ML19322E341
| ML19322E341 | |
| Person / Time | |
|---|---|
| Issue date: | 01/10/1980 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-1693, NUDOCS 8003270217 | |
| Download: ML19322E341 (69) | |
Text
_ _ _
.. > =
Y50*] f's1 k., ', O Oj II hf EATE ISSUED:
1/10/30 9 1.j: d :.J w wi d HJ ':. Q UM MIN 11rES CF BIE
(
/
C
.a ja j p/ $0 ACRS SUBCCMMI'ITEE MErrING ON V
/
FLUID DYNAMICS NOVFPSER 16, 1979 SAN FFANCISCO, CALIFCRNIA "he ACRS Subcoraittee on Fluid Dynamics held a meeting on November 16, 1979, at the 3arrett M: tor Hotel, 501 Post Street, San Francisco, California. ':he purpose of this meeting was to develop information far consideration by the ACKS in its review of the Mark I containment long-term program. Notice of this meeting was published on November 1,1979, in the Federal Register, Volume 44, Nunber 213; a copy is included as Attachment A.
Dr. Andrew Bates was the Designated Federal Employee for the meeting. A list of meeting attendees is included as Attachment B.
A tentative presentation schedule is included as Attachment C.
EXEC'JTIVE SESSION Dr. Plesset, the Subco m ittee Chairman, convened the meeting at 8:30 a.m. and reviewed briefly the schedule for the meeting. Prior to holding discussion with the NRC Staff and the Mark I Owners Group, he solicited comments from the Subce:mittee and its consultants on the subject matter. Dr. Catton raised the following questions to be arsered during the course of the meeting:
1.
How important is the condensation loading of the torus?'; what are the locatfor- -f the pressure transducers?; how the relationship between the maximum and the torus bottom pressure is arrived at?
2.
What is the relationship between the pressure loading and the time between the actuations of Safety Relief Valve (SRV)?
3.
Why does the NRC Staff require more Full Scale Test Facility (FSTF) tests?, and what is the nature of those tests?
MARK I I.ONG-TERM PROGRAM STATUS - MR. GRI.VIS, NRC STAFF Mr. C;imes provided a brief sunnary of the current status of the Mark I containment long-term program indicating that the NRC Staff's acceptance criteria for the long-tern program were transmitted to the Mark I licensees on October 31, 1979 so as to enable them to perform plant unique analysis.
'Ihe NRC Staff and its consultants are in the process of preparing the Safety 20 6003270
_,.4
Fluid Dynamics Noveber 16, 1979 Evaluation Report (SER) for the Mark I containment long-term program which is scheduled to be issued in December 1979.
He pointed out that there are still some differing technical opinions between the NRC Staff and the Mark I Cwners on several aspects of the NRC Staff's acceptance criteria; the NRC Staff and the Mark I Owners have been working together to resolve these dissenting technical issues.
'1he NRC Staff intends to show a film on the pool swell phencmena at a later part of the meeting.
PCCL SWELL TESTING PROGRAMS AND LOAD DEFINITICN METHODOLOGY Pool Swell Loads - Mr. V. Tashilan, General Electric Concany (GE)
Mr. Tashjian reviewed briefly the pool swell phenomena and indicated that' the following pool swell leads are of coin interest:
1.
Torus vertical loads 2.
"brus submerged pressure loads 3.
Torus airspace pressure loads 4.
Vent system impact and drag loads 5.
Subnerged structure impact and drag loads 6.
Vent header deflector loads Mr. Tashjian pointed out that, since there are some dissenting technical opinions between the NRC Staff and the Mark I owners Group on the technical assessment of 3-D/2-D upload multiplier, he would like to concentrate his discussion on this issue.
He would also like to present some technical justi-fication for the Mark I Owners' position on the pool swell shape.
Mr. Tashjian stated that several tests were conducted at the GE 1/4 Scale 2-Dimensional Test Facility to determine the pool swell loads; the data obtained frcn these tests were used to define the pool swell loads as delineated in the Mark I plants Icad Definition Report (LIR).
On behalf of the Mark I Owners Group, GE also performed an assessment of the 3-Dimen-sional effects on the Mark I plants by using the data from the tests con-t ducted at Electric Power Research Institute (EPRI) 1/12 scale 3-Dimens Test Facility.
A comparison of the GE 1/4 Scale 2-Dimensional test data and the EPRI 1/12 Scale 3-Dimensional test data (Attachment D, page 1) indicatec h
that the torus uploads observed in the 2-Dimensional tests are consistently higher than those observed in the 3-Dimensional tests.
Based on this comparison,
Nevernber 16, 1979 Fluid Dynamics he believes that the torus upload obtained frcrn the 1/4 Scale 2-Dimensional tests is conservative and, therefore, there is no need to apply an uncertainty f-factor, as required by the NRC Staf f, to account for the 3-Dimensional effects.
i Indicating that he was given to understand in one of the previous Fluid Dynamics j
subcommittee meetings that the orifice in the EPRI 1/12 scale test choked, but E
the one in the GE 1/4 Scale did not choke, Dr. Catton asked whether this fact has been factored into the comparison of the 1/4 Scale 2-Dimensional and the 1/12 Scale 3-Dimensional test data.
Mr. Tashjian stated that he believes that consideration has been given to this fact; however, he will confirm whether it has been factored into the comparison of the 2-Dimensional and 3-Dimensional test data at a later part of the meeting.
Mr. Tashjian indicated that the Mark I owners Group also looked at the results'of the 2-Dimensional and 3-Dimensional tests conducted at the Iawrence Live
'Ihese tests conducted at the 1/5 Scale Test Facility were to Laboratorf (LLL).
provide data to the NRC Staff to aid in their evaluation of Mark I containment A comparison of the LLL 2-Dimensional and 3-Dimensional test pool swell loads.
data indicates that the peak torus uploads are higher in the 3-Dimensional case than in the 2 *:imensional case, thus giving an average 3D/2D multiplier somewhat Based on the review of the LU data, the Mark I owners Group greater than one.
believes that the following factors influence the differences in the peak uploads between the 2-Dimensional and 3-Dimensional tests:
Structural oscillations of the 3-Dimensional test facility 1.
