ML20002C853
| ML20002C853 | |
| Person / Time | |
|---|---|
| Site: | Sequoyah  |
| Issue date: | 12/24/1980 |
| From: | Tedesco R Office of Nuclear Reactor Regulation |
| To: | Murley T, Ross D, Vollmer R Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 8101120176 | |
| Download: ML20002C853 (45) | |
Text
_ __
[
REBUBIOT! D]C.k-i tL CM g'I
F *'
."F DISTFilBUTION:
Docket File (50-3 7/328)
LB d2/ File DEC 2 31980 M. Service C. Stable MEMORANDUM FOR:
R. Vollmer. Director, Division of Engineering D. Ross, Director, Division of Systems Integration T. Murley, Director, Division of Safety Technology M. Ernst. Assistant Director for Technology, Division of Safety Technology J. Knight, Assistant Director for Components & Structures Engineering Division of Engineering V. Noonan, Assistant Director for Materials & Qualifications Engineering, Division of Engineering L. Rubenstein, Assistant Director for Core and Containraent Systems, Division of Systems Integration F. Schauer, Chief. Structural Engineering Branch, Division of Engineering Z. Rosztoczy, Chief. Equipment Oualification Branch, Division of Engineering T. Spets, Chief, Reactor Systems Branch, Division of Systems Integration W. Butler, Chief, Containment System: Branch, Division of Systems Integration A. Thadani, Acting Chief, Reliability & Risk Assessment Branch, Division of Safety Technology FROM:
R. Tedesco, Assistant Director for Licensing, Division of Licensing
SUBJECT:
SEQUOYAH SUPPLEMENT ND. 4 - HYDROGEN CONTROL l
Enclosed is a draft of the SER, Supplement No. 4 that incorporates all staff inputs to date on hydrogen control. Your coments are needed C.O.B.,
December 29, 1980. This brief review period is necessary in order for us to mail the draft SER to ACRS members on December 30, 1980.
The ACRS Subcomittee meeting is scheduled for January 6, 1981 (1-5:00 PM).
We expect each contributing branch (CSB, EQB, RRAB, SEB) to formally present their findings (use vu-graphs) to the ACRS. Approximately 30 minutes is allocated for each branch including ti e for questions for ACRS members.
The presentation should include appropriate references to the McGuire plant to ensure an optimum generic assessment of the hydrogen control systent. design.
(
. es e
8
..... a n i
f O**T) r
+7 u_;
C F?iCI A '..:IC O R D C O PY c: ;m;,
z e m @%
D;
. DEC 2 31980 With respect to the open issues on the survivability of equipment. TVA will provide further infor: nation on December 24, 1980. We expect TVA to present this material at the ACRS and it is hoped the staff can coment on this submittal prior to the meeting. Most likely an additional supplement will be necessary to cover this issue.
~Oriefmf etmarf hv Robert L. Tedesco Assistant Director for Licensing Division of Licensing
Enclosure:
Draft SER, Supplement No. 4 i
l t
2,. co. LB #2/DL
! LB #2/0L lA DL l
I s mo CSEaf%/if ^AEch2n' er j ll
' do" l '
'j c
12/$3/80" l T$/
780' [ 12/ 9 ) 780
~,'
.n
.= : - : w ; 3 :
2.
a : s,. ; -
C ??!CIAL R E CO R O C O ?'t ac w-:-
M u c c G-Ocl l t
SUPPLEMENT NO. 4 i
l TO THE SAFETY EVALUATION REPORT BY THE j
OFFICE OF NUCLEAR REACTOR REGULATION l
U. S. NUCLEAR REGULATORY COMMISSION i
IN THE MATTER OF TENNESSEE VALLEY AUTh0RITY SEQUOYAH NUCLEAR PLANT, UNITS 1 AND 2 l
DOCKET NOS. 50-327 ANO 50-328 i
l i
i 1
s 1
TABLE OF CONTENTS 5
- 1. 0 Introduction and General Discussion................
1-1 1.1 Introduction............................................
22.0 TMI-2 Requirements......................
22-2-1 II. 8. 7 Analsys of Hydrogen Central.........................
22.2-1 Position.............................................
Discussion and Conclusions...........................
System Description.................................
Testing of the 10IS................................
Analysis 10I5......................................
Probability of Core Damage Events..................
Hydrogen Generation and Containment Pressures......
Containment Pressures..............................
Surviality of Essential Equipment..................
Conclusions........................................
TABLES............................................................
22.2-1 Core Melt Probabilities..............................
22.2-2 Containment Analysis Sensitivity Studies.............
22.2-3 Mean and Standard Deviation of Material Properties...
9'
"*a+=4n= ant Ana'veie <=aci'4u4*;
Itadier.............
FIGURES 22.2-1 Estimated Time Sequences for Potential Core Damage Events...............................................
i 22.2,)JiEnclosureofFoamInsulation..........................
l 2 1. 1 - 2 /g b L -d :m. - A Ds.-l u. a,t t.4.1 6 3 t
V l
s 1
- 1. 0 INTRODUCTION AND GENERAL DISCUSSION 1.1 Introduction On September 17, 1980, the Nuclear Regulatory Commission (NRC) issued the facility operating license OPR-77 to the facilit) :; ~ + "1 *ea"^ 2 M to the Tennessee Valley f,uthority, for the Sequoyah Nuclear Plant Unit No. 1, located in Hamilton County, Tennessee.
The license authorizes operation of
.4 Unit No. I at 100 percent power, subject 4w a number of license conditions.
The purpose of Supplement No. 4 to the SER is to further update our Safety Evaluation Reports on the hydrogen control measures (Section 22.2, II.B.7);
and to comply with license conditions 22D(1) and (3) which are as follows:
D.
Hydrogen Control Measures (Section 22.2, II.B.7)
(1 By January 31, 1981, TVA shall be testing and analysis show to the satisfaction of the NRC staff that an interim hydrogen control system will provide with reasonable assurance protection against breach of containment in the event that a substantial qcantity of hydrogen is generated.
42.s I
During the interim period of operation, TVA shall continue a research program on hydrogen control measures and the effects of hydrogen burns on safety functions and shall submit to the NRC quarterly reports on the research program.
- N4,
W cmhb.,(N M O-F= *-
d'
- Ig 1-1 R
s t
In compliance with license condition 220 (3), above, TVA submitted on December 1, 1980, the first quarterly report on the research program for hydrogen control.
