by the air return fans.
In assessing mixing in the latter portion of the test, however, it should be noted that for tests both with and without forced recirculation, the test compartment volume is well mixed with less than 1 volume percent concentration difference between points within 20 minutes after stopping the hydrogen-steam or helium-steam source.
Based on review of the HEDL results, the staff concludes that the formation of significant hydrogen concentration gradients in containment is unlikely if the air return fans survive the acci. dent environment.
The operation of the deliber-ate ignition system near the lower hydrogen flammability limit in conjunction with the mixing by the air return fans ensures that hydrogen concentrations at or below the flammability limit will be maintained throughout containment for the duration of the accident.
In this regard, the formation of detonable pockets of hydrogen is precluded.
5 Detonations The TVA position regarding detonation is that detonation is not a credible phenomenon in the containment because:
(1) there would be no rich concentra-tiens throughout the containment because the distributed igniters would initi-ate combustion as the mixture reached the lower flammability limit and because effective mixing would occur; (2) there are no high-energy sources to initiate a detonation; and (3) there are no areas of the containment with sufficient geometrical confinement to allow for the flame acceleration necessary to yield a transition to detonation.
The staff agrees with the TVA position.
Because of the well-mixed atmosphere in containment, as confirmed by the HEDL mixing tests, the potential for localized accumulation of significant concentrations of hydrogen is unlikely.
Even given that a high concentration might be formed locally, detonation of the cloud is extremely remote because this would require that the cloud encounter 11/27/82 22-17 SEQUOYAH SSER6 SEC 22
l an ignition source of sufficiently high energy to initiate a detonation before it passes through a region in which an igniter is located or before combustion is initiated.
The staff concluded in SSER 5 that the energy level of the ther-mal igniter is not sufficient to initiate a detonation.
This conclusion is supported by test data, including several of the tests recently conducted at Whiteshell and LLNL. Although these tests do not show conclusively that detona-tion or transition to detonation cannot occur, they do illustrate the difficulty involved in producing the phenomenon even using stoichiometric hydrogen-air mixtures such as those present in tests.
In the staff's view, the only scenario in which large concentrations of hydrogen might accumulate is one in which all igniters in a given region fail, along with the air return fan.
TVA has provided redundant igniters on separatt power trains in each region of the containment to preclude such an occurrence.
The staff thus concludes that detonation of local pockets of hydrogen is extremely unlikely.
Another concern related to the detonation issue is that of flame acceleration.
The phenomenon of flame acceleration as a possible mechanism for producing a detonation or large overpressures in containment was discussed in SSERs 4 and 5.
The concern, expressed by Sandia National Laboratory, was that obstructions in the ice condenser region of the plant may serve to accelerate combustion to the point that a transition to detonation would occur.
Utility consultants previously concluded and still contend that there are no areas in the contain-I ment that provide sufficient geometrical confinement to allow for the extreme l
l flame acceleration necessary to result in a transition to detonation.
For j
example, the vertical ice baskets in the ice condenser are not sufficiently confined radially and the circumferential upper plenum above the ice condenser is not sufficiently confined for transit on to detonation to occur.
d With regard to the ice condenser region of contair. ment, the utility consultant's view was that, for an S 0-type scenario, the upper plenum igniters would ignite 2
l the mixture as it first becomes flammable; then, as a richer mixture is vented to the upper plenum, the igniters will produce a horizontal standing flame.
If l
11/27/82 22-18 SEQUOYAH SSER6 SEC 22
I l
the mixture is further enriched, the flame will propagate downward into the ice bed until it settles to an equilibrium point where sufficient steam has been condensed.
TVA concluded that even if an inerted mixture with a high hydrogen concentration were introduced to the ice bed, which is highly unlikely because of operation of the lower compartment igniters and the air return fans, the flame front would simply propagate to an equilibrium elevation where sufficient steam was condensed to support combustion.
The flame propagation will not allow the hydrogen-steam-air mixture to dry out to the point where detonable mixtures would develop.
The staff previously considered these matters, as discussed in SSER 5, and concluded that a transition to detonation in the ice condenser region was not likely.
Results of recent research conducted at McGill University as part of the NRC Hydrogen Research Program support the TVA position that flame acceleration will not occur in an ice condenser containment.
In laboratory-scale studies of flame propagation through obstacle fields, McGill researchers have investigated the rate of flame acceleration as a fa ction of obstacle configuration and hydrogen concentration in dry air.
In these tests, noticeable flame accelera-tion and transition to detonation were observed only at hydrogen concentrations in ex ess of 13 to 15 volume percent.
This limit is lower than the often-quoted value of 18%, but is still well above the concentration expected in the containment building.
The requisite concentration may shift upward if steam is added to the mixture.
Furthermore, Sar.dia tests have confirmed tnat confine-ment of the gas mixture is a requisite condition for producing a transition to detonation.
The composite evidence of these relatively recent tests has led Sandia to conclude that a transition to detonation in the upper plenum region is unlikely.
The McGill findings are preliminary in nature, and additional tests are planned at both McGill and Sandia to address the effects of steam addition and scaling on the requisite concentration for flame acceleration.
However, the staff believes that the preliminary findings by McGill will not be significantly altered by additional tests and that they provide an adequate basis for licensing decisions.
11/27/82 22-19 SEQUOYAH SSER6 SEC 22
l l
Although the potential for detonation and flame acceleration is extremely remote, TVA has calculated the espense of the containment shell to a postu-lated local detonation of a 6-ft-diameter gas cloud and showed that a margin of safety of 3 exists before material yield would be reached.
The results of this analysis were reported in SSER 4.
At that time, further studies were thought to be necessary to bound the variation in pulse shapes to confirm the TVA findings. TVA was therefore required by license condition to address the potential for local detonation.
TVA has considered the potential and has concluded, based on the results of its research program, that detonations and transitions to detonations are not credible in Sequoyah.
TVA thus considers further studies of containment response to detonations unwarranted.
The staff agrees with TVA that detonations are extremely unlikely in Sequoyah and therefore feels the TVA position is rea'sonable for the licensing decisions related to the PHMS.
Even though the staff's view is that sufficient information exists for closure of the detonation issue, the staff with the support of Sandia has initiated an independent calculation of containment response to postulated local detonations.
Sandia, using the CSQ computer code in conjunction with a simple structural failure criterion, has calculated the effects of various postulated local deto-nations on the containment structure.
Results of early calculations for the upper plenum of an ice condenser plant indicate that containment integrity can be threatened if the requisite conditions for detonations were attained. As previously stated, however, it is the view of the staff that the conditions that must prevail to produce detonations are extremely unlikely.
Moreover, even with the presence of detonable mixtures, as assumed in the Sandia analysis, there has been no demonstration that a detonation would occur.
Subsequent calcula-tions performed by Sandia using a detailed structural model indicated the con-tainment would survive upper plenum detonations.
The Sandia investigation, which is not yet complete, is viewed by the staff as a confirmatory item to provide further insight into the consequences of local detonations.
The results of this effort are not expected to alter the staff's findings on the hydrogen control capability at Sequoyah for the aforementioned reason.
11/27/82 22-20 SEQUOYAH SSER6 SEC 22
l 6 Degraded Core Accidents and Hydrogen Generation As discussed in SSER 4, a small-break LOCA followed by a failure of emergency core cooling (ECC) injection (S 0) was selected by TVA as the base 2
case for evaluation of the hydrogen mitigation system.
Hydrogen release rates are a time-varying function whose average is of the order of 20 lbs per minute.
The staff considered these rates to be representative of releases that might be encountered in typical degraded core accidents less severe than total core melt or vessel failure, and considered them an acceptable upper limit basis for use in the interim evaluation; however, several concerns remained open.