(Attachment D, page 2).
Non-Simultaneous vent clearing between the 2-Dimensional and 2.
3-Dimensional test facilities due to variations in the initial conditiers.
Interaction between the 2-Dimensional and 3-Dimensional test 3.
facilities (Attachment D, page 3).
Capacitance and FL/D (flow resistance) differences between the 4.
2-Dimensional and 3-Dimensional test facilities due to the variation in the location of the orifices.
- ~
Fluid Dynamics November 16, 1979 Mr. Tashjian pointed out that, if the above factors are taken into account, he believes that the LLL 3-Dimensional and 2-Dimensional peak uploads would be essentially equal.
Mr. Tashjian stated that, based on the comparisons of the test data obtained from various test facilities, the Mark I Owners Group arrived at the follow-ing conclusions:
1.
A comparison of the EPRI 3-Dimensional and GE 2-Dimerisional test data shows that the 3D/2D upload nultiplier is < 1.
2.
A comparison of the LLL 3-Dimensional and 2-Dimensional test data confirms that the 3D/2D upload multiplier isT 1 when facility and test conditions are matching.
Pool Swell Shace and Vent Header Imoact Timing Mr. Tashjian provided justification for the Mark I Owners position that the vent header irpact sweep times defined usir.' the Mark I IDR methods are sufficiently conservative and bounding.
Mr. Tashjian stated that the LDR definition of pool swell displacement, velocity, and vent header impact timing in the longitudinal direction were obtained by interpolating halfway between the results given by the EPRI 3-Dimensional vent orifice and downcomer orifice tests. Subsequent to the developnent of the LDR definition, EPRI performed a series of split orifice tests with orifices placed both in the main vent and downcccers. A ccznpari-j son of the interpolated vent header impact times with the results of the EPRI split orifice tests indicates that the interpolated vent header impact times as specified in the Mark I IDR are conservative.
Dr. Catton asked whether the Mark I Owners Group has run any tests by usirg j
several orifices (4 or 5 orifices) and made a comparison of these results with the split orifice test results.
i Mr. Kennedy frem GE stated that they did run such tests at the same scale and l
compared the results with the split orifice test results; they observed that the results are equivalent. However, he believes that the results may be f
different when one goes from a smaller scale to the full scale because of the compressibility effects.
[
Fluid Dynamics mber 16, 1979 In response to another question from Dr. Catton, Mr. Kennedy stated that the 4
compressibility effects at small scales (1/4 Scale and below) are negligible.
Dr. Catton wondered how the compressibility effects can be neglected at small scales. He stated that he would like to see the analysis pertinent to this issue.
Indicating that there is a potential for either low or medium cycle fatigue, due to some repetitive pressure loads associated with certain safety systems (such as safety relief valve discharge), which would lead to the degradation of the pressure boundary, Dr. Bush asked whether they have looked at the implications of strh low-probability accident type loads when superi= posed on a degraded pressure boundary.
Mr. Grimes stated that they will look into this issue.
NET VERTICAL PRESSURE LCAD IN THE TORUS - MR. J. RANLET, BROCKHAVEN NATIONAL I.ABORATORY (NRC CONSULTANT)
Mr. Ranlet stated that the NRC Staff's acceptance criteria for the Mark I containment long-term program requires that the downward and upward net vertical pressure wds on the torus shall be derived from a series of plant-specific 1/4-Scal Test Facility (OSTF) tests. However, based on the review of the pool swell tests (2-Dimensional and 3-Dimensional) performed by the Mark I Owners Group and the confirmatory tests performed for NRC at the Iawrence Livermore Iaboratory, the NRC Staff believes that the following margins should be applied
~
to each loading phase:
1.
For the net upward load, a margin equivalent to a value of 21.5% (15%
to account for the uncertainties of 3D/2D comparisons plus 6.5% as derived from the statistical analysis of the entire QSW data base) should be acclied to the average upward loads of the QSTF plant-specific test results.
2.
For the net downward load, a margin equivalent to a value of 6.3 to 15.54 (derived from the statistical analysis of the entire QSTF data base) should be applied to the average d,.unward loads of the I
Q37 plant-specific test results.
Mr. Ranlet provided justification for requiring that a margin equivalent to a value of 15%, to account for the uncertainties associated with the 3D/2D l
Novenber 16, 1979 Fluid Dynamics comparison, should be applied to the net average upward load obtained frcm He pointed out that the Mark I Owners Group, the plant-specific OSTF tests.
af ter comparing the results of the GE 1/4 Scale 2-Dimensional tests and EPRI 1/12 Scale 3-Dimensional tests, chose the GE 1/4 Scale 2-Dimensional tests r
However, based en the data as the basis for defining the peci swell leads.
review of the comparisen of the GE-E?RI test results, the NRC Staff has con-3 c)uded that it should not be used to assess the possibility of a 3-Dimensional effect on pool swell uploads for the following reasons:
FRI 1/12 Scale 3'-Dimensional Test Facility represents the Browns 1.
Ferry Plant configuration; the NRC Staff believes that Browns Ferry gecmetry is not prototypical of Mark I Plants; the 45 downcomer configuration causes an early bubble break* rough; such an early breakthrough phenomena attenuates the torus up-loads because the wet well airspace is not compressed suffi-m is early bubble breakthrough did not occur in any ciently.
)
of the other plant configurations.
(For different plant con-figurations, see P.tachment D, page 4).
W e tests were conducted at full Ap and at 3 feet 4 inches 2.
cergence; such test conditions would minimize the reduced st pol swell effects.
EPRI tests were conducted at higher values of flow resistance 3.
than the GE tests; such high flow resistance would reduce the net uploads.
We downcomer orifice size variation caused a distorted pool 4.
swell, thus resulting in reduced uploads.
Mr. Ranlet stated that in order to obtain additional data base and also confirm the 3-Dimensional effects on gol swell vertical loads, confirmatory tests were conducted at the LLL 1/5 Scale 2-Dimensional and 3-Dim Rese tests were performed using Peach Bottom plant con-Test Facilities.