Also, TVA revised volume 2 of the Sequoyah Degradation Program Report to
. t
- i. s,. 2. s 1 x #w t " ".
incorporate accitional information on the overall program.
k '-----' "-
q' -
_l
,, M L C a-da=1s, solely wHh license condition,22D(1).
Each section is supplementary to and not in lieu of discussion in the Safety Evaluation Report and Supplements Nos.1, 2, and 3, except where specifically noted.
1-2
s s
22.0 TIM-2 REQUIREMENTS II.B.7 Analysis of Hydrogen Control Position Reach a dfcision on the immediate requirements, if any, for hydrogen control in small containments, and apply, as appropriate, to new OLs pending ccmple-tion of the degraded core rulemaking in II.B.B of the Action Plan.
Discussion and Conclusions it3 Supplement $No & to the Safety Evaluation Report provided a basis of concluding that the full power licensing of Sequoyah Unit No.1 need not afait completion on ongoing work on hydrogen control measures.
This supplement concludes that, 3. r u o,.o.t d v subject to the satisfactory resolution of equipment suggaetttty in containment during the postulated hydrogen burns, operation of the IDIS for an interim period of 1 year is appropriate.
System Oescriotion i
The Tennessee Valley Authority (TVA) has installed within S.he Sequoyah, Unit 1 containment a system of igniters and ancillary equipment designated as the interim distributed ignition system (IDIS).
The igniters are designed to insure a controlled burning of hydrogen in the unlikely event that excessive quantities of hydrogen, well beyond the design bases required by 10 CFR i
22-1 l
L
s Section 50.44, are generated as a result of a postulated severly degraded core accident.
The igniter selected by TVA is a glow plug commonly used in diesel engines that is manufactured as Model 7G by General Motors AC Division.
The igniter is powered directly from a 120/14V ac transformer.
The igniter assembly consists of a steel box with 1/8-inch thick walls, which houses the trans-former and all electrical connections and partially encloses the igniter.
The glow plug side of the box is covered by a spray shield; and the glow plug face of the igniter assembly has a copper heat 3 ink.
The igniters are powered from the standby lighting system which has normal and alternate ac power supply from offsite sources.
In the event of a loss of offsite power, the igniters would be powered from the diesel generators.
The IDIS was not designed as a safety grade system and as such is not a seismic Category I system; but it is seismically designed not to damage other safety-related equipment inside containment. This is accomplished by security the igniter assemblies with a steel cable attached to an anchoring bolt.
The interim distributed ignition system presently installed in the Sequoyah Nuclear Plant, Unit 1 consists of 32 igniter assemblies distributed throughout the ucper, lower, and ice condenser compartments. There are a total of 20 igniters in the lower compartment, and 3 igniters, which are suspended 35 feet from the top of the containment in the upper compartment. There are a total of 9 igniters located in the ice condenser compartment; 5 in the lower plenum and 4 in the upper plenum of the ice condenser. As a supplement to the 32 22-2
9 igniters presently installed, TVA has committed to install 13 additional igniters, all to be located in the upper compartment, before the first refuel-ing outage.
The interim distributed ignition system is designed to be actuated following the start of an accident and to remain actuated until the unit reaches cold shutdown. The system is to be manually initiated by switching on three light-ing circuits at the Standby lighting panel located in the auxiliary building.
To insure that the interim distributed ignition system will function as intended TVA has proposed a pre-operational and surveillance testing program.
Pre-operational testing, to be performed upon installation of the system, will verify:
- 1) that the output voltage of the transformer is greater than or equal to 12 volts and is less than or equal to 14 volts; and 2) temperature of the igniter is at least 1500*F.
During the pre-operational 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 surveillance testing wnich will consist of energizing the IDIS at the standby lighting panel and taking current readings of the circuits.
If the current readings do not compara fcvorably with current measurements taken during pre-operational testing, then all igniters will be visually and otherwise checked to instre their operability.
22-3
i Q
Testing of the IDIS TVA has 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 TVA igniter were performed by Fenwal, Inc. to study igniter performance under various environmental cor.ditions.
l The principal purpose of the testing at Singleton Laboratory was to evaluate igniter surface temperatures and determ he the effects of overvoltage condi-tions and extended operation.
Testing of a GM plug was conducted at 12, 14, and 16 volts ac with surface temperatures of 1480,1550, and 1650!F as measured by a thermocouple. Since thermoccuple heat losses were estimated to be signi-ficant the surface temperature of another plug at 14 volts was measured with an optical pyrometer obtaining at least 1800*F.
Subsequent testing on several GM glow plugs has been performed to insure adequate surface temperatures at the minimum design voltage. The minimum acceptance temperature of 1500*F was reached within 1 minute, with all plugs reaching 1600*F in 3 minutes.
Voltage tests on 5 plugs verified reliable operation at 14 volts but two plug failures were recorded after a minute's operation at 16 volts.
Initial endurance tests were conducted on a plug by operating it continuously at 14 volts for 148 hours0.00171 days <br />0.0411 hours <br />2.44709e-4 weeks <br />5.6314e-5 months <br />. After the successful completion of the test, the plug was then used in hydrogen burning tests.
22-4
After the GM glow p'.ug was selected, TVA began more extensive endursnce testing to select the igniters which would be installed inside containment.
From a lot of 302 plugs which were subjected to a screening test to eliminate those with manufacturing defects, 50 plugs were randomly sampled.
These plugs were then tested by cycling the plug and then by continuous testing for 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> to verify successful operation.
Based on the results of these tests, we find that the selected igniters should perform satisfactorily in terms of reliably achieving the desired surfact temperature of at least 1500*F for the required duration.
To evaluate the efficacy of the igniters, TVA in cooperation with Duke Power, American Electric Power and Washington developed a two phase testing program, which was conducted by Fenwal, Inc. The testing was conducted using a single igniter assembly in a spherical vessel approximately six feet in diamater (134 ft ).
Features of the test assembly include external electrical heaters, internal fan and water spray. The vessel atmosphere pressure and temperature conditions were measured, as was the vessel surface temperature.
Instrumentation was also added to measure the surface temperature of equipment samples located in the test vessel.
The hydrogen concentration was determined by taking samples for gas chromatograph analysis.
The phase 1 tests were conducted to determine the ignition efficancy of the igniter for varying hydrogen concentrations under various test vessel atmos-pnere conditions of pressure, temperature, steam concentration and turbulence.
22-5
t 4
Turbulence was simulated by the internal mixing fan to produce estimated flow velocities of 5 and 10 feet per second past the igniter.
The phase 2 portion of the test program was designed in four parts to investi-gate separate effects which may affect igniter performance and tne response of typical equipment in the Sequoyah plant.