Among these were: (1) the possibility that other scenarios might present schedules of steam and hydrogen release not covered by the analysis chosen; (2) that steam inerting might occur at some time during the sequence allowing large concentrations of hydrogen to develop; (3) that the recovery period might produce an exceptional burst of steam or hydrogen; or (4) that hydrogen might be released after the loss of the ice heat sink.
TVA was therefore asked to broaden the studies of steam and hydrogen releases.
In the follow-on CLASIX studies that were submitted by the applicant, steam and hydrogen releases were varied to correspond to higher release rates and releases after the ice had melted.
It was shown that a representative selection of sce-narios would be bounded by the calculated release rates, and thus it was claimed that a satisfactory group of alternative scenarios had been encompassed by the calculations. TVA states that the scenarios encompassed included an intermediate-break LOCA with a loss of ECC (5 0), a small-break LOCA with a loss of contain-1 ment heat removal (S G), a transient loss of main feedwater and loss of all ac 2
power (T B ), and a transient loss of main feedwater, loss of auxiliary feedwater, B2 and loss of the ECC (T LD).
B The staff has compared the release rates and sequences used in TVA's calcula-tions to those developed in an independent study of degraded core accidents in ice condenser plants carried out at Brookhaven National Laboratory (Yang and Pratt, 1982).
It is clear from this comparison that TVA's choices of hydrogen and steam release rates do indeed cover the above range of accident scenarios.
The highest rate of hydrogen release calculated by Brookhaven was of the order 11/27/82 22-21 SEQUOYAH SSER6 SEC 22
I of 1 lb per second. The Brookhaven calculations did not indicate that these rates would be exceeded during quenching or recovery from the degraded core conditions as well as in the initial core uncovery phase.
On the other hand, TVA has calculated the effect of hydrogen release rates as 6 lb per second under representative steam conditions, with and without ice.
In addition, the staff has compared the release rates chosen by TVA to those suggested in a proposed rule (" Notice of Interim Requirements Related to Hydrogen Control" (46 FR 62281)).
In this comparison, the release rates used by TVA were again found to be an adequate representation of the scenarios considered important ir, these degraded core situations.
The licensee's core reflood studies using MARCH, WFLASH, and LOCTA did not disclose any conditions that would be more adverse than the high release rates used in CLASIX.
The staff therefore finds TVA's treatment of scenarios to develop steam / hydrogen source terms in conformance to the requirements of existing hydrogen degraded core rules acceptable.
7 Sequoyah Containment Structural Capacity In support of the initial licensing of the plant, the ultimate pressure-retaining capacity of the Sequoyah steel containment was calculated by five different investigators.
These pressures ranged from a low of 27 psig to a high of 50 psig, as listed in column 2 of Table 22.1.
The variation was the result of the difference in the material properties used in the analysis, the stress limit criteria, and the manner of incorporating the horizontal and vertical stiffeners. When the material properties and the stress limit criteria are normalized to actual mean material properties and Von-Mises criteria, respectively, to form a uniform basis for coraparison, the ultimate capacity then varies from a low of 40 psig to a maximum of 60 nsig as listed in column 3 of Table 22.1.
To provide an adequate safety margin, the staff reduced its ultimate mean value of 60 psig by 3 standard deviations.
The standard deviation computation incorporated the variations in the material 11/27/82 22-22 SEQUOYAH SSER6 SEC 22
l l
Table 22.1 Internal static pressure capacity for hydrogen burning, psig Column 2, Column 3, Column 1, Reported Normalized Service ultimate ultimate Investigator Level C*
capacity **
capacityt TVA 38 40 Staff 30.0 36tt 60 (Ames Laboratory)
Franklin Research 30 51 Offshore Power 50 53 R&D Associates 27 40
- Based on ASME Code methods and Code allowables; 1/2-in. steel plate controls.
- Reported by individual investigators and summarized in NUREG/CR-1891.
tCapacity values normalized using actual mean material properties instead of Code values and Von-Mises yield criterion.
ttBased on actual material properties and Von-Mises yield criterion; this value is the mean value minus 3 standard deviations.
properties, material sizes and thicknesses, stiffener spacing, and containment shell diameter.
The standard deviation of the containment pressure was calculated to be 8 psig.
Therefore the ulitmate capacity of the containment adopted by the staff was 36 psig, which represents a lower bound value. An assessment of the containment penetrations was also made at the initial licensing stage and showed that the penetrations were not the controlling item for the containment ultimate pressure capacity, as reported in SSERs 3 and 4.
The proposed rule, " Interim Requirements Related to Hydrogen Control," was published after these analyses. The proposed rule would require that the 11/27/82 22-23 SEQUOYAH SSER6 SEC 22
l l
hydrogen control system perform its function without loss of containment structural integrity.
For the PHMS installed at Sequoyah, the rule would require that the containment pressure throughout the accident transient remain at or below that which corresponds to Service Level C limits of the Boiler and Pressure Vessel Code of the American Society of Mechanical Engineers (ASME Code).
The staff's consultant, Ames Laboratory, computed the value of the internal pressure that would produce stresses in the steel shell corresponding to Service Level C Limits as specified in the ASME Code,Section VI, Division 1.
This value is 30 psig and is shown in column 1 of Table 22.1.
This value is based on the finite element analysis model used in computing the containment ultimate capacity reported earlier.
The limiting section in the Ames Laboratory analyses is the 1/2-in. thick cylindrical plate between elevations 756 ft 3 in.
and 610 ft 3 in.
TVA has,also made an evaluation of the reinforced concrete floor that divides the upper and lower compartments.
This evaluation showed the reinforced concrete floor differential pressure capacity to be equal to or greater than the containment shell capacity.
The staff concludes that the estimated pressure retention capability for ASME Boiler r d Pressure Vessel Code Service Level C limits is 30 psig with e.11 of the inherent safety margins of the code implied.
8 Containment Analysis t
l 8.1 Containment Codes Calculations of containment atmospheric pressure and temperature have been performed using the CLASIX computer code developed by Westinghouse Offshore Power Systems (Westinghouse OPS-36A31).
Descriptions of the earlier version of CLASIX have been previously reported in SSERs 3, 4, and 5.
As noted in SSER 5 and as part of the license condition, the staff asked TVA to provide improved calculational methods for containment pressure and temperature response to 11/27/82 22-24 SEQUOYAH SSER6 SEC 22
l hydrogen combustion.
Specifically, TYA was to refine CLASIX to permit the addition of structural heat sinks and the separate modeling of the upper plenum.
In addition, TVA was to provide additional verification of the CLASIX code by comparison with results from other accepted codes and combustion tests. The present and latest version of CLASIX incorporates those changes requested by the staff.
The CLASIX code is a multivolume containment code that calculates the contain-rrent pressure and temperature response in the separate compartments.
CLASIX has the capability to model features unique to an ice condenser plant--including the ice bed, recirculation fans, and ice condenser doors--while tracking the distribution of the atmosphere constituents-- oxygen, nitrogen, hydrogen, and steam. The code also has the capability of modeling containment sprays. Unlike the earlier version, the present version of CLASIX includes heat sinks and models the upper plenum as a separate model.
Mass and energy released tc the contain-ment atmosphere in the form of steam, hydrogen, and nitrogen is input to CLASIX.
The burning of hydrogen is calculated in ti.e code with provisions to vary the conditions at which the burn initiates and propagates to other compartments.
CLASIX input for each compartment consists of the net free volume, temperature, contents by constituent, burn control parameters, and passive heat sink data.
The burn control parameters include the hydrogen concentration and oxygen con-centration required for ignition, the hydrogen concentration for propagation, the hydrogen fraction burned, and the minimum oxygen concentration required to support combustion and the burn time.
The flow area, flow loss coefficient, and propagation delay time for each intercompartment flow path is also required.
Additional input data are supplied to describe the ice condenser, fans, and sprays.
l l
The major difference between the present and earlier version of CLASIX is in the heat sink model.