W e results of these mergence.
figuration at zero Ap and at 4 feet reduced st tests (Attachment D, page 5) indicated that torus uploads are higher in the tey have also compared 3-Dimensional case than in the 2-Dimensional case.
the results of some GE (1/4 Scale 2-Dimensional) and EPRI (1/12 S Dimensional) tests which were not used by the Mark I owners Group in the GE-EPRI ccuparisons; the results of this emparison show that the torus
i Fluid Dynamics November 16, 1979 i
uploads are higher in the 3-Dimensional case than in the 2-Dimensional case (Attachment D, page 5).
Mr. Ranlet pointed out that, in order to determine whether the experimental trend as indicated by the LLL test data was due to a 3-Oimensional effect on pool swell or possibly a mis-match of the 3-Dimensional and 2-Dimensional sectors, a 1-Dimensional transient pool sell analysis for boe.h the LLL 2-Dimensional and 3-Dimensional sectors ses conducted. W e results of this analysis have shown that the LLL 2-Dimensional and 3-Dimensional sectors were indeed mis-matched due to differences in capacitance and resistance.
'Iherefore, the NRC Staff believes that, to account for the uncertainties associated with the 2-D and 3-D ccxnparisons, the Mark I owners should apply a margin equivalent to a value of 15% to the average uploads of the CSTF plant-specific test results.
Mr. Steiner, from GE, expressed concern indicating that the NRS Staff's acceptance criterion for the torus upload has excessive conservatism and he believes that it will have significant impact on the torus tploads.
Mr. Grimes, NRC Staff, stated that, on the basis of analyses and nodel tests, they have developed the criterion for the torus uploads.
tey have performed several assessments on the basis of the existing knowledge without giving too much consideration for its potential impact or consequences.
He pointed out that their main aim is to restore the margins of safety in the plant designs and they have incorporated appropriate techniques in their criterion to reduce the excessive conservatism without affecting the margins of safety.
POOL SWELL FLOW DISTRIBtfrION EFFECTS - DR. KOSSAN, SC CONSULTANT Dr. Kossan stated that the main objectives of his p.esennation are to:
1.
provide justification to the NRC Staff's acceptance criterion' pertinent to the vent header impact timing which requires that the vent header and vent header deflector timings should be derived from the 3-Dimensional test data using orifices only in the main vent line, and 2.
show that the techniques employed to develop Mark I LIR definition of the vent header impact timing may not be appropriate.
t l
l
November 16, 1979 Fluid Dynamics h
t header Dr. Kossan indicated that the Mark I LCR definition of t e ven h EPRI impact timing was obtained by interpolating the results of t e He stated that 3-Dimensional vent orifice and downcomer orifice tests. techniques employed h
there are several uncertainties associated with t e for the vent by the Mark I Cwners Group in developing the definitionBased on sever header impact timing (Attachment E, page 1).
tions:
and model tests, the NRC Staff has made the following observa by Se orifice sizes for the EPRI test model were establi He believes that 1.
running tests in 1/12 and 1/31 Scale models.
it is very difficult to obtain the exact orifice size experi-Without using the appropriate orifice size, it is mentally.
s very difficult to obtain the desired flow; such concern wa h
the confirmed by the observation made in the EFRI tests t at d
ratio 'of the highest to the lowest downcomer flow rates seeme ll and excessive.
Flow calibration tests were run with no water in the wetve However, he be-2.
with uniform exit pressure at all downcteers.
lieves that during early bubble growth, bubble pressure can vary from one downcomer to another, thus causing a non-uniform e flow pressure.
Le split orifice and the downcomer orifice provide the same 3.
distribution and sweep time.
i Le split and downcomer orifice tests probably had an excess ve 4.
flow ratio.
NRC Staff believes Dr. Kossan stated that for the reasons given above, thek I owners Gro that several of the technicues used by the Mar vative. Wey are also the vent header impact timing do not seem to be conser in the EPRI tests pro-not sufficiently confident that the flow distribution Based on vides a prototyoical representation of a pool swell response.
h t the vent several analyses and test results, the NRC Staff believes t a i
ibution and th: best orifice tests provide the most prototypical flow d strD erefore, the N estimate of vent header impact timing.
should be derived i
that the vent header and vent header deflector tim ngs from the main vent orifice tests.
1 Fluid Dynamics November 16, 1979
)
I Dr. Plesset asked whether the Mark I owners Group has strong reservations about the NRC Staff's criterion on the vent header impact timing.
Mr. Steiner str'ed that the Mark I Owners Group believes that the NRC Staff's criterion has excessive conservatisn and it wuld have significant impa
.a 'he design of the pool structure as well as other structures above the pool.
VENT HEADER DEFLECTCR LCAD DEFINITION - MR. KEhWEDY, ACUREX CCRPCRATICN Mr. Kennedy stated that the vent header deflector is located between the pool surface and the vent header for the purpose of deflectino the risina surface of the pool water thus preventing the high velocity impact of the
. water on the vent header (Attachment F, page 1).
He discussed briefly different types of vent header deflectors that are beirg considered for use in Mark I plants.(Attachment F, page 2).
Mr. Kennedy stated that there are two types of methodology used in the prediction of deflector loads:
1.
Direct use of deflector load data obtained from the QS'IT plant-specific tests.
2.
Analytical Methods.
With regard to the prediction of vent header deflector loads by using experimental data, Mr. Kennedy stated that scaled models of actual deflectors, which would be eventually used in the actual plant configura-tions, were installed in the CGTF Test Facility and tests were run to obtain necessary data (Attachment F, page 3) for vent header deflector load predictions.
Mr. Kennedy stated that for those plants for which the vent header deflector has not been tested through plant-specific CSTF tests, a semi-empirical methodology has been used to calculate the vent header deflector loads.
The load is assumed to consist of impact transient, steady drag, and acce-1eration drag; all these components are defined and added together to obtain the vent header deflector loads. '1he empirical correlations used to calculate the impact transient, steady and acceleration drags are included in Attachment F,
i.
Fluid Dynamics mvember 16, 1979 page 4.