The part 1 tests were performed to evaluate the igniter performance in the lower hydrogen concentration range.
The part 2 tests were conducted to determine igniter efficiency under tran-sient conditions whereby steam and hydrogen were introduced continuously into the test vessel. The part 3 tests were conducted to evaluate the effects of water spray on igniter performance. The part 4 tests were designed to provide information on the effects of a hydrogen burn on typical equipment, e.g.,
limit switch and solenoid valve, by measuring the equipment temperature response.
The results of the Fenwal test program were generally consistent with the applicable published information on hydrogen combustion.
The tests confirmed that limited combustion occurs over hydrogen concentrations of 6-B% although the completeness of combustion is influenced by the ability to promote mixing or thereby exposing the igniter to sources of fresh atmosphers.
In the regime of hydrogen concentrations of 8-9%, test results indicate the combustion process is altered; this range of concentrations represents a transition zone where comcustion may proceed to a nearly complete reaction.
Again, this is consistent with published data and findings regarding the general limits of upward and downward flame progagation. At the higher hydrogen concentrations of 10-21% that werc tested, the results indicate that the likely scenario is 22-5
S 4
that the combustion process will proceed to consume all the hydrogen present in the atmosphere.
The test vessel atmosphere pressure measurements reflected the completeness of combustion at the various hydrogen levels.
Prcssure measurements showed an increase of as little as approximately one psi for single hydrogen burr.s at 6%
concentra= tion and an increase of approximately 70 psi for the burning of a mixture of 12% concentration. The phase 2, part 2, portion of the test program dealing with transient hydrogen injection with the plug pre-energized resulted in a series of eight burns for the test with continuous injection of both steam and hydrogen.
The injection flow rates were scaled to simulate the flow rates calculated to enter the lower compartment of an ice condenser containment for a small break loss-of-coolant accident (LOCA).
The results of that test indicate relatively small pressure increases of approximately 2-7 psi for each of the eight discrete burns.
These lower pressure excursions result from the burning of hydrogen at lower concentrations and allowing for heat removal via heat transfer to heat sinks during the transient.
It is to be reasonably expected that this type of sequence simulates the transient that would occur if hydrogen were to be introduced inside Sequoyah in a like manner.
The effects of spray operation on igniter performance appear to be minimal in terns of affecting the ability of the igniter to initiate combustion. There is some evidence, however, that spray operation, by promoting mixing, and turbulence, may actually improve the efficiency of the IDIS in burning lean hydrogen mixtures.
Similarly, the effect of fan flow across the glow plug, within the considered flow rates, indicates minimal alteratidn of the ignition 22-7
time and some improvement in the ability to burn a larger fraction of the hydrogen in lean hydrogen mixtures.
Based on the results of these tests as described in the applicant's suomittal, we find that the glow plug igniter will serve its intended function to initiate I
comcustion flammable mixtures under various conditions.
To independently evaluate the efficacy of the ISIS, the NRC staff arranged for the testing of the GM glow plug igniter at Lawrence Livermore National Laboratory.
The test program was designed to examine the performance of the igniter under a spectrum of test conditions.
The principal parameters of concern in this testing were varying nydrogen and steam concentrations.
The Livermore tests were conducted in an insulated pressure vessel having a 3
volume of 10.6 ft. Unlike the vessel used in the Fenwal test program, this vessel was unheated, which allowed a faster steam condensation rate and heat transfer to the environs. Primary data for the Livermore tests included atmosphere pressure and temperature measurements and gas concentration analysis.
Gas analysis was accomplished by drawing samples from the vessel; samples were taken before and after each hydrogen burn.
Mass spectrometric analysis was the method by which volume concentrations were determined.
The test matrix included dry tests with hydrogen and air mixtures ranging from l
6% to 16% hydrogen, and steam tests with 30% and 40% steam fractions, each with varying hydrogen concentrations. A total of 43 tests was run.
22-8 l
The Livermore combustion tests confirmed the ability of the proposed TVA igniter to ignite gas mixtures over a range of conditions.
In the dry air tests we were able to partially burn the hydrogen at lower concentrations (7-6%) and completely burn the hydrogen at higher concentrations, similar to what was seen in the Fenwal test program.
For the tests with 30 and 40% steam concentrations, it appears that the flammability limit for downward flame pregagation was shifted upward to higher hydrogen concentrations.
The results of a test at 30% steam and 10% hydrogen indicated a partial burn.
- However, the igniter never failed to initiate combustion for any of the 30 and 40%
steam fraction tests.
The measured pressure increases for tests with hydrogen concentrations over the range of 8-12% varied from approximately one psi to 65 psi, showing close agreement to the meassured pressure increases during the Fenwal tests.
There was no detectable deterioration in the ability of the glow plug to initiate combustion through the test series using the same plug for all the tests.
It consistently initiated burning in dry mixtures at plug temperatures of 1310 to 1370*F for the dry tests and at 1360-1480*F for the steam tests.
As previously discussed for the Fenwal test program it was apparent that promotion of mixing or turbulence enhanced combustion at ower concentrations.
This phenomenon, while not specifically addressed in the Livermore tests, was apoarently manifested as secondary burns for tests when the circulating fan was activated after the initial burn and the igniter temoerature remained high enough to reinitiate combustion.
22-9 l
Several tests were run with a nominal steam fracticn of 50%. Although the degraded core accident scenario considered in the Sequoyah review would not result in such high steam concentrations, we performed the tests to determine what steam fractions would be necessary to prevent ignition of the mixture.
In all the tests with a 50% steam fraction we were unable to initiate combus-tion.
This steam concentration which was observed to effectively inert the vessel atmosphere also approximates the values quoted in the reference litera-ture.
Two tests which began at nominal steam concentrations of 50%, with no initial combustion, were allowed to continue with the steam fraction being gradually reduced by condensation on the vessel wall.
Even though the steam fraction was eventually reduced to levels where combustion should have occurred, no substantial pressure increase, as a result of hydrogen burning, was recorded.
We have been unable at this time to conclusively resolve why a pronnounced burn did not occur and plan to continue our investigation of this matter.
Because the initial steam concentration is outside the spectrum of conditions calculated for the Sequoyah plant during the accident used as the basis for review, we see no immediate cause to consider these particular test results as l
l a bases for rejection of the TVA ignition system as an interim solution to hydrogen control for degraded core accidents.
The staff is planning to continue the Livermore tests for several months to investigate the effects of containment spray operation on igniter performance i
and to further study hydrogen combustion in steam environments. We will j
report on the continuation of this testing in a future supplement to the Safety Evaluation Report.