The analytical model of the structural heat sinks repre-sents all heat sinks as multilayered slabs.
Heat transfer to the exposed sur-faces by both convection and radiation is modeled.
Radiation is assumed to occur only between the water vapor in the containment atmosphere and the l
11/27/82 22-25 SEQUOYAH SSER6 SEC 22
l surface of the heat sinks. A conventional finite difference formulation is used to model internal heat transfer.
The staff, with the support of Los Alamos National Laboratory (LANL), has completed a preliminary assessment of the CLASIX code. This assessment involved an evaluation of the validity and adequacy of the assumptions and models employed and review of the TVA-supplied comparisons between CLASIX results and those for other containment codes and combustion experiments.
A number of technical concerns were identified during the code review. With regard to the CLASIX radiation model, the staff requested that TVA clarify the expression used to compute the net radiant heat exchange.
Specifically, the staff questioned the inclusion of gas and wall emissivities as multipliers on the temperature terms.
TVA has reviewed the development of radiation model and has concluded that use of the emissivities is inappropriate.
However, TVA notes that use of the emissivities results in an underestimate of radiant heat flux to the walls and thus leads to conservative containment temperature and pressure predictions. Based on an independent review and on the TVA clarifica-tion provided, the staff and LANL concur in the TVA finding that CLASIX under-predicts the radiation heat transfer.
LANL, as part of its review, also identified a number of questions regarding l
the fluid flow equations used in the code.
LANL's concerns centered on the use of (1) steady-flow equations to describe the transient phenomena and (2) constant loss coefficients for subsonic flows. The rationale provided by TVA l
for the CLASIX flow equations is that the Mach number for all CLASIX cases analyzed to date has been less than the commonly accepted criteria for assuming l
incompressible flow.
On this basis, the staff and LANL agree that the CLASIX approach is valid.
To increase the level of confidence in the CLASIX code, TVA has validated CLASIX by comparing calculated results with the calculated results of the Westinghouse C0C0 CLASS 9 code (Westinghouse, 1981), the Westinghouse Transient Mass Distribution (TMD) code (WCAP-8077,-8078), and the measured results of selected l
Fenwal and LLNL tests.
1 11/27/82 22-26 SEQUOYAH SSER6 SEC 22
1 C0C0 CLASS 9 is based on the NRC-accepted code C0CO (WCAP-8326, 8327), and has been used in support of licensing activities for dry containments. The C0C0 CLASS 9 analytical model has the capability to simulate heat transfer to passive heat sinks and containment sprays as well as high enthalpy water mass and energy addition.
However, the C0C0 CLASS 9 model provides only a single volume representation of containment, and does not allow spray evaporation as CLASIX doen. Also C0C0 CLASS 9 does not have the capability to model the addition of hydrogen during a burn.
This limitation precludes comparison of a transient burn case.
Comparative runs were made with CLASIX and C0C0 CLASS 9 assuring no heat sinks, heat sinks, and heat sinks with radiation.
The comparison indicates that despite the previously cited discrepancy in the CLASIX radiation model, the two codes produced almost identical results for all cases considered.
TVA attributes the excellent agreement to the use of a similar heat transfer model in C0C0 CLASS 9.
The TMD program was developed for analyses of the ice condenser containment response during the initial few seconds following the onset of a design-basis LOCA. TM0 contains a multicompartment analytical model but does not include models for containment sprays, air return fans, heat sinks, or hydrogen addition.
Therefore, the comparison of CLASIX and TMD results is limited to multicompartment pressure and temperature responses to high enthalpy water mass and energy addition.
This comparison provides verification of CLASIX pressure.
And temperature response calculations, flow path calculations, and certain aspects of the ice condenser model.
Four CLASIX-TMD comparison runs were made covering the anticipated range of the blowdown energy from saturation to superheat conditions. A containment similar to the Sequoyah ice condenser plant was modeled in all cases.
Direct compari-sons were made between the calculated temperature and pressure responses.
Comparisons indicate that the two programs are in excellent agreement with the CLASIX-calculated values for both temperature and pressure being generally more conservative.
CLASIX is expected to be conservative relative to TMD because of differences in the treatment of breakflow as the flow enters containment.
11/27/82 22-27 SEQUOYAH SSER6 SEC 22
l For the final part of the verification, CLASIX was used to model hydrogen combustion experiments conducted at Fenwal and LLNL. These comparisons provide limited verificat,fon for such features in CLASIX as the hydrogen burn model, the models for hydrogen and high enthalpy water mass and energy addition, and to some extent the passive heat sink and containment spray models.
The apornach taken to establish CLASIX input data for the experimental simulation was to utilize t.s the fullest extent possible all reported test measurements for the selected experiments. This included the CLASIX initial conditions as well as burn parameters such as the fraction of hydrogen burned and the burn time.
A total of 17 tests were selected for CLASIX verification.
These included six dry tests'and nine steam tests reported by Fenwal and LLNL.
Hydrogen concentra-tions for both the dry and steam tests ranged from 8 to 15 volume percent.
Steam concentrations for the latter tests ranged from 5 to 10 volume percent.
In addition, one transient test and one test with spray were analyzed.
Conparison of CLASIX-calculated results with those measured in tbe tests in-dicated that CLASIX predictions for peak pressure are consistently higher than those measured in the tests. Temperature comparisons were not attempted because of the slow response time of the thermocouples used in the tests. Only in a few cases were the CLASIX-calculated pressures higher than those measured.
This was attributed to inaccuracies or inconsistencies in the estimated burn fractions.
In addition to its limited assessment of the CLASIX code and the TVA-supplied comparative runs, the staff directed its contractor, LANL, to develop the capability to independently model containment response to degraded core accidents. The ultimate purpose of the LANL effort was to perform confirmatory l
calculations for Sequoyah and other ice condenser plants; however, comparison of the models and results for the LANL-developed code with those for CLASIX l
provides an additional basis for evaluating the adequacy of the CLASIX code.
A modified version of the NRC COMPARE code was developed by LANL to model containment response to degraded core accidents.
COMPARE was previously 11/27/82 22-28 SEQUOYAh SSER6 SEC 22
developed to perform confirmatory subcompartment analyses, and included cap-abilities required to analyze ice condenser containments, heat transfer to passive heat sinks, and the thermodynamics of atmospheres composed of steam, water, and ideal gases.
To apply the COMPARE code to the analysis of hydrogen burning in containments, several capabilities were added, specifically a new ice condenser door model, a fan cooler model, a sump recirculation heat ex-changer model, and a hydrogen burn model.
A complete and total evaluation of the hydrogen burn version of COMPARE was not performed during the LANL effort.
However, the applicability of the COMPARE code for the performance of subcompart-ment analyses has been evaluated rather extensively.
Verification of the models added to the subcompartment version of COMPARE were performed performed by LANL.
These evaluations show that the models provided results that are consistent with the original objective of the model.
A verification of the hydrogen burn analysis capabilities of the hydrogen burn version of COMPARE is also provided by the comparisons of calculated results with those obtained using the CLASIX code.,These comparisons, discussed below, indicate that similar calculated values of pressure and temperature are obtained even though the codes were developed independently and utilize different models.
Based on its assessment of models used in the CLASIX code, a review of com-parative runs provided by TVA, and the reasonable agreement bound between CLASIX and the hydrogen burn version of COMPARE, the staff concludes that use of the CLASIX code to predict ice condenser response to a degraded core accident is acceptable, if appropriate input values are used. Approval of the CLASIX code for application to this particular class of accidents does not, however, constitute NRC endorsement of CLASIX for applications involving other classes of accidents, or variations of CLASIX to model other containment types.
The staff will continue to assess the adequacy of the CLASSIX code as part of its ongoing confirmatory effort.