Se instantaneous poel velocity necessary to evaluate the empirical expression is obtained from the QSTF test movie data. We results of the comparison of the QSTF test data and the calculated data (Attachmen't F, pages 5-7) show that the calculated data is conservative and bounds the measured data.
He stated that the comparison of QSW test data and analytical data obtained with different deflector types and locations indicate that the analytical data over-predicts the deflector load about 33% (Attachment F, page 8).
Mr. Kennedy pointed out that the NRC Staff has some concerns about the methodology used by the Mark I Owners Group in predicting the vent header deflector loads. S e NRC Staff believes that the Mark I Owners Group has over-predicted the pool velocity component, but tmder-estimated the drag coefficient; they expressed concr,rn that this mis-match may produce some non-conservative vent header deflector loads. As a result, the NRC Staff has developed some drag coefficients and suggested that the Mark I owners Group should use those in calculating the vent header deflector loads.
Consequently, the Mark I Owners Group calculated the vent Nader deflector loads by applying the drag coefficients developed by the NRd Staff alorg with the pool velocity obtained fra the 052 movie data and observed that the loads were about 20-30% higher than it was predicted earlier; they believe that it is overly coriservative and will have significant impact on the structural design.
Mr. Kennedy indicated that as a resolution to EC Staff's concerns on this issue, the Mark I owners Group intends to redefine the pool velocity com-ponent; they are in the process of doing this and the results will be dis-cussed with the NRC Staff in the near future.
CEFLECTCR IDAD DATA ASSESSN - DR. SOiIN, MIT, N'C CONSULTANT Dr. Sonin reviewed briefly the methodologies proposed by the Mark I owners Group in predictirg' the vent header deflector loads and the NRC Staff's position in accepting those methodologies. He stated that the Mark I Owners Group' has proposed two methodologies to determine deflector loads:
1.
Alternative A i
In this methodology, the Mark I owners Group has proposed to use the data obtained fra the plant-specific 037 tests for predicting the vent header deflector loads.
November 16, 1979
. Fluid Dynamics Based on the review of the plant-specific CSTF tests results and other appropriate information, the NRC Staff believes that this methodology may be used to estimate the vent header de-flector loads subject to the following redifications:
For cylinderical types of deflectors (Pipe, Pipe with Angles a.d Pipe with Tees), the loadirg transients should be adjusted to a.
E include the empirical impact spike that is derived frcm the impact tests of cylinders conducted by EPRI.
We 3-Dimensional pool swell effect.s should be interpreted conservatively as required by the NRC Staff; the QSW plant-b.
f specific loads must be adjusted to account for the effects o impact time delays and pool swell velocity and acceleration differences which result from the uneven spacing of the down-come'rs.
Mien applying the load to a Mark I containment deflector, the inertia due to the added mass of water impacting the deflector c.
should be accounted for in the structural assessment.
Alternative B_
Sis methodology, proposed by the Mark I C ners Group, consists 2.
de-of a semi-empirical approach to calculate the vent header flector loads for those plants for which the deflectors were not We load is assumed tested by the plant-specific OSTF tests.
i to consist of impact transient, steady drag, and accelerat on drag and all these components are defined and added together to obtain the deflector loads.
Based on the review of this methodology, the NRC Staff believes t the cylinderical
~
.that the steady drag coefficient'used to compu e B ey also believe type deflector loads are non-conservative.
that an appropriate force transient for the wedge-type deflectors Unless all the components associated with has not been specified.
Staff believes that this methodology are conservatively defined, the NRC Derefore, this approach may not provide acceptable deflector loads.
d in this the PRC Staff requires that the steady drag coefficient use
~
Fluid Dynamics
/e sr 16, 1979 1
methodology should be redefined. Rey also recuire that, when applying the load to the deflector, the inertia due to the added mass of the water impacting the deflector must be accounted for in the structural assessment.
Dr. Sonin discussed briefly the NRC Staff's definiricts for the impact transients and steady drag coefficient (Attachment F, pages 9-13) and stated that the NRC Staff believes that use of these values in the Mark I owners Group semi-empirical methodology may produce conservative deflector loads.
Dr. Hanauer, NRC Staff, commented that he believes that Mark I owners semi-empirical methodology for calculating the deflector loads provide inappropriate results not mainly because of the non-conservative drag coefficients but because of the use of the overly conservative pol velocity. He believes that the pool velocity is not at all representa-tive of the actual situation.
Dr. Sonin pointed out that, after realizirg the excessive conservatism associated with the prediction of the pool velocity, the Mark I owners Group has proposed to redefine the velocity component and use that refined velocity companent along with the drag specifications provided by the NRC Staff in the Mark I Owners Group semi-empirical methodology. He believes that this proposed technique may be a reasonable solution to this issue;
~ however, the NRC Staff has to review the results of this technique to assure that adequate conservatism exists in this techinque.
Prior to hearing the other scheduled presentations, the Subecznmittee and its consultants viewed the following films:
1.
Computer simulation of the response in the Mark I torus, developed by the Iawrence Livermore Laboratory, to look at the fluid struc-ture interaction effects.
2.
Summary of tests conducted at the Full Scale Test Facility (FSTF) -
W is film was developed by the General Electric Company for the benefit of Mark I plant Owners to give an overview of the FSTF tests.
i m.
-~
Fluid Dynamics Novmber 16, 1979 FSTF TORUS SHELL PRESSURES AND LCAD DEFINITION BASES Description of ESTF - Mr. Torbeck Mr. Torbeck stated that the main cbjective of the FST. program is to perform appropriate tests using a representative Mark I containment torus and obtain data to define hydredina.ic 1 cads and dyn=nic structural re-spense resulting frem steam condensation phenomena. He discussed briefly the main characteri'stics of the FSTF (Attachment G, pages 1 and 2).
He also discussed the instrteentations used in the FSTF and their locations (AttacM.ent G, pages 3 and 4).
He pointed out that in the FSW tests a prototypical segment of a Mark I torus and vent system were subjected to ten steam and liquid blowdowns (Attachment G, page 5) simulating a range of toss-of-Coolant Accidents (LOCAs). %e parameters which were varied in the FSTF tests include downeomer sutrnergence, initial pool temperature, blowdown of liquid and stem, and initia?. wetwell pressure (Attachment G, page 5).