22-10 2
Analyais of IDIS To evaluate the role of igniters in accident mitigation, TVA has initiated an analytical effort to determine the effectiveness of distributed ignition systems in reducing the threat to containment integrity due to the combustion of that hydrogen generated following a spectrum of postulated degraded core j
accidents.
It is expected that thorough analyses including sensitivity studies on critical parameters for a range of accident scenarios will continue for approximately one year.
Currently, TVA has prcvided the results of analyses for a single degraded core accident scenario, designated S20 in WASH-1400, which is a small break LOCA accompanied by the failure of emergency core cooling injection. The 520 sequency leads to the production of hydrogen from the zirconium-water reaction as a result of the degraded core conditions, i.e., lack of core cooling.
TVA has concluded as a result of studies to determine accident sequence pro-babilities that a small break LOCA followed by a failure of emergency core
[
cooling injection is one of the more probable sequence among the spectrum of serious degraded core accidents. TVA has cited calculations to indicate that the S20 sequence is a factor of 20 more likely than a large break LOCA.
Another sequence initially considered for the purposes of evaluating the IDIS was a loss of feedwater transient following by a loss of offsite power; however, TVA has concluded that the good availability history for offsite power near l
the Sequoyan plant lessens the likelihood of this accident.
22-11 l
probability of Core Damage Events The sequence of interest in evaluating a hydrogen control system were limited to those that result in only core damage with associated significant hydrogen generation from the zirc water reaction, as opposed to complete core melt events with associated pressure vessel failure.
In general, these sequences result from arrested core melt events where vital equipment / functions are restored after their initial failure of result from dynamic situations caused by man-machine interactions similar to the TMI-2 accident.
The frequency of these core damage sequences has not been estensively reviewed previously because of the difficulty in discriminating between core damage and core melt events.
TVA has presented estimated probabilities for core melt sequences based on an earlier study of Sequoyah performed by Sandia under the reactor Safety Study Methodology Applications Program sponsored by the Office of Nuclear Regulatory Research. A list of potential events where operator action may limit the l
event to only core damage by restoring failed equipment is presented in Table 1.
These events were selected because of the relatively long time to total core melt (1-1/2 to 2-1/2 hours) which could allow potential beneficial operator action.
In order to have core damage with s1gnificant hydrogen generation (and not l
total core melt), vital equipment would have to be restored in a short time period (15 to 45 minutes) prior to complete core malt as shown by the shaded areas in Figure 22.2-1.
Earlier recovery would result in small hydrogen 22-12 I
l
~
generation which would not deopardize containment integrity while later recovery would be too late to preclude complete core melt and pressure vessel failure.
The probabilities of core melt in Table 22.2-1 are estimates by Sandia and the licensee. Without vouching for the correctness of either set of estimates, the tocal probability of potentially rectifiable core melt sequences (i.e.,
limited to core damage) is approximately 50 percent of the probability of all core melt sequences.
The conditional probability for restoring a vital component / system function by unplanned operator action in the time frame noted above given a potentially rectifiable event has not been determined.
Such a determination would require a decomposition of the ccmponent/ system unavailabilities into their root causes and a probabilistic analysis of their rectification under a temporal fame of reference.
Mean repair times for failed pumps, valves, diesels, and instrumentation range from 6 to 21 hours2.430556e-4 days <br />0.00583 hours <br />3.472222e-5 weeks <br />7.9905e-6 months <br /> in WASH-1400.
On the other hand, opening valves locally or jury rigging electrical connections or bypasses may only take minutes to correct misaligned systems.
Considering the above repair times, a crude estimate of the conditional probability for restoring vital component / system function in the 15 to 45 minute window shown by the shaded area in Figure 1 is assumed to be 0.1 to 0.2.
Combining this assumed con-ditional probability with the probability of having a rectifiable event, we would estimate that the probability of having an event with core damage and significant hydrogen generation (i.e., an aborted core melt) is approximately 5 to 10 percent of the total core melt probability.
22-13
TABLE 22.2-1 Core Melt Probabilities
- SANDIA TVA
~
~
Small LOCA, loss of ECCS injection **
9 x 10 8 4 x 10 4 (S 0) 4 2
Small LOCA, loss of ECCS recirculation **
2 x 10 2 x 10 8
~
Small LOCA, loss of ECCS and spray **
3 x 10 7 recirc; no containment drain effect
($ HF) 2
~
~
Loss of all feedwater with AC power **
7 x 10 7 2 x 10 8 available (TML, TB, TLD) 2
~
~
Station Blackout **
3 x 10 7 8 x 10 3 (TML3 ' )
~
~
All other events 2.8 x 10 5 4 x 10 8 Total 5.8 x 10 s 1,2 x 10 s i
I
- Per reactor year.
- Potential core damage events with significant hydrogen generation.
l l
22-14
5 S
FIGURE 22.2-1 22-15
i 1
FIG. I P
l Esfimorto Tir4E SEQoEp4E5 FoA l'oftinflAL CohG, DAMME EVf HTS ll uan t sanoto Mus B7//////////,
sunucare cou. canus 514Af Ec4$
cont coat wsTtt slGAilr K gaer
"* #5
" * "I8V hy0AoG6As C Ev64nisou 4
SMALL LOCA, no EdC6 IU J E CTioD 3
e l
t s
6
$14AT E45 coa 4 coat.
stats vecouthi
VO/M/Al unu unt.
cokt uncnsa1 M6r LOS$ OF FE Eb U>ATRk, sTnnon V, tall < our IE 0
i J
t-3 o
/
I 3
q Til4E - Houlu 4
A second mode of achieving core damage with significant hydrogen generation is for operator interruption (error of commission) of vital furctions and subse-quent late restoration of the functions in a dynamic man-machine interaction situation such as the TMI-2 accident. We are not aware of any detailed event tree / fault tree analysis of such man-machine interactions, however.
Some insignt may be gained by considering some potential events qualitatively.
There are essentially two primary situations that may lead to core damage with significant hydrogen generation, namely, a hole in the primary system with insufficient coolant makeup or insufficient feedwater flow to remove reactor heat leading to dry out of the steam generators and subsequent boil off of the primary coolant. These events would have to be slow enough, like those in Table 1, for the operator to even contemplate potential inappropriate control (error of commission) of the vital functions of core cooling and/or feedwater flow.
Based on our review of small LOCA's and loss of feedwater events in Westinghouse plants (NUREG-0611), these vital functions could be interrupted for significant lengths of time (about 40 to 60 minutes) before operator corrective action would necessary to preclude significant hydrogen generation.