8.2 Containment Pressure and Temperature Calculations The approach taken by TVA to establish the acceptability of the hydrogen control system was to select an accident sequence based on its significance and charac-teristics from the standpoint of hydrogen threat, and to then parametrically 11/27/82 22-29 SEQUOYAH SSER6 SEC 22
l vary key aspects of the containment analysis. As in previously reported anal-yses, a small-break LOCA with failure of safety injection, the S D event was 2
chosen as the base case.
TVA has performed calculations of the containment pressure and temperature response to the base case scenario using the latest version of CLASIX and the releases calculated from the MARCH code.
For the base case calculation, TVA assumed a lower flammability limit of 8 volume percent hydrogen, a burn frac-tion of 85%, and a flame speed of 6 fps. Test data fro.li Fenwal and Whiteshell, as well as the literature on combustion, indicate that ignition in the turbu-lent post-accident environment will occur around 5 volume percent hydrogen, with a burn completeness of 30 to 40%. Test data and the literature also show that at an 8% hydrogen concentration flame speeds are between 1 and 3 fps rather than 6 fps. The assumptions of ignition at the higher concentrations with a faster flame speed result in a greater amount of energy being released over a shorter period of time, and thus are conservative. Another conservatism in the CLASIX analysis is the assumption that ignition will occur simultaneously at all igniter sites in a compartment. This assumption will act to further in-crease the calculated pressures and temperatures.
The results of the CLASIX base case analysis indicate that the hydrogen will be ignited in a series of 7 burns in the lower compartment and 30 burns in the upper plenum.
The burns occur over a 2500-second interval, with the 7 lower compartment burns intermixed, some concurrently, with 15 upper plenum burns over the first half of the interval. The peak calculated containment pressures and temperatures are 18.7 psig and 1245'F for the lower compartment, 18.1 psig and 257 F for the dead-ended region, 13.1 psig and 1220 F for the upper plenum, I
and 10.4 psig and 163 F for the upper compartment. The pressure in containment before the first burn was approximately 5 psig.
As a result of the action of engineered safety features such as the ice con-denser, air return fans, and upper compartment spray, the pressure and temper-i ature spikes were rapidly attenuated between burns. After the last hydrogen I
burn, which occurs at approximatley 7100 seconds into the accident, roughly 11/27/82 22-30 SEQUOYAH SSER6 SEC 22 i
l 780,000 lbs of ice are calculated to remain in the ice condenser section (representing at least 110 x 108 BTUs in remaining heat removal capacity).
In summary, the results of the TVA base case analysis show an increase in containment pressure as a result of hydrogen burns on the order of 13 psi, with the containment remaining well below the lower bound ultimate capacity of 36 psig. The analysis predicts the burning will occur in the lower compartment and upper plenum, thereby gaining the advantage of heat removal by the ice bed and venting to the large upper compartment volume.
It should also be noted that each burning cycle involved the combustion of only 30 lbs of hydrogen, or roughly 2 x 108 BTUs of energy addition.
By burning at a given concentration in the lower compartment (and upper plenum), there is also the advantage of burning less total hydrogen at a time because the combined volumes account for less than one-third of the total containment volume.
To more realistically assess the efficacy of the PHMS, a best estimate calcula-
, tion was performed by TVA assuming a lower flammability limit of 6 volume percent, a burn fraction of 60%, and a flame speed of 3 fps.
The best estimate case results in a peak containment pressure of 10.6 psig, which is below the 12 psig containment design pressure.
TVA has also performed sensitivity studies to determine the effects of CLASIX burn parameters, safeguards performance, and reduced igniter performance on the containment response.
To bound reported data regarding hydrogen combustion, a number of cases were analyzed in which burn parameters such as hydrogen concen-tration for ignition, burn completeness, and flame speed were varied either throughout containment or in selected compartments.
Ignition criteria analyzed ranged from ignition at 4% hydrogen with 40% burn completeness, to complete combustion at 10% hydrogen.
Flame speeds were varied from 1 to 12 fps.
Additional cases were run to assess the effects of partial operation of the containment air return fans and sprays, heat removal by ice, and hydrogen release rates.
In some of these cases several parameters were varied simultane-ously such as a case with partial fan and spray operations, and modified igni-tion criteria (see Table 22.2).
Finally, there were investigations of the l
11/27/82 22-31 SEQUOYAH SSER6 SEC 22
4 e
J Table 22.2 Containment sensitivity studies
- Calculated peak Calculated peak pressure (psig) temperature (*F)
LC UC LC UC Base case 18.7 10.4 1245 163 (14.2)
(14.4)
(1262)
(236)
Ignition criteria All ignition at 6% H,
12.8 8.9 805 148 2
60% burned All ignition at 10% H,
8.0 9.7 214 171 2
100% burned (8.6)
(8.9)
- (237)
(175)
Flame speed 1 fps flame 10.1 9.6 884 150 12 fps flame 23.5 10.8 1306 182 (12.4)
(13.2)
(1243)
(205)
Safeguards 1 fan, 1 spray 17.6 18.0 1159 606 operational, UC and DE ignition at 6% H,
2 60% burned No ice, 22.8 26.9 1132 548 UC ignition at 6% H,
(18.3)
(25.3)
(1236 (545) 2 60% burned Hydrogen release 3 x base case H:
19.1 15.3 1578 498 release rate Same as above with 24.9 25.3 1310 542 6 lbs/sec spike, no ice Reduced igniter performance UP ignition at 8% H,
17.6 10.5 1284 157 2
40% burn No LC ignition 7.8 9.2 214 153 LANL mechanistic burn model Conservative (see text)
(26.1)
(24.2)
(1585)
(513)
Best estimate (18.5)
(20.0)
(1382)
(360)
(see text)
- LC = Lower compartment; UC = upper compartment; DE = dead-ended region; UP = upper plenum All cases assume base case parameters except as noted; (
)=
results predicted by LANL using hydrogen burn version of COMPARE.
11/27/82 22-32 SEQUOYAH SSER6 SEC 22
l ffccts of such postulated phenomena as fogging reducing the burn completeness in the upper plenum and steam inerting the lower compartment.
As discussed in SSER 5, the staff requested that TVA quantitatively assess the formation of fog and its effect on the performance of the igniter system. With regard to the effect of fogs and sprays on combustion, analytical studies of tha requisite fog density and droplet size for inerting have been conducted by Wastinghouse, Sandia, and others.
Based on considerations of the heat of combustion and fog / spray droplet vaporization, these studies show that to fog insrt an otherwise flammable mixture, two conditions must exist simultaneously:
the fog density must be sufficiently high and the droplet diameter sufficiently small. The requisite fog density increases approximately as the square of the droplet diameter.
Both of these parameters vary as a function of the hydrogen concentration of the mixture.
In general, fog droplets on the order of 10 microns or less in diameter are capable of vaporizing completely in the flame front and quenching the flame.
However, if the majority of the droplets in the population are larger than 10 microns, the fog is not expected to significantly influence the flame structure and may in fact exhibit beneficial effects such as the suppression of combustion pressure and any detonation waves.
To determine the significance of fog with regard to the PHMS installed in Sequoyah, TVA conducted a study to identify the major fog formation and removal m chanisms within an ice condenser containment. Analysis revealed that the upper and lower compartments maintained lower fog concentrations than the upper plenum. When the hydrogen concentration reached the lower flammability limit in the lower and upper compartments, the calculated fog concentrations were wall below the calculated inerting limit.
For the upper plenum, the fog is predicted to increase the flammability limit slightly. When the hydrogen concentration reaches 8.0 to 8.5 volume percent hydrogen in the upper plenum, the calculated fog concentration is two times smaller than the required concentration for inerting.
Tha staff has reviewed the TVA analysis and the results of the fog / spray tests conducted in support of the deliberate ignition system.