Mr. Torbeck discussed briefly the result of the Condensation Oscillation (CO) tests conducted in the ESTF. Se result of the tests conducted in the FSTP to determine the CO loads indicate that the splitude of the pressure oscillations induced by condensation oscillations on the torus shell is dependent on the break si::e and the phase of the blewdown fluid (liquid or steam). Se highest pressure splitude was observed during the large liquid break test which simulated the design basis accident conditionse h erefore, the results of the large liquid break test were used as a conservative basis for CO load definition for the design basis accident.
He pointed out that in some of the tests conducted with break size equal to 25 percent of the design basis accident area, strong condensation oscillation did not persist. However, in those tests, they observed some pressure oscillations during air carry-over, but the pressure splitudes were I
l bounded by the peak values of the chugging pressures which occurred.following the air carry-over period.
l l
With regard to the test results pertinent to the torus wall pressure amplitude, Mr. Torbeck stated that pressure measurments obtained from various locations on the torus shell show that the longitudinal pressure
reovenber 16, 1979
-n-Fluid Dynamics i lly escillation amplitude distribution alorg the torus centerline is essent a he test results also indicate that the maximum pressure occurs
- uniform,
._l at the bottom dead center of the torus.
In response to a question from Dr. Catton regarding the location of pressure located transducers, Mr. Torbeck stated that there is a pressure transducer innediately below one of the vents; however, that pressure transducer did no k'
register the maximum pressure amplitude.
1 Dr. Catton wondered how a pressure transducer located very close to the ve exit failed to show the maximum pressure amplitude.
Mr. Torbeck pointed out that the dominant frequencies (about 5 and 10 H:)
Bey were observed during the large steam and large liquid break tests.
B erefore, also observed that these frequencies vary during other tests.
tively the load specification frequency ranges, which were selected to conserva and 8 to 16 Hz.
bound the dominant frequency variances, are 4 to 8 H:
Mr. Torbeck discussed briefly the FSTF test results pertinent to CO loads on suinerged structures, downcomers and vent systems.
Mr. Torbeck reviewed the results of the FSTF tests conducted to gather He pointed out that data for use in the definition of chuggirq loads.
four among all the FSW tests conducted, they observed chugging only during te FSTF blowdowns which simulated the snall steam break of the tests, accidents (Attac!nent G, page 6) produced the most severe chugging loads, and the data fra these tests wre used as a basis for chugging load He p:>inted out that the FSTF tests conducted with large specifications.
steam and large liquid break accidents did not produce large chugging like behavior.
f Mr. Torbeck stated that by emparing the psol temperature at the bottom o ble to the downcmers with the average downcomer steam mass flux they were a come up with a bounding value; this comparison also shows that chtsgin not normally occur when the pool temperature is high (Attacinent G, pag He pointed out that a cmparison of the dynamic stresses obtained fra t ll condensation oscillation (Large Liquid Break) test and the chugging (Sma
Fluid Dynamics Novemba 16, 1979 Steam Break) test indicate that the condensation oscillations produce significant loads on the torus and downcmer structures and the chugging produces leads that are considerably lower than the condensation oscillations (Attachment G, page 8).
Indicating that the stiffness of the 22.5 sector of the torus which was used to run the FSTF tests is much different than that of a full-scale torus, Dr. Bush wndered how the stresses obtained by running tests in a 22.5 sector can be extrapolated to obtain stresses for the full-scale torus.
Analysis of Full Scale Test Facility for Condensation Oscillation Loading -
Mr. Broman, Bechtel Power Corporation Mr. Broman stated that, based on the comparison of the analytical data and test data, they, observed that poor correlation exists between these two; sub-sequent evaluation of the test data and the structural analysis techniques indicated that the Foor correlation was due to the effects of fluid structure interaction on measured wall pressures. Consequently, in September 1978, they have started the structural analysis of FSW to:
1.
extract rigid wall pressures from test data, 2.
develop analytical techniques which will predict test results for structural response, 3.
assess structural response based on LIR load definitions.
Mr. Broman discussed briefly the basic concepts of the structural analysis that was performed (Attachment G, page 9).
He provided also a brief descrip-tion of the overall procedure used in performing the analysis and of the work performed with respect to fluid structure interaction (Attachment G, page 10).
Mr. Broman reviewed the ESIF analytical model indicating that it is a. finite element coupled fluid-structural model of the torus developed using NN cmputer program (Attachment G, page 11). He indicated that the FSW analytical model was verified by the following methods (Attachment G, page 12):
1.
Static check cases.
2.
Comparison against the results of the shake tests which were performed l
using an eccentric mass shaker.
Fluid Dynamics Novenber 16, 1979 3.
Comparison against the FSTF test data to determine tha ability of the FSTF analytical model to predict FSTF structural respnse to condensation oscillation loading.
Mr. Eroman pointed out that t".e results of the analysis indicated that the maximum difference betwen the flexible and rigid wall pressure occur in a frequency range of 16 to 17 Hz (Attachment G, page 13). He indicated that the same tyoe nf behavior was also observed during the shake tests.
With regard to the results of.the verification performed by using the FSTF test data, Mr. Broman stated that such a verification indicated that the maximum contribution of the source to the cumulative axial membrane stress (total load) is negligible beyond a frequency range of about 30 Hz (Attachment G,.page 14).
Mr. Broman also pointed out that a comparison of the test data with the analytical data and the UR data indicated that the analytical method is conservative with respect to the test data (Attachment G, page 15).
Condensation oscillation Load Ocfinition - Mr. Saxena, General Electric Company Mr. Saxena reviewed breifly the approach used to develop the condensation oscillation load definition for Mark I torus shell (Attachment G, page 16).
~
He stated that data from the entire condensation oscillation tests of the FSTF were examined and the large' liquid and large steam break test runs w re selected to obtain data base. Se highest pressure amplitude was observed to occur during the large liquid break test.
From the large liquid and steam break tests maximtzn pressure anplitude data segments were selected as data base for use in the load definition (Attachment G, page 17). He discussed briefly the steps taken to reiuce the FSTF test data for use in the Mark I torus shell load definition. (Attachment G, pages 18 and 19).