Thus, it is reasonable to consider operator corrections of earlier inappro-i priate operator termination of the vital functions. Operator recognition and maintenance of these vital functions has been stressed by the past TMI-2 l
actions.
i Emergancy operating procedures and operator training for small LOCA's now emphasize the need to maintain high pressure injection flow if subcooled conditions cannot be maintained in the primary system.
In addition, reactor 22-16
subcooling meters have been installed and a technical advisor added to the control room complement which would further enhance the fulfillment of the safety function of keeping the core covered.
Similarly, emphasis has been placed on the operation and control of the auxiliary feedwater system needed to provide the vital function of a heat sink under accident situations.
Although no formal quantification has been made of the human errors associated with these vital functions, the staff does not believe that human errors (uncorrected errors of commission) associated with controlli g primary systems represent major contributors to the unavailabilities of the core cooling and feedwater delivery functions on a best estimate basis because of the actions initiated since TMI-2.
Our qualitative review of potential human errors has been limited; it did not consider human control errors associated with support equipment like component cooling water, ventilation, etc., which could ulti-mately impact the vital safety functions and, in the absence of a systematic event tree / fault tree evaluation, the completeness of the sequences considered is uncertain.
The foregoing discussion of man-machine interaction and the necessity of maintaining certain vital functions assumes a complete rectification of the human-related deficiencies exhibited by the TMI-2 accident. However, on an upper-bound basis, some undefined event may occur because'of an unforeseen design deficiency and/or operator error of commission.
The severity of an upper-bound event cannot be predicted a priori because of the unknown circum-stances leading up to the event and the time window for an event to progress -
i from core damage with significant hydrogen generation to total core melt.
22-17 g
.-.,w-n3 s
The occurrence of the TMI-2 event does not mean that an upper-bound event will more likely result in substantial core damage rather than core melt because of the correction of TMI-2 deficiencies.
Since all core melt events are preceded by core damage with significant amounts of hydrogen generated, hydrogen control may have a significant impact on containment failure (and societal consequences) regardless of the relative frequency of core damage (only) events. We have not reviewed the potential benefit of hydrogen control under total core melt conditions.
Hydrogen Generation and Containment Pressures In order to perform analyses of the containment atmosphere pressure and tempera-ture response due to a loss-of-coolant accident, it is necessary that the releases from the reactor coolant primary system be known, including steam and hydrogen release rates.
The TVA containment analyses was based on the reactor coolant system response and releases using results from the MARCH computer ccde as provided by Battelle Columbus Laboratory.
The MARCH code was developed by Battellee Columbus Laboratory for the NRC in order to provide the capability for the analysis of the thermal-hydraulic re-sponse of the reactor core, primary coolant system and containment to core melt accidents. 'n:th regard to hydrogen evolution the code provides a method of incorporating the Baker-Just and Cathcart metal water reaction rate models with a history of the uncovering and overheating of the core to obtain hydrogen generation rates. The extent of production of hydrogen is dependent on the 22-18
degree of uncovering of the core and on the amount of water (steam) available for oxidation, but the reactor pressure at the time of release is a dominant factor in determining the amount of hydroga? rel ased from the primary system according to MARCH.
MARCH then models the release of hydrogen with the steam from whatever openings in the primary system may be appropriate or the scenario (PORV, small break, or large break).
The talease of hydrogen is assumed to accompany the steam release according to the mass average composition calculated for the steam and hydrogen.
Under small break LOCA conditions, MARCH predicts the generation and release of hydrogen prior to core melt to be spread out over a 30 to 100 minute period.
Average release rates of 8 to 30 pounds per minute cover these cases.
In some instances, a Lurst of hydrogen into containment when the core melts completes the release.
Another mode of hydrogen release is possible through the high point vents that are to be required on LWRs.
These are to be sized such that their discharge can be compensated for by the charging pumps.
It is estimated that vents of this size will release hydrogen at the rate of about 20 lb/ min.
These considerations have led the staff to use 20 lb/ min as a base value in scoping calculations of the release of hydrogen prior to core melt or vessel failure. This release rate is typical of small break LOCAs, up to 2" in diameter, high point venting, and TMLB accidents modified to have small break characteristics.
22-19
The hydrogen release rates emphasized in calculations by TVA are those taken from MARCH code interpretations of a small break LOCA.
These MARCH release rates are a time varying function whose average is of the order of 20 lb/ min.
The staff considers these rates to be representative of a significant group of releases that might be encountered in typical degraded core accidents short of total core melt or vessel failure, and are an acceptable upper limit basis for calculations.
Although there are opposing view points, the consensus interpretation of the TMI-2 accident is that from 30 to 60% of the core zirconium cladding was oxi-dized generating 500-1100 pounds of hydrogen.
The applicant, TVA, has submit-ted analyses that persue the course of the accident up to the time when approximately 80% of the core cladding has been oxidized, and has justified the termination of the analyses at that point because further oxidation would result in majnr melting or rearrangement of the core.
The staff notes that these analyses do not provide for any supplementary oxidation of ferritic materials.
The ferritic materials, however, are,
unlikely to be oxidized substantially until the final phases of core rearrangement or melting.
The quantity of hydrogen chosen by the applicant for his principal analyses is, therefore, acceptable.
TVA then, using the releases calculated from the MARCH code, calculated the containment atmosphere transient using the CLASIX code, which was developed by Westinghouse /0PS.
The CLASIX code is a multi-volume containment code whicn calculates the containment pressure and temperature response in the separate 22-20
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 out presently does not include a model for structural heat sinks.
Mass and energy released to the containment atmosphere in the form of steam, hydrogen and nitrogen is input to the code.
The burning of hydrogen is calcu-lated in the code with provisions to vary the conditions under which hydrogen is assumed to burn and conditions at which the burn will propagate to other compartments.
Tne conditions inside the containment prior to the onset of hydrogen geners-tion were determined from LOTIC analyses; LOTIC being the Westinghouse long-term ice condenser analysis code previously reviewed and approved by the staff. The CLASIX calculations then begin at the onset of hydrogen produc-l tion, which occurs at approximately 3500 seconds following onset of the 1
accident.
The base case CLASIX analysis utilized the assumption that hydrogen l
ignition within a compartment was initiated at a 10% hydrogen concentration l
and that burning is assumed to propagate the other compartments with a 10%
hydrogen concentration.
l l
The hydrogen release to the containment was terminated, for the containment analysis, after approximately 1550 lbs of hydrogen were released. This mass of hydrogen corresponds to the reaction of approximately 80% of the zirconium 22-21
cladding (including plenum in the core.