Based on the informa-tion provided as a result of these investigations, the staff concludes that the 11/27/82 22-33 SEQUOYAH SSER6 Fec 22
presence of fogs and sprays in a post-accident atecsphere may affect the opera-tion of the PHMS by increasing slightly the concentration at which ignition is initiated, but will not preclude satisfactory operation of the PHMS, because ignition is still expected to occur with acceptable consequences.
The staff notes that even though there is still reasonable assurance that reliable igni-tion will be achieved with the PHMS, reduced ignitor performance has been as-sumed by TVA in CLASIX containment analysis, and the results have been found acceptable.
After the issuance of SSER 4, TVA performed sensitivity studies of the hydrogen release rates and has computed the hydrogen release rates for a number of other accident sequences using the MARCH code. Two different sensitivity cases were considered.
In the first, a hydrogen release rate three times that of the base case was assumed for the period up to and including the maximum release rate (spike). To provide equivalent hydrogen mass additions, the duration of blow-down following the spike was correspondingly decreased.
Fcr conservatism, the steam releases were not changed, because additional steam would act as a burn heat sink.
In the second case, the hydrogen release rate was similarly assumed to be three times the base case; however, a maximum release rate of approximately six times the base case value was assumed.
The CLASIX code was used to analyze the containment response for the two cases:
first assuming ice to be present, and assuming all the ice melted.
The highest peak pressure predicted by CLASIX for all the sensitivity runs was 27 psig.
This pressure is well below the lower bound pressure capacity for the Sequoyah containment.
The results of selected CLASIX sensitivity analyses are summarized in Tsole 22.2, along with the results predicted by LANL using the hydrogen burn version of the COMPARE code.
Comparison of the CLASIX and COMPARE results indicates excellent agreement between the two codes.
The peak containment pressures calculated by COMPARE are consistently lower than comparable CLASIX values, illustrating the conservatisms in CLASIX.
The peak temperatures calculated by COMPARE are gen-erally equivalent to those calculated by CLASIX but in some cases are slightly higher.
11/27/82 22-34 SEQUOYAH SSER6 SEC 22
In conclusion, the results of the CLASIX sensitivity analyses demonstrate that (1) the effect of ignition criteria on containment pressure is dominated by the corresponding changes in burn location and sequence, but within the parameter ranges considered it does not result in peak pressures significantly greater than for the base case; (2) flam speed has a considerable effect on contain-ment pressure but does not pose a threat to containment integrity even for conservative flame speeds; (3) partial versus full operation of the air return fans makes little difference in the calculated results; (4) ice condenser heat removal is effective in reducing containment pressure; (5) the rate of hydrogen release has little effe t on the peak containment pressure; and (5) even with reduced igniter efficiency or lower compartment inerting, the PHMS will continue to perform its intended function.
It should be noted that the cases with no ice are not mechanistic, i.e., they are not representative of the S D scenario.
2 However, these cases importantly demonstrate that even without ice, the contain-ment pressure with the assumed igniter operation remains below the containment pressure capacity. This serves to indicate some insensitivity to whatever accident scenario is chosen.
8.3 Confirmatory Analysis and Conclusion At the request of the staff, LANL has performed confirmatory analyses for the base case and several other cases using the hydrogen burn version of COMPARE.
Code input equivalent to that for the CLASIX code was used ;, the confirmatory analyses with one exception.
In the LANL analyses, the ice condenser section was represented by four separate nodes each accounting for one-fourth of the ice condenser volume; this is a finer model representation of the ice bed than l
used in CLASIX. The hydrogen burn parameters for the ice condenser and lower plenum nodes were specified to preclude the initiation of independent burns but to permit burning by propagation if the hydrogen concentration exceeded 8 volume percent.
Agreement between COMPARE and CLASIX was quite good, with COMPARE predicting peak pressures throughout containment of 14 psig.
The mass of ice left in the ice condenser after the last burn is estimated at 289,000 lbs.
This value is less than predicted by CLASIX because more burning in the ice bed region is 11/27/82 22-35 SEQUOYAH SSER6 SEC 22
predicted by COMPARE, but is not a safety concern because the remaining ic'e reprasents adequate heat removal capacity.
The TVA sensitivity studies indicate that containment integrity will be main-tained for the base case and all sensitivity variations considered;'however, upper compartment burns occurred in only two of the TVA cas'es.
The subject of burning in the upper compartment was previously identified as a staff concern.
Staff interest in this area lies intne fact that ignition in the large, relatively open upper compartment. conceivably represents the largest-energy release rate by combustion and thus the greatest threat to containment.
Although the TVA upper compartment burns did not result in excessive pressures, the ftaff asked LANL to further investigate this phenomenon.
In response to the NRC request, LANL performed a number of additional sensitivity analyses using the modified COMPARE code.
The approach taken by LANL was to s
identify the combination of burn parameters required to produce a maximum contain-ment pressure, and then to assign paratieter values based.on a mechanistic burn model that is substantiated by test.
Independent burn initiation in the uppe,r,-,
comapartment was identifed as necessary to produce maximum pressures._
i The model used by LANL to establish parameter values for the, COMPARE contain-ment analyses is based on estimates of turbulence levels arid fluctuations and their relationship to eddy diffusivity and burn velocity.
The controlling rate mechanism for the transport of the hydrogen from its source to~an igniter can,
\\
in general, be estimated by using turbulence theory.
The rate of burning for the lean mixtures under consideration is also controlled by the turbulence level. The level of turbulence is estimated by summing all of the disiipation sources (sprays, fans, jets, natural convection, etc.) and by using a formula-tion that relates the turbulent kinetic energy, mixing ' length, and eddy diffusi-vity to the rate of dissipation of kinetic energy.
The turbulence model was used to estimate the mean concentration at the initiution of burning and the flame speed for the ice condenser containment burn analyses in which the first burn occurred in the upper compartment.
s 11/27/82 22-36 SEQUOYAH SSER6 SEC 22
Two COMPARE calculations were performed to assess the significance of upper compartment burning.
Burn' parameters for these runs were specified so that burning could only initiate in the upper compartment, but could propagate into any compartment in which the hydrogen concentrations is greater than 4.1 volume percent.
The first COMPARE run conservatively assumed ignition at 5% hydrogen with 40% burn completion, and a flame speed of 30 fps.
The second run assumed the best estimates for these pa'rameters based on the mechanistic burn model, i.e., ignition at 4.2% hydrogen hith 10% burn completion, and a flame speed of i _
,16 fps.
Results of these calculations, summarized in Table 22.2, show that for L*
bota cases peak pressures will remain below the estimated failure pressure.
A The staff concludes that the CLASIX containment analysis performed by TVA and confirmed in part by LANL provides an adequate basis for concluding that hydrogen combustion associated with the operation of the PHMS will not pose a
- threat to the integrity of the Sequoyah containment.
]-
9 Survivability of fsential Equipnient 4
By letters dated June 2, 1981, December 1, 1981, and November 29, 1982, TVA submitted an evaluation of survivability of the essential equipment exposed to the thermal env*coiimg t postulated in the containment during hydrogen burns
'~
initiated by tne PHMS.
Although' this' system was designed to prevent high hydrogen concentration buildup by, deliberate ignition of relatively low concen-trations of hydrogen in hydrogen-air. steam mixtures, the resulting release of thernal energy may still be sufficient to signficantly increase the temperature of the equipment located in the conta)nment.
Because some of this equipment is needed to ensure maintenance of a a safe shutdown condition and containment l
integrity, TVA was required to demonstrate that the essential equipment located inside the containment will survive the hydrogen burn environment resulting from operation of the PHMS.
TVA determined analytically and experimentally the thermal response of selected pieces of essential equipment exposed to a
(
hydrogen burn environment.
Cdeparing the resulting temperatures with the
(
qvalification1 temperatures for this equipment, TVA provided information to l
demontrate the survivability of the equipment.