Mr. Saxena pointed out that, based on the evaluation of appropriate FS*IT test data, they have selected a bounding frequency range to cover the range of frequencies expected in all Mark I plants. Se load specification fre-quency ranges, which were selected to bound conservatively the dczninant fre-quency variances are 4 to 8 Hz and 8 to 16 Hz.
Ebr a plant unique structural l
F"'
~
November 16, 1979
. Fluid Dynamics d between 0 and 50 evaluation, the structural response from each 1 H: banW e O to 50 Hz has been analyzed and sumed to get the total response.y spectrtn of H:
total range analyzed would include the frequenc range which produces the maximum response.
in plant unique Mr. Saxena indicated that in order to apply the FSTF data geometries, adjustments were made to: account for fluid struc FSW l.
data, and account for the differences in the ratio of the paol surface area to vent cross sectional area among the Mark I plants.
2.
h t the condensation Mr. Saxena stated that the Mark I Owners Group believes t abeen developed co oscillation load definition for Mark I torus shell has h i ues.
servatirely using appropriate P5W test data and analytical tec n q h load Indicating that the fluid structure interaction factor used in t e k d how definition analysis was tmique to the FSTF facility, Dr. Zudans as e d
such a factor could be used in developing plant-specific loa s.
ture Mr. Broman responded that the main objective of the fluid struc ll load interaction analysis performed on FSTF was to develop a rigid wa i
ffects that are which would not include the fluid structure interact on eBerefore, e unique to a specific Mark I plant.
i to that speci-should include the fluid structure interaction effects un que fic plant.
C CCNSULTANT_
TORUS SHELL CCNDENSATICN LCAD ASSESSMDTT - DR. BRDNEN, N Dr. Brennen reviewed the M1C Staff's position with regar d
ncy of the data oscillatin load definitions and their concerns on the a equa d
tion oscilla-base used by the Mark I CWners Group in developing the con ensa tion load definitions.
f the oscillatory Dr. Brennen stated that condensation oscillation loads re i t behavior of pressure loads imparted to structures due to the unsteady trans e i
~
Fluid Dynamics Novenber 16, 1979 the steam released during a 14CA, occuring near the end of the downcaners. We phenomenon of unsteady condensation involves an unsteady turbulent two-phase flow. He believes that it is very difficult to model such flows throt.gh analy-tical methods. Werefore, the Mark I Owners Group has developed the condensation oscillation load definition based en the results of seme tests conducted in FSTF.
Dr. Brennen pointed out that the maximt.sn condensatio'n oscillation loads in FSTF were found to occur during the large liquid break test. The Mark I owners Group has conducted only one such large liquid break test and based on the results of that one test, they have developed the condensation oscillation load definit'.ons.
The NRC Staff believes that the large liquid break test conducted by the Mark I owners Group provides only one data point; therefore, they believe that statis-tical variance or load magnitude uncertainty cannot be established with adequate accuracy from a single test run. We EC Staff believes that the data base used by the Mark I owners Group for defining the condensation oscillation loads is inadequaate to establish a reasonable measure of the uncertainty in the loading functions.
Dr. Brennen stated that the EC Staff's position on the condensation oscillation load definition is that they accept the loads developed by the Mark I Owners Group with the condition that each Mark I licensee should perform additional FSTF tests to establish the uncertainty in the condenration oscillation loads and confirm the adequacy of the load specifications.
In response to a question from Dr. Zudans as to whether the NRC Staff expects to get significantly different data from the additional tests, Dr. Brennen stated that there may not be any significant difference; however, until additional tests are run, they may not be able to assure that adequate conservatisn exists in the l
data base used by the Mark I Owners Group in developing the condensation oscilla-tion loads.
Dr. Zudans expressed his personal opinion indicating that the large liquid break test conditions are prototypical for the Mark I design and therefore, he believes that additional tests are not necessary.
Novembog 16, 1979 Fluid Dynamics his time is that each Mr. Grimes stated that the NRC Staff's position at t stablish the un-Mark I licensee should conduct additional FSTF tests to e servatism and certainty levels in the condensation process to assure cen However, if i
minimum required level of safety in the containment des gn.l inc this matter, the ACRS provides other types of guidance in reso v NRC Staff will give consideration.
A h type of additional In response to a question from Dr. Catton regarding t eh t the Mark I licensees tests required by the NRC Staff, Mr. Grimes stated t a in the FSTF.
should conduct two additional large liquid break tests Staff would de if In response to a question from Dr. Bush as to what the NRCloads, Mr. G the additional two tests show lesser condensation proposal for reducing i
that they may ask the Mark I owners Group to subm t.ai h the additional test the condensation oscillation loads in accordance w t data.
d his concern about Mr. Logue, dairman of the Mark I owners Group, expresseindicating that the a the NRC Staff's requirement for additional testis ificant impact on the i
tional tests required by the NRC Staff will have s gndifications. He cost and schedule for completion of the Mark I plant modditional large believes that the N:ic Staff has never specified that only two liquid break tests ate necessary.
lts of the test already the results of the additional tests differ from the resuh differences and may eve conducted, the IRC Staff may ask for reasons for sucindefinite process.
ask for more tests; it seens like this is going to be an C Staff will be Dr. Plesset stated that he does rut believe that the NR h Mark I Owners.
unreasonably requiring more and mou FSTF tests from t e there is a need for Mr. Sobon from GE commented that he does not believe t following factors:
l additional FSTF tests if consideration is given to the h
In the ETE tests and analytical methods for predicting t e d the containment response, the Mark I Owners Group has ignore 1.
F
. Novmber 16, 1979 Fluid Dynamic:
contribution of heat sink in the drywell to the mass flux of He believes that, during the initial phase, the vent system.
the heat. sink will absorb some of the energy thus reducing the mass flux of the vent syste for a short period of time.
(
S e configuration of the FSTF is made in such a way to purge 2.
In the air in the drywell in a very short period of time.
I view of the fact that any air content in the steam condensation phase would tend to reduce the pressure amplitudes, the quick purging of the air increases the pressure aplitude.