At this point in the scenario the core is dry, thus there is no steam to produce a further zirconium-steam reaction.
Extending the accident scenario to the point of reactor vessel melt through will be the suoject of future analysis in conjunction with TMI Action P1an Item II.B.8, i.e., the reulemaking proceeding.
The results of the CLASIX base case analysis indicate that the hydrogen will be ignited in a series of nine burns in the lower compartment with the last of the burns propagating upward into the ice condenser compartment.
The total interval over whicn the series of burns occurs is approximately 3300 seconds.
For the first burn, a peak pressure of 26.5 psi was calculated for the lower comaartment, and 28.5 psi for the ice condenser and upper compartment.
The pressure in the containment before the first burn was approximately 22.5 psia.
Subsequent burns resulted in successively lower pressure peaks.
Peak tempera-tures of 2200*F, 1200*F and 150*F were calculated in the lower compartment, ice condenser and upper compartment, respectively.
As a result of the action of engineered safety features, such as the ice condenser, air return fans, and upper compartment spray the pressure and temperature spikes were rapidly atenuated between burns.
The pressure was decreased to its preburn value roughly 2 minutes after the burn occurred.
After the last ignition of hydrogen, which occurs approximately 6800 seconds after onset of the accident, there were roughly 300,000 pounds of ice left in 6
the ice condenser section (representing at least 40 x 10 BTU's in remaining heat removal capacity).
22-22
In summary, the results of the TVA base case an.. lysis show only a modest increase in containment pressure, on the crder of 4-6 psi, with the containment remaining well below the estimated failure pressures.
The burning criterion used in the analysis caused virtually all of the burning to occur in the lower compartment, thereby gaining the advantage of heat removal by the ice bed.
It should also be noted that each burning cycle involved the comoustion of only 6
100 pounds of hydrogen, or roughly 6 x 10 Btu's of energy addition.
By burning at a given concentration in the lower compartment (where one might naturally assume hydrogen concentrations to be higher since this is the area of hydrogen release) there is also the advantage of burning less total hydrogen at a time because the lower compartment volume is only aroung 1/4 of the total cor.tainment volume, which allows for expansion of the hot gases to the rest of the containment free volume.
.i TVA has also performed preliminary sensitivity studies to determine the effects of ignition criteria and safeguards performance on the containment response.
Results of several of these studies are shown in Table 22.2,X 2 The sensitivity analysis performed to date demonstrate that 1) the ignition criterion, within the bounds chosen, has little effect on the containment pressure; 2) partial vs full operation of the air return fans makes little difference on the results; 3) ice condenser heat removal is effective in reducing pressure; and 4) without any fan operation to assure mixing, the containment pressures due to burning rise dramatically to the point where the containment can be expected to lose structural integrity.
It should be noted 22-23
fa TABLE 22.2i{}
CONTAINMENT ANALYSIS SENSITIVITY STUDIES Peak Press (Psia)
Lower Upper Comp.
Comp.
1.
Base Case 26.5 28.5 2.
H Ignition and y
Propagation 8%
28.5 30.5 3.
1 Air Fan 26.5 29.5 4
No Ice
- 41 41 5.
No Air Fans 46.4 92.4 l
- Ice exists only for the first two of seven burning cycles.
l t
22-24
that the case which considered only enougn ice exists to reduce th e pressure sptke for two burns (out of seven) is non-mechanistic; i.e., it is not representative of the acual S20 scenario.
However, it does importantly demon-strate that even without ice, the containment pressure, with the assumed igniter operation, remains below the estimated the estimated containme failure pressure.
This serves to indicate some insensitivity to whatever accident scenario is chosen.
The calculations performed to date by TVA are the first efforst to anal t y cally define to value of the distributed ignition system.
TVA plans to refine the analytical models in the CLASIX code perform additional parametric a evaluate other accident sequences, in assessing the effctiveness o ignition system.
These additional analyses will be discussed in a future supplement to the SER.
TVA has initiated efforts to verify the computer code CLASIZ which wa e
o perform the preliminary containment transient analysis of hydrogen distrib u on and deflagration.
CLASIX, which was developed by Offshore Power Systems (OPS)/ Westinghouse, has been described as a code under development Therefore, in order to increase confidence in the code's calculations
, OPS has begun efforts to verify the code by comparison with the results of other Westin containment codes, namely the TM0 and COCO codes.The CCCO code, which is the Westinghouse dry containment code, has been used for several years and most recently was used to perform containment pressure calculations with hydr burning in the Zion / Indian Point (Z/IP) studies.
22-25
4 A comparison of results has also been made for selected cases using the TM0 code. The TM0 code is the Westinghouse subcompartment and short-term transient ice condenser code, which has been reviewed and approved by the staff.
For both two phase and superheated mass and energy releases, the CLASIX and TM0 codes predict pressure transients in close agreement.
In summary, the efforts to date to verify CLASIX using familiar licensing codes has demonstrated that the CLASIX code adequately predicts the containment transient.
To independently assess the role of the IDIS in mitigating the consequences of a degraded core accident, the staff has obtair.ed technical assistance from the Battelle Columbus Laboratory to analyze the containment atmosphere response to the combustion of hydrogen. The calculations were performed using the MARCH code with a 2-volume model of the Sequoyah containment and assuming a small LOCA consistent with the TVA analysis (S20).
The MARCH code model consistend of a lower and upper compartment, with the ice bed modeled as a junction and nct as a separate volume. The MARCH code features include models of ice bed heat removal, structural heat sinks, return air fans and containment sprays.
The sprays in ice condenser model, however, were presently assumed, due to code constraints, to have heat removal capacity only after the ice is co.1-pletely melted.
The results of analyses performed using the MARCH code were similar to those calculated by TVA using the CLASIX code in that hydrogen combustion was cal-culated to originate in the lower compartment in a series od burns.
Following each burn and concomitant pressure spike, the containment pressure was rapidly reduced until the next burn was calculated to occur.
22-26
The majority of cases analyzed, assuming various ignition setpoints, indicated a peak containment pressure of 23 psi. With a initial containment pressure of approximately 20 psia prior to burning, the pressure rise following a hydrogen burn is approximately 3 psi.
Similar to previously discussed CLASIX analysis, a MARCH analysis was performed for a $2D transient with the arbitrary assumption that the ice bed had completely melted before the onset of hydrogen burning.