! is ps i
s y
L_
t11/27/82.
s
' 22-37 SEQUOYAH SSER6 SEC 22 g '.
?
jf
'J' s
_ };l
/
9.1 Essential Equipment The selection of equipment that.mue'csurvive a hydrogen burn was based on its function during and after an accident.
In general, all the equipment in the following four categories of systems located in the containment was considered to be essential for safety of the plant:
(1) systems mitigating the consequences of the accident (2) systems needed for maintaining integrity of the containment pressure boundary (3) systems needed for maintaining the core in a safe condition (4) systems needed for monitoring the course of the accident TVA's selection of safety-related equipment was based on the shutdown and safety function diagrams (letter from R. T. Cross, TVA, to R. L. Tedesco, December 15, 1980).
The list of safety-related equipment is in Table 22.3.
TVA restricted the survivability evaluation to the equipment which is most sensitive to temperature change.
This reduced considerably the number of thermal response analyses and/or experiments that had to be performed.
The following equipment items were selected for an evaluaton of their thermal response to the hydrogen burn environment:
(1) igniter assembly (2) Barton transmitter (3) igniter power cable in conduit (4) thermocouple cable (5) resistance temperature detector (RTO) cable The staff has compared TVA's list of equipment selected for survivability evaluation with the lists of essential equipment prepared independently by the staff, and finds that the TVA list contains the equipment essential for safe operation of the plant under accident condition.
The staff has also reviewed the criteria used by TVA in selecting the equipment for analytical and experi-mental investigations.
Determination of the survivability of these pieces of 11/27/82 22-38 SEQUOYAH SSER6 SEC 22
Table 22.3 Essential equipment 1.
Mitigating Systems 1.1 Hydrogen igniters 1.2 Air return fan 1.3 Associated power and control cables 1.4. Hydrogen recombiner 2.
Systems Maintaining Containment Pressure Boundary 2.1 Air locks and equipment hatches 2.2 Containment isolation valves including hydrogen sample valves 2.3 Electrical penetrations 2.4 Gaskets and seals for flanges 2.5 Electrical boxes 3.
Systems Maintaining Core Safety 3.1 Reactor vessel vent valves (PORV) 4.
Monitoring Systems 4.1 Steam generator, pressurizer and sump water level transmitters 4.2 Core exit thermocouples 4.3 Reactor coolant system pressure transmitter 4.4 Hot leg RTD 4.5 Cold leg RTD 4.6 Reactor vessel level system 4.7 Associated cables (in conduits and expcsed) 4.8 Junction boxes 4.9 Operators on solenoid valves 4.10 Hydrogen analyzer equipment will be sufficient for establishing survivability of all the equip-ment listed in Table 22.3, provided these pieces of equipment have been included in the TVA equipment qualification (EQ) program.
For pieces of equipment that are not in the EQ program, TVA has provided separate bases for the survivability finding.
l l
9.2 Thermal Environment Response Analysis i
The thermal environment for evaluating equipment survivability was determined by the CLASIX computer code.
It corresponded to energy release from burning hydrogen which was generated during the accident resulting from a small-break l
l 11/27/82 22-39 SEQUOYAH SSER6 SEC 22
l LOCA with a loss of emergency core coolant injection (5 0 sequence), but with 2
both trains of sprays and air return fans operating. The hydrogen was assumed to be ignited by the PHMS when it reached 8 volume percent concentration, with each burn being 85% complete.
It was further assumed that the flame propagated throughout the containment with a velocity of 1 fps and its temperature remained constant at the adiabatic flame temperature of 1400*F. The CLASIX code predicted 6 burns in the lower compartment and 26 burns in the upper plenum of the ice condenser for this scenario.
No burns were predicted in the upper compartment.
The average time between the burns in the lower compartment is about 200 seconds, and the highest temperature reached by the gas is 884 F.
In the upper plenum, the average time between burns is about.. seconds and the highest temperature reached by the gas is 1114*F.
In addition, for the analysis to demonstrate thermal stability of the ice condenser foam insulation, the licensee has referenced the Duke Power Company's analysis (Parker,1981) in which it was assumed that hydrogen was burning continuously for 45 minutes at the midpoint of the ice condenser baskets; the resulting flame was conserva-tively assumed to be 1-in. thick with a temperature of 1600 F.
The thermal responses of the igniter assembly, Barton transmitter, and igniter power cable in conduit were analytically predicted for the thermal environment described in the previous section.
The igniter assembly was analyzed using the upper plenum temperature profile that is considered to be the most severe thermal environment for igniters.
It should be noted that the TVA analysis was l
done for the ign ter assembly used for the IDIS.
TVA has now decided to use a i
i different igniter assembly for the PHMS, one that does not employ a transformer.
Because the transformer was the most sensitive component of the previous igniter assembly, the staff concludes the same analysis could be applied to the new Igniter assembly.
The Barton transmitter was analyzed using the lower compartment tecperature profile, and the igniter power cable in l
conduit was analyzed for both the upper plenum and the lower compartment temperature profiles.
The staff has reviewed and concurs with this choice of thermal profiles for analysis, because these profiles conservatively represent I
the thermal environments to which the given equipment would be exposed during an accident.
11/27/82 22-40 SEQUOYAH SSER6 SEC 22 t
l
The analytical models used in predicting thermal responses of equipment considered thermal energy transfer from the moving flame by radiation and from the hot gases by natural convection only.
Standard heat transfer equations were used to calculate this heat transfer.
The heat transfer inside the equipment was determined by TVA using the HEATING 5 computer code (ORNL). This code was applied to solve heat transfer equations for two-dimensional models of different components.
Therefore, these components had to be represented by relatively simple geometries.
TVA prepared such simplified models which, despite their simplicity, included significant heat transfer characteristics.
The models used in the analysis were verified by comparing calculated results with the results derived from other accepted computer programs or obtained experimentally. The validity of the pressure transmitter model was determined by comparing its response to the results of the temperature transient analysis performed for equipment qualification using the C0C0 computer program (WCAP-8936).
This program was previously verified by the staff.
The agreement between temperature responses predicted by these two programs is satisfactory.
TVA verified the model for thermal response of thermocouple cable by comparing it to the results of the test performed by Fenwal (Fenwal, 1980).
Analytical results predicted the melting of the teflon insulation that was observed in the experiments.
The staff has reviewed the methodology used by TVA and finds that in general the models conservatively overestimate heat transfer from the flame because it is assumed to move in the containment with an artificially slow velocity and at an adiabatic temperature, despite its loss of energy to different heat sinks. On the other hand, the transfer of heat by radiation from the hot gases was neglected by the licensee.
The staff's consultant, Sandia, performed independent verification of TVA's analyses (McCulloch, 1982) and concluded that although they do not reflect true mechanisms of energy transfer for the hydrogen burn environments used, they y! eld conservative results.
Thermal responses for the thermocoupole and RTD cables were determined experimentally at TVA's Singleton Laboratory.
The cables were exposed to the simulated hydrogen burn environment in a Lindberg Tube furnace, and the temperatures reached by cable insulation were measured. The cables were 11/27/82 22-41 SEQUOYAH SSER6 SEC 22
exposed to 1400'F for five 30-second cycles.
Between the cycles (170-second period), the temperature was reduced to 300'F.
The staff concluded that this environment conservatively represents the condition existing in the lower compartment during hydrogen burn.
Thermal response of the igniter cable used in the IDIS was determined experi-mentally at Singleton Laboratory. The cable was placed in a conduit with both ends sealed. The cable in the conduit was placed in a Blue M oven and was exposed to about 700*F for about 45 minutes.
The staff concluded the environment conservatively represents the condition existing in the lower compartment or in the upper plenum of the ice condenser.