We load specification includes the summation of the amplitude 3.
from each 1 Hz frequency band between 0 and 50 Hz range; the loads defined in this way are about three times higher.than those observed in the FSTF tests.
Mr. Sobon stated that all the above factors coupled together will form a basis to preclude the need for any additional FSTP tests.
DCMCOMER CONDENSAT'T LOAD DEFINITION - MR. BROMAN, BECHTEL POWER CORP Mr. Broman reviewed the work that is underway to reevaluate the downcomer loads during condensation oscillation. He pointed out that this work is being carried out as a result of the concerns expressed by the NRC Staff about the FSW test data used in defining the downcomer condensation oscillation loads.
~
Mr. Broman stated that the main approach is to postulate a load definition, based on pressure data measured during the large liquid break FSW test, for the downcomers during condensation oscillation; this postulated load defini-Analytical tion will be analyzed using NASTRAN finite element computer model.
model will be developed to simulate the downcomer configuration as used in W e analytical model will be verified the large liquid break FSW test.
using " Jacking" and " Snap" tests of the downccriers (Attachment H, page 1).
W e results obtained through the analytical method will be campared with the large liquid break FSTF test data to determine the appropriateness of the postulated downcomer. condensation oscillation load definition; if there seems to be an improper correlation, the Mark I Owners Group will look at phasing between pressures in adjacent downcomers.
I
Fluid Dynamics November 16, 1979 In response to a question from Dr. Zudans with regard to the effects of bending, caused by pressure imbalance outside of downcomers, on the down-comer loads, Mr. Torbeck stated that they do not expect much pressure variation around the downcemers as a result of condensation oscillation.
DC'*M CCNDENSATICH IDAD ASSESSMNT - MR. GRIS, NRC STAFF Condensation Oscillation Loads on Untied Downcemers Mr. Grimes stated that, based on the review of the information provided in the Mark I IIR with regard to the condensation oscillation loads for " untied" downcmers, the NRC Staf f requires that a more accurate determination of the FSTF downcemer response characteristics (natural frequency and damping) should be developed to assure a conservative dynamic load factor scaling.
In view of the fact that a frequency of 5.5 H: is observed to be the natural swinging mode of the downcmers, the driving frequency for the FS1F plant unique dynamic load factor should be assumed to be 5.5 Hz. The NRC Staff believes that, with the correction mentioned above, the proposed Mark I Lm specification will provide a conservative estimate of the condensation oscillation loads on " untied" downecmers.
Condensation oscillation Loads on Tied Downcomers Mr. Grimes stated that the results of the comparison of loading conditions observed in " tied" and " untied" downcomers in the FSIT indicated that the strain measurements observed in the " tied" downecmers were significantly lower than those for the " untied" downcmers. Based on a detailed analysis of the downcomer-vent header system, the Mark I owners Group provided the
,(
reasoning for the differences indicating that the downecaer loads during
~
condensation oscillations were primarily an in-phase vertical thrust load caused by the pressure oscillations inside the downcm.er, with only a small lateral loading contribution. However, the NRC Staff believes that the existing information seems to be inadequate to provide answer to the question about how well the pressure inside the downecners are phased to establish
November 16, 1979 Fluid Dynamics terefore, they require that an improved load definition a load definition.
m for the " tied" downectners be developed from the FSTF data, Chuacina Leads on Untied Downcomers Based on the review of the infor ation provided by the Mark I Omers Group
,y with regard to the " untied" downcomer chtgging leads, the NRC Staff believes lied that the proposed Resultant Static Equivalent I. cad (RSEL) spectra, as app to the " untied" downcomers, is acceptable with the 'following exceptions:
We load specification should be based on the maximum measured 1.
RSEL load in the FSTF.
Se fatigue loading analysis for each downceer shall be based on 2.
a statistical loading with a 95% non-exceedance probability.
S e multiple-downcomer loading to assess statistical directional 3.
dependence shall be based on a probability of exceedance of 10 per LOCA.
Chuecine Loads on Tied Downcomers With regard to the chtqging loads on " tied" downecners, Mr. Grimes stated the NRC Staff's criteria require the following:
1.. %e Mark I IIR should specify adequately a procedure for deriving the strain in the tie bar between a downecaer pair.
We load direction shall be taken as that which results in the 2.
worst loading for the tie bar and its attachments to the down-Comers.
MARK I OMIERS PERSPt.uiVE - m. LOGUE, (MARK I CWERS GROUP CHAIR PHIIADELPHIA POWER & ELi.uxIC COMPANY f
Mr. Logue, Chairman of the Mark I owners Group, reviewed briefly the status the Mark I containment program and the efforts taken by the Mark I Owners He Group to resolve several of the concerns expressed by the NRC Staff.
stated that Part A of the Mark I LIR was issued in December I
Since the issuance which includes some revisions 'es issued in March 1979.
of Part B of the LER, the Mark i Owners Group has met with the NRC Sta several times to resolve the dissenting technical views between the NR Subsequent to the issuance of NRC Staff's draft and the Mark I owners Group.
m e M
6 e
+6 m
Fluid Dynamics November 16, 1979 acceptance criteria for Mark I containment program in October 1979, the Mark I owners have planned to make several modifications to their plants (Attachment I).
With regard to the !EC Staff's acceptance criteria on Mark I contai. Tent program, Mr. Iogue indicated that the Mark I owners Group has several dissenting views. They have been trying to resolve these technical differences. He believes that the additional requirernents of the NRC Staff will have significant impact on the overall cost and schedule for completion of the plant modifications.
He believes that the Mark I containment loads are conservatively defined in the Mark I I.31 and continuous additions o' conservatism are unwarranted; adoption of overly conservative critr7.ic as proposed by the NRC Staff would be counter-productive to the empletion of Mark I containment program. A stanmary of Mark I Owners Position is included in Attachment I, page 3.
SUBCOMMI'ITEE REMARKS Subsequent to hearing the scheduled presentations from the Mark I owners Group and the NRC Staff, Dr. Plesset solicited cements from the Sub-committee and its consultants.