Again, this assumption conservatively neglects calculations which demonstrate a large portion of the ice bed would be remaining for this accident.
Nevertheless, calculations show that without ice, the pressure rises to approximately 50 psia which is sufficiently low that containment structural integrity should not be seriously threatened.
The control of hydrogen generated during a severly degraded core accident through the use of a deliberate ignition technique such as the IDIS proposed by TVA requires the consideration of the effects of the environment on struc-tures and equipment.
In an environment where hydrogen deflagration is taking j
place two pronounced effects on the containment atmosphere, namely the con-I comitant pressure and temperature increase, may adversely affect containment structures and internal equipment.
I l
We have discussed the analysis which has been performed to demonstrate that the advantage of deliberate ignition is to limit the amount of hydrogen burn such that relatively low pressure increases are experienced.
The burning of hydrogen, however, was calculated to result in extremely high temperatures for l
the compartment where deflagration occurs. High temperatures would naturally 1
22-27
t follow the burning of hydrogen due to the relatively low thermal capacity of the atmosphere.
However, it is TVA's view, which is shared by the staff, that the CLASIX analysis which neglects of effects of structural heat sinks is too conservative in the calculation of the atmosphere temperature transient.
- TVA, therefore, has an intensive effort underway to modify the code to incorporate heat transfer to structures.
In the interim, however, TVA has provided informa-tion to demonstrate that vital equipment (i.e., equipment needed to safely shutdown the plant and maintain shutdown status) will survive the nydrogen burn environment and perform its intended function.
Survivability of Essential Eouioment The results of testing of equipment at Fenwal laboratories clearly indicates to the staff the need for a more %er ~) investigation on equipment surrvivability ru dA*eugh in containment.
The efforts of TVA are noteworthy in that they clearly recog-nize that " hardening" will enhance the reliability cf equipment subsequent to a hydrogen burn.
For example, all essential safety-related equipment will be sealed to preclude gas penetration to the inner parts of equipment.
Some Tw a
degradation [has been noted in the test resultg\\if miner, it should be fully investigated.
E"--"d' Additional information is needed in the following areas beforisthe NRC staff review is completed on this subject matter:
A.
TVA is to relate the test date to actual equipment and conditions since l
surface heat loads on equipment could be significantly lower in the smaller chambers used at the Fenwal laboratories.
22-28 u
\\
B.
TVA is to assess the appropriateness of the protection provided for the safety-related equipment to determine survivality when exposed to high temperature resulting from the high intensity hydrogen flame.
TVA is proceeding in conformance with our requests for additional information.
Another aspect related to use of deliberate ignition as a mitigation technique for control of excessive hydrogen release is the possible adverse impact of a local detonation on structures and equipment.
TVA has concluded, and the staff concurs, that large detonable mixtures of gas will not be formed inside containment.
The return air fan system, and containment spray system, designed as safety grade systems, operate to insure mixing of the containment atmosphere.
The natural flow path, resulting from pressurization effects, is from the lower compartment through the ice condenser into the upper compartment.
Following depressurization, reverse flow along this path will also occur.
As a supplement to this natural flow path, the return air fan system at a total flow rate of 80,000 cubic feet per minute draws suction from various regions of the containment, principally the upper compartment, and discharges into the lower compartment. This flow rate is sufficient to recirculate the entire lower compartment volume atmosphere (one air change) in the order of 5 minutes, which is a comparatively short time period in relation to the cal-culated interval of hydrogen release of approximately 45 minutes.
Furthermore, the analysis performed with the CLASIX code verifies that large volumes of a detonable mixture will not exist for the postulated accidents.
22-29
-~
w
j j
~
1 l
l The staff, however, believes that there is potential for small volumes of a detonable mixture to be formed. We have therefore required the applicant to consider the effects of local detonations on structures and equipment.
TVA is in the process of evaluating the consequences of detonations, and has stated that investigation o# this phenomenon will continue.
Nevertheless, TVA has provided an estimate of a pressure loading resulting from a local detonation with a peak pressure increase of 180 psi. Although the pressure from this detonation is high the time duration for which the pressure is brief, on the order of 0.5 milliseconds.
Sequoyah Containment and Structural Capacity The use of the IDIS has raised two questions which are addressed herein:
First, will a local hydrogen burn within the ice condenser ace damage the air return duct systetr so that the insulation between the dues and the containment shell would be exposed to hydrogen burning.m Secondly, will local hydrogen i
burning produced by activation of the IDIS pose any threat to the integrity of I
the primary containment structure?
The Sequoyah containment structure is a free standing steel containment shell which is composed of a right circular cyli1 der with a hemispherical dome. The cylinder is 115 feet in diameter and 105 feet tall ans is comprised of eight l
different thickensses of steel plates, 1-1/4 inches at the base and stepping down to 1/2 inch at the top. The dome is comprised of four thickness, 7/16-inch at the intersection with the cylinder are reinforced with horizontal stiffener rings of steel plates placed about 10 feet apart vertically.
Also there are l
22-31 l
l
4
,,,,.. t :. ;.2 s t. w --. - 3
' ' ' ~ -
s..m... o -
-s
.~.o, r,,
vertical stringers spaced around the cylinder and spanning from the base slab up tc.
'op stiffener plate on the dome.
The air return duct system, which provides refrigeration for the ice condensers, runs vertically along the containment shell.
Between the duct and the containment shell is a layer of foam insulation which,is encapsulated by the containment 8
shell and the duct systerr. ' Figure 22.2-2 showy a plan view of the duct system at the containment shell.
TVA estimates that there is approximately 30,000 pounds of polyurethane form in the sealed panels adjacent to the duct.
TVA reports that the foam insulation can withstand a compressive pressure of 30 psi.
The maximum pressure on the outside of the duct work from a hydrogen
~ burn in the area is estimated by TVA to l: Setween 13 to 15*psid.
Th'e duct system was originally designed to withstand an external pressure of 15 psid and was actually tested in compression of 19 psid. Therefore, it can be concluded that the duct system will not be damaged by a hydrogen burn in the ice condenser compartment and that the foam insulation will not be exposed.
Independent analyses were conducted by several organizations and concentrated on predicting the ultimate static pressure for the steel containment shell.
The containment static pressure capacity is based on the actual material properties for the shell material as shown in Table.M.1 - 3 l
Based on these analyses and the statistical variations of the material properties the staff has con-cluded that the lower bound of the containment static pressure retaining capacity is 36 psig due to an overall pressurization of the whole containment.