The IDIS cable was not part of the NUREG-0588 qualification program, although the cable used for PHMS is qualified to meet NUREG-0588 requirements.
Also, the materials used in the construction of the IDIS cable are more sensitive to heat than the materials used in the PHMS cable.
The acceptance criterion used for evaluating survivability of essential equipment is based on the qualification temperature of the equipment and the duration for which the temperature is maintained.
The equipment located in the containment will survive the hydrogen burn if the temperature reached by its most sensitive component will not exceed the temperature reached by this component during qualification tests.
Because the actual temperature reached by the tested equipment during these tests was not measured and qualification temperature was the temperature of thermal environment to which the test equip-ment was exposed, there is no direct way to determine the actual qualification temperature reached by the limiting components.
However, TVA claims that i
environmental qualification tests are typically conducted for extended periods of time and the equilibrium surface temperature should achieve thermal equilib-rium with the test chamber during the tests.
Because of several conservative assumptions in the thermal response analysis, the staff is of the opinion that use of the qualification temperature by TVA as a criterion for evaluating the survivability of limiting components is acceptable. To confirm equipment surviv-ability at elevated temperatures, TVA has performed tests in Singleton Laboratory in which the igniter power cable in conduit was exposed to 700 F for 45 minutes.
Although some degradation of the insulation was observed, the cable qualified in t
the subsequent high voltage test.
11/27/82 22-42 SEQUOYAH SSER6 SEC 22
v The analytically calculated thermal responses during hdyrogen burn are compared with the qualification temperatures in Table 22.4.
In all cases, the qualification temperatures are not exceeded.
It is the opinion of the staff
[
that this equipment will survive a hydrogen burn.
The survivability of thermocouple and RTD cables was determined experimentally l
by actually verifying their behavior in a simulated hydrogen burn environment.
The temperatures reached by the cable insulation are listed in Table 22.4.
Only slight degradation of cable insulation was observed.
Both cables successfully passed high voltage tests.
Table 22.4 Comparison of analytically calculated thermal responses during hydrogen burn and qualification temperatures Maximum temp, *F Component (calculated),
Design / test temp, *F Igniter (used in IDIS) i Interior box air 227 428 (transformer)
Cable 171 Transformer core 157 Bartol transmitter 1
Interior air 231 310 Case surface 245 1
Cable in conduit (used in IDIS) i Copper 251 tested to 700 Insulation 260 Conduit surface 332 j
Thermocouple cable insulation 1126 RTD cable insulation 1013 l
{
11/27/82 22-43 SEQUOYAH SSER6 SEC 22
4 All equipment except the core exit thermocouple, reactor vessel level thermo-couple, and vessel vent valves has been included in the TVA E0 program.
The j
core exit thermocouples are located inside the vessel head t ;a are not exposed to the hydrogen burn environment.
The reactor vessel level thermocouple and l
vent valves will be included in the EQ program when they are added to the plant.
i In a submittal dated November 29, 1982, TVA stated that all the equipment listed in Table 22.3--except for thermocouple and P.TD cable--reaches the equilibrium temperature.
Based on this statement and the experimental verification of RTD and thermocouple cables, the staff concludes that all the j
equipment listed in Table 22.3 will survive the hydrogen burn environment.
t It should be noted, however, that the tests conducted by the licensee were performed in a relatively small oven.
In NUREG-CR/2730, the staff's contractor (Sandia) has stated that on the basis of some preliminary test results, scaling (volume of containment building vs. volume of the test chamber) may be a 1
significant factor in analyzing the survivability of the equipment.
During i
fiscal year 1983, Sandia will be performing some additional confirmatory tests I
to address this concern.
But, based on the conservative assumptions and available margins in the work done to date, the staff finds that the essential I
equipment will survive the hydrogen burn environment.
The results from l
Sandia's upcoming tests will be relied on to confirm the findings made above.
I Gecondary fires in the Sequoyah plant may originate either when combustible l
materials located in the containment reach their ignition temperature or when i
the insulation on the ice condenser cooling ducts is heated to the point at which polyurethane foam starts to decompose and emit combustible gases.
After reviewing different possible sources of combustible materials, TVA identified organic cable insulation as the only significant suurce.
In most cases, how-ever, cables are completely enclosed in conduits or cable trays, and are not directly exposed to the hydrogen burn.
Those cables that'have e posed insula-tion have been tested to ensure their flame resistance.
In evaluating the thermal stability of insulation at ice condenser cooling ducts, TVA referenced 11/27/82 22-44 SEQUOYAH SSER6 SEC 22
?
the analysis performed by Duke Power Company for the McGuire plant (Parker,1981).
Because the ice condenser designs are similar in both plants, the analysis performed for McGuire is applicable to Sequoyah.
This analysis indicates that the polyurethane foam will not reach temperatures at which pyrolysis could generate combustible gases.
The staff has reviewed this analysis and concurs with TVA's conclusion.
9.3 Pressure Effects For the pressure profile inside the containment during the hydrogen burn, the conservative pressure profile was obtained from the CLASIX analysis with a 12 fps flame speed.
This analysis is identified in TVA's submittal of December 1, 1981.
With the PMHS, the highest predicted pressure in the containment does not exceed the pressures used during the qualification testing of equipment.
However, a pressure differential could be developed between the lower and upper compartments of the containment that could strain the blades of the air return fans.
TVA has indicated that the fans are protected by backdraft dampers; hence this pressure differential would not affect their performance.
In addition, TVA performed a structural analysis that indicates that the fans could take static loads in excess of those produced by the predicted pressure differential.
9.4 Staff Conclusions Regarding Equipment Survivability After reviewing TVA's analysis and/or experimental investigation of equipment survivability, the staff concludes that TVA has provided sufficient evidence that all the equipment required to ensure safe shutdown conditions and containment integrity will survive the environment created by burn of the hydrogen generated during a postulated accident. This conclusion is based on the following:
(1) The list of equipment provided in the submittal included all the essential ecuipment.
11/27/82 22-45 SEQUOYAH SSER6 SEC 22
(2) The equipment selected for the analytical and experimental investigations adequately characterizes the essential equipment on the list.
(3) The analytical methods used by the applicant adequately calculate thermal response of equipment, based on the postulated thermal environment.
(4) The comparison of analytically determined thermal responses to the corresponding qualification temperatures for some sample components has indicated that these temperatures will not be exceeded during a hydrogen burn.
(5) Experimental determination of survivability of the thermocouples, RTD cables, and igniter cable in conduit in the test chambers conservatively predicts their behavior in a hydrogen burn environment.
1 (6) It was satisfactorily demonstrated that burning hydrogen will not initiate secondary fires in the containment by igniting combustible materials by generating combustible gases from the decomposition of polyurethanc foam insulation.
10 Overall Conclusions The operating licenses for Sequoyah Units 1 and 2 contain a condition requiring that, " prior to startup following the first refueling outage, the Commission must confirm that an adequate hydrogen control system for the plant is i
installed and will perform its intended function in a manner that provides adequate safety margins." The licenses include another condition dealing with the TVA research program which provides, among other things, that "..TVA shall... evaluate and resolve any anomalous results occurring during the course of its ongoing test program."
The staff has concluded its review of the matte of hydrogen control for postulated degraded core accidents at the Sequoyah phnt and finds, subject to satisfactory resolution of the Tayco igniter temperature response during operstion of the containment spray system, that l
1 l
11/27/82 22-46 SEQUOYAH SSER6 SEC 22 l
l l
l
c The peak pressures as a result of igniter-induced burns will be less than
?
the containment pressure capacity.
The results of many accident analyses suggest that the peak containment atmosphere pressure will be close to the design pressure of 12 psig.
Even considering a broad range of accident scenarios and combustion assumptions that is more conservative, it is expected that the containment pressure will remain below 30 psig.
With adequate margins, the containment pressure capacity is 36 psig.
The essential equipment has been identified and the peak temperatures during a hydrogen burn for the most sensitive piece of equipment have been shown to be less than its qualification temperature.
The contingency identified in the above findings deals with a design feature of the PHMS.
Specifically, it concerns the capability of the Tayco igniter to maintain a surface temperature sufficient to initiate combustion in a spray environment.
Recent tests conducted by TVA indicate that the igniters will function as intended.
However, the temperature margin provided by the igniters appears to be small under spray condition.
The staff will require that TVA complete certain additional tests
'to verify that the Tayco igniter will maintain an adequate surface temperature in a spray environment such as that expected in the upper compartment of the ice condenser containment.
This work can be performed at the Nevada Test Site in early 1983, as part of the EPRI/NRC hydrogen research program.
As part of its PHMS evaluation, the staff al'so identified a number of technical l
concerns that it intends to investigate further as confirmatory items.
These will be pursued within the context of the staff's Unresolved Safety Issue Program, Task Action Plan Item A-48.
The confirmatory items are 1
local detonations CLASIX/ COMPARE code work equipment survivability for a spectrum of accidents combustion effects at large scale combustion phenomena including flame acceleration in the upper ice bed 11/29/82 22-47 SEQUOYAH SSER6 SEC 22
The subject of local detonations in confined regions of the containment is
?
currently under investigation at Sandia under a staff technical assistance contract.
This work is considered confirmatory in nature because:
(1) mixing of the containment atmosphere, in conjunction with igniter operation at low hydrogen concentrations, will preclude the formation of detonable mixtures, and (2) recent analyses performed by Sandia using the CSQ code and a refined structural analysis indicate that the Sequoyah containment can withstand the postulated detonation of a 20 volume percent hydrogen mixture in the upper plenum of the ice condenser.
The Sandia investigation should be completed by mid-1983.
The staff will continue to assess the adequacy of the CLASIX code as part of its technical assistance program with the Los Alamos National Laboratory. This containment code work is considered to be confirmatory in light of the staff's findings regarding the adequacy of the CLASIX models and the reasonable agree-ment obtained between CLASIX and the hydrogen burn version of COMPARE.
The code work will be an ongoing effort.
The staff will also continue to investigate equipment survivability for a spectrum of degraded core accidents.
This investigation will be carried out as part of the NRC Hydrogen Burn Survival Program already in place at Sandia.
The results of the hydrogen release rate sensitivity analyses and the substantial margins between predicted and qualification temperatures for the more temperature-sensitive pieces of equipment provide the bases for classifying this item as confirmatory.
The staff will monitor the results of other ongoing NRC and EPRI hydrogen research programs to:
(1) confirm the adequacy of the number and location of igniters in the upper compartment of containment; and (2) confirm the lack of significant flame acceleration at large scale.
Research programs to address these concerns will be performed at the Nevada Test Site and the Sandia FLAME facility, respectively.
These programs are considered confirmatory because similar test programs have been completed at smaller scale with acceptable results.
l Accordingly, subject to satisfactory resolution of the open item dealing with the Tayco igniter surface temperature, the staff finds the license conditions l
11/29/82 22-48 SEQUOYAH SSER6 SEC 22 l
[
dealing with hydrogen control during postulated degraded core accidents to be
~
l satisfactorily resolved.
l i
1 a
1 I.
I i
i 4
t es s
1 11/29/82 22-49 SEQUOYAH SSER6 SEC 22
APPENDIX B BIBLIOGRAPHY American Society of Mechanical Engineers, " Boiler and Pressure Vessel Code,"
Section VI, Division 1.
Berman, M., Sandia, letter to J. T. Larkins, NRC, enclosing April and May 1982 status reports for the Sandia Hydrogen Programs, July 20, 1982.
Cross, J.
L., TVA, letter to R. L. Tedesco, "Sequoyah Nuclear Plant Hydrogen Study, Volume II, Revision in Response to NRC Questions," December 11, 1980.
Fenwal Laboratory, "Sequoyah Nuclear Plant Core Degradation Program, Volume 2, Report on Safety Evaluation of the Interim Distributed Ignition System,"
December 15, 1980.
Kammer, D. S., letter to E. Adensam, NRC, " Seventh Quarterly Research Report,"
July 28, 1982.
l McCulloch, W. H., Sandia, letter to K. Parczewski, NRC, dated October 29, 1982.
Mills, L. M., TVA, letter to A. Schwencer, "Second Quarterly Research Report,"
March 16, 1981.
--, letter to E. Adensam, NRC, "Fifth Quarterly Research Report," January 22, 1982a.
--, letter to E. Adensam, NRC, " Sixth Quarterly Research Report, " April 23, 1982b.
l
--, letter to E. Adensam, NRC, " Summary of Testing to Determine Suitability of Tayco Igniter for Use in the Permanent Hydrogen Mitigation System at Sequoyah and Watts Bar Nuclear Plants," June 14, 1982c.
11/26/82 B-1 SEQUOYAH SSER6 APP B l
o 1,d*
Oak Ridge National Laboratory (ORNL), " HEATING 5, an IBM 360 Heat Conduction Program," ORNL/CSO/TM-15.
Parker, W. O., Duke Power Co., letter to H. Denton, NRC, dated December 31, 1981, transmitting " Analysis of Hydrogen Control Measures at McGuire Nuclear Station,"
Vol. II, dated October 1981.
Roller, S. F, and S. M. Falacy, Sandia, " Medium-Scale Tests of H : Air: Steam 2
Systems," paper prepared for NRC for submittal to the Second International Workshop on the Impact of Hydrogen on Water Reactor Safety," Albuquerque, October 3-7, 1982.
U.S. Nuclear Regulatory Commission, " Notice of Interjm Requirements Related to Hydrogen Control," 46 FR 62281.
--, NUREG-75/014, " Reactor Safety Study, " October 1975.
--, NUREG-0588, " Interim Staff Position on Environmental Qualification of Safety-Related Electrical Equipment," December 1979.
--, NUREG/CR-1891, " Reliability Analysis of Containment Strength--Sequoyah and McGuire Ice Condenser Containment," Ames Laboratories, August 1982.
--, NUREG/CR-2486, " Final Results of the Hydrogen Igniter Experimental Program," Lawrence Livermore National Laboratory, February 1982.
--, NUREG/CR-2730, " Hydrogen Burn Surv: val Thermal Model and Test Results,"
l Sandia, August 1982.
l l
Westinghouse Electric Corporation, " Zion Probabilistic Safety Study,"
Module 4, Section 4, 1981.
--, WCAP-8077, " Ice Condenser Containment Pressure Transient Analysis Methods" (Proprietary Class 2), March 1973; WCAP-8078 (Proprietary Class 3), March 1973.
l 11/26/82 B-2 SEQUOYAH SSER6 APP B
o,.
--,WCAP-8326 (Proprietary Class 3), F. M. Bordelon and E. T. Murphy,
" Containment Pressure Analysis (C0CO)," July 1974; WCAP-8327 (Proprietary Class 2), July 1974.
o ddtI.
--- WCAP-8936, " Environmental Qualification Instrument Transmitter Temperature Transient Analysis."
Westinghouse Offshore Power Systems, OPS-36A31; "The CLASIX Computer Program U
TdA for the Analysis of Reactor Plant Containment Response to Hydrogen Release and-Deflagration" (non proprietary; OPS-0735, proprietary).
Yang, J. W., and W. T. Pratt, "A Study of Hydrogen Combustion During Degraded Core Accidents in PWR in Condenser Plant." Brookhaven National Laboratory, Department of Nuclear Energy, prepared for NRC under Interagency Agreement DE-AC02-76CH00016, January 1982.
tidt 7p+W SECY-82-11
- jett 11/27/82 B-3 SEQUOYAH SSER6 Af? B