22-32 l
l
TABLE 22.2-3 MEAN AND STANDARD DEVIATION OF MATERIAL PROPERTIES PROPERTY MEAN STANDARD DEVIATION Modulus of elasticity (normal) 29,000 ksi 174 ksi Poisson's Ratio (normal) 0.3 0.009 3
Yield Stress (lognormal) 47.2 ksi 2.50 ksi Ultimate (lognormal) 66.2 ksi 1.80 ksi Fracture Stress (lognormal) 197 ksi 102 ksi Bolt Yield Stress (lognormal) 105 ksi 2.50 ksi l
22-30
ICE CONDENSER AIR DUCTS AND INSULATION I
i TOP DECX 6
UPPER 8 _ PLENUM U
(
CONTAINMENT]
g WALL AIR HANDLER (TYP.)
l J
i A
g INTERMEDIATE DECK N
i URETHANE ] k iNSEEiCN
/-
^t" "C' (
-)
N /
S sN l
ICE BED N
REcION NNN N
a N
s<
y N '
N\\
t N\\
)LOWERSUPPORTSTRUCTURE N
g LOWER N
PLENUM
\\ U h
FLOOR I
e a
v:me-22 '> b en
\\f
- -i
\\_' e :
n '.
i i
C 44T.
D 6 5 C-T t...
~ ~ ~ ~
I
\\
J l.
i I.
'.k f
[;.,~i 7 l-
~
h*'~t'I _ h, J.
t r,
'a H)
..l.(~
.f.'I'hh*3,.1, g., a b I't d
b I
L *i m.
W B ".^
l '.\\
P 2 f fr,:j, I
' 's
\\
c f
\\
_\\\\
gjf3.'
ll}l f
,,a a
b
'A
,l-
...s..
e
/
J p* / / J Jf r
.J J'
y s '
a
- .r] A
?&J,
s !
/ f
/,fl l *I $.
- -.m'=r-% ;'-
-( i--
- m
%/ I/
U Lo
'/-
// j
/ J f "f "OJ) /l y
- .,*-.\\
~
h,S f
S /
/
.s-
/ f-
/
/
/,
j jw.fl.JbLi5 $in l 1kt.fApLJ I ' '
\\,\\
f-l' F t O C F A C C.
C 'r~-
'MWTt o u A7 p g 997
%ew y==.
- y, n m l
k
[.I(3 b
,1 '
_- ~~~~-, u ~.-~ m m :-
= = m~nm e
g 12
C z ).,
Figure 22.2-2p r
i 22-33 &
With respect to potential hydrogen explosions, TVA has proposed a worst-case pressure profile for local detonation as a 6-foot diameter gas cloud wnich resuits in peak overpressure in a local area of 180 psid with a rise to this peak pressure in 0.10 milliseconds and a disipation back to zero pressure within 0.5 milliseconds. An analysis has been performed by the staff, I
aooroximating the response of the containment to the aforementioned local detonation by modeling the containment as responding in a breathing mode. The staff assumed 40 H instead of the 20 H estimated by TVA in the breathing 7
7 mode.
Damping was conservatively neglected.
The results of this analysis indicate that the effective equivalent statis pressure on the steel contain-ment is approximately 14 psi. This pressure compares favorably with the 36 psi static capability of containment.
The above described approximate analysis shows that the structural integrity of the primary containment would be maintained in the event of a local hydro-
.. b e.
gen detonation of the type postulated.
However,itisnotedthatthisMLba4c p t s o, shape used in this evaluation is very general in Nature.
The actual puba4c l
i shape will depend on individual events.
Further studies are needed to bound the variation in pulse shapes.
Conclusions On the basis of the analyses and testing performed to date, [and subject to satisfactory resolution of the equipment survivability and other related issues], the staff concludes that the IDIS would serve to greatly reduce the consequences of a severly degraded core accident.
The operation of the igniter 22-34
system acting in concert with existing heat removal mecnanisms in the plant, i.e., the ice bed, sprays and passive heat sings, would sufficiently reduce, for certain accident scenarios, the increase in containment atmosonere pressure resulting from the burning of hydrogen.
The staff has previously determined that without a reliable ignition system ice condenser plants could tolerate the burning of only that amount of hycrogen that would be re! eased by appro-ximately a 25% zirconium clading water reaction.
The analyses presented herein which have credited operation of the IDIS have been performed for an accident which was allowed to proceed to the point where 80% of the core cladding reacted.
This increased containment capability to accommodate hydrogen releases represents a substantial improvement over the design basis capatility.
The increased capacity of the containment to tolerate the energy addition due to hydrogen burning is a rate dependent phenomenon such that a distributed ignition system allows for a more controlled burning of lean hydrogen mistures in the containment. Moreover, the actuation of igniters prec'udes the formation of large vlumes of hydrogen gas at a detonable concentration in the event of large hydrogen releases.
Accordingly, we conclude that TVA has satisfactorily shown that the IDIS will l
provide protection against breach of containment in the event that a substantial 1
quantity of hydrogen is generated.
For these reasons, we conclude that, subject to satisfactory resolution of the equipment survivability and related issues, operation of the Susquoyah Nuclear Plant,UnitIwiththeIpSforaninterimperiodof1yearisappropriate.
22-35 t
We have, however, identified items that we believe require f Irther investiga-tion before longterm operation of a distributed ignition system can be endorsed.
We are targeting resolution of these items by January 31, 1982.
TVA has instituted a degraded core task force program inorder to systematically study the benefits and options related to mitigation of the consequences of severe accidents which are beyond the current design bases.
The first quarterly report on this program was filed on December 1, 1980.
The NRC staff has also begun evaluating future design requirements for accident control as part of a process which is anticipated to cintinue for several years. Because many of the proposed mitigation concepts represent state of the art technology the i
staff expects that information needs will evolve throughout this process.
The staff therefore believes that it is prudent to continue the detailed review of the distributed ignition system proposed by TVA for the Sequoyah plant.
As part of this ongoing effort the staff will continue to consider issues related to controlled ignition and evaluation of this technique.
The staff believes that issues such as small local detonations 2.-d quip:=t :urvivabiMty up..
should be studied furtner beofre a long-term commitment to deliberate ignition is endorsed even though our current evaluation leads us to conclude that these issues do not represent an apparent deterrent to use of such a system.
i Consideration of various accident scenarios should also be a part of the staff and licensee efforts over the next year.
Furthert.: ore, combustion phenomena for the environments in a post-accident containment will be studied by both the industry and the staff over the ensuing year ano this future study demands our consideration before final approval of the Sequoyah distributed ignition system can be reached.
22-36
_