Dr. Bush suggested that clear idenuzication of the symbols and acronyms used in the Mark I containment program reports would be helpful.
Mr. Grimes stated that they plan to include identification of all the acronyms and symbols pertinent to Mark I containment p-ngran in a separate NlAEG document.
Indicating that several statistical studies show that the probability of occurrence of the IBA break is about two to three orders of magnitude less than that of the intermediate break, Dr. Bush suggested that serious consi-deration should be given to the relative probabilities of break sizes in analyzing the overall load situation.
1 y
' ~ ~ ' ' ~ -
Fluid Dynamics November 16, 1979 Dr. Plesset stated that the Fluid Dynamics Subemmittee will give its report on the Mark I containment larg-term program to the ACRS full Committee in the near future.
Tae meeting was adjourned at 4:45 p.m.
NME: For additional details, a complete transcript of the meeting is available in the NRC Public Doctanent Room,1717 H St., N.W.,
Washington, D.C. 20555, or from Ace-Federal Reporters, Inc.,
444 North Capital Street, N.W., Washington, D.C.
1 p
.....~... -
OCS lhareau of Land Mana emeni.
published Octob r 4 2W9 (44 FR L
D Federst Consu, wncy Isme Department of the Intsmr. Wcshineton, etn7s),
f 10CS's.nde Act trnoument: tion (D) CCS Lasse 5.4ts tJpdats D.C.or de cpown.ite km oi Land In accordanc3 with pivM-.
t Management OCS r teld O!! ices.
outlined in the Federal Resister. Oct.1 b
7 g
gy Avadability of the minntes will be eight 1979 64 FR mal cral or written I
(33 OCS Advisory hard (Pfenary Sessient weeks after the meetmg.
statements may be presented by I
(A)OCS and the r. :en3 Eserry Cnals Desed: October 25.19 9.
members of the public, recordings will f
(BIRate of the OCS Advtsory Board Ales D. Powers.
be permitted only durtng those portions i
(C) Mexican Od spill Direeser. Offierof OCK Progmer of the meeting when a transcnot is being h,
anamoon contace Alan Powan,ad Caenimecos-bept, and questfons may be asked only g
(315ckaufle comminee Fu oma.mame ru=8 ***d* *as==8 by members of the Subcommittee. its (A) Use of Envirnver:tsi Studie, 8EA8'8 C888 88*'* "
consultants. and Staff. persons desinng
[
alonnauon m the CCS Least9e Pme.sa to mais oral statements should nonfy i
- (B) Roles of the Scientific Comr attee the Designated Federal Emplovee as far i
(C1 Relationship t:etween " Effects Studies INTERNAT10NAt. CEVT OMtENT ied Socio-Econorme Evsivanons COOPERATION AGETCY in advance as practicable so that appropriate arrangements can be made (D) Coorges Bank Bioict: cal Task Form yaIrinfonnanon coat:2 Piet deWitt.:tzt/
Agency WInternadonal Cepent Mw 6 comy W hg b i
pg meeting for such statements.
l (4} North and.%Gd.Atisatic Technical Prtvecy Act of 1974. Systen's of The agenda for sub(ect meeting shall t
7terkins Croves Records; Annust Put!!catton be as foHows-se Relating to OCS Ci! and Ces AcsMcT: Agency for Interaational Development. internauenal Fdde Naae M 1M9-
[",,
""i"fs Development Coopersuon Agency.
A:17a.m. Ucalthe Canch.mr of L
Cmunds harCCS C
s (C) FT.81 RegocalStudies Ptar Action: Systems of Records: Annual Business 6
)
yee informanos cor. tact Dick Wildersnamn.
Pubucadon.
The Subcommittee may meet la
{
The Privacy Act of19 4 (S U.S.C.552a Executive Session. with any ofits t
ue ud CdfTedimcal Marking Croups (e)(4)) requi' s agencies to cublish consultants who may be present. to 9
(A) Proposed FY 81 RettonalStudies Plan, annually m tse Feceral Register a notica explore and exchange their preliminary h)
(B) Review of the Proposed CY 198o IPp of the existence and character cf their opmions regarcing matters which should systems cf records.The Agancy for be considered durmg the meetmg and to E
gsttvtues (C1Roleof the RegionalTechnicalWorking International Development last formulate a report and 1
l nospe poblished the full text of its systems of recommendauens to the full Committee.
L' Forinformation contacc Sydney Venades, records at 4 FR 47371. Sentember 20-At the conclusion of the Enecutive ousas-mt.)
1977.No further changes have occurred.
Session. the Subcommittee will hear I'
effect as pub sh presentations by and hold discussions j
(Al se St d es P! :
(31 Regional Wortung Group Schedule fo' "P
41su The full text of the Agency for the Maru" Owners Cmp. Se Geral di International Development systems of a
Per Infonmatimi metace Esa Aronson. 213/
rarnrds also appears in Privacy Act Elecic Company, and their consultants.
a?e a$e =eo age Thf v e
e ord d on. ma necmary for ng gfr Q Sale Notice Overnew for Seaufort Sea through the Supennterdent of the Subcommittee to hold cne or more leis Notics for Laese Sale Documents. U.S. Covernr: ent P-mting closed sessions for the purpose of g
(Cl Recap of Call for Nominations for Lease OfSce. Washington. D.C. 2402. The explonng matters involving proprietary g
! ales c37 and =10 price of this volume is $10.25.
Informanon. I have determmed. In I
ForInformation contact Cordy Erler.907/
Jassee 1. Harper.
eccordance with Subsection IC(d) of
.T 7 Ness )
prfvocyAct CWicer. Ofice a/PublicA/ fairs.
Pub. L 92-453. that. should such
- j De testing is open to the public.
tru oma mamme rums te.es ru e.e.=
sessions be required. It is necesssary to I
those persons aho are interested may saa.sme caos ermeus close these sessions to protect nake oral or wntten statements to a proprietary information (5 U.S.C.
- ommittee. Such recuests should be 55:b(c)(41).
. I nade to the contact listed for each NUCt. EAR RECULATORY Further information retarding topics ierticular committee. Requests should CCMMISSION to be discussed. whether the meeting
- i. i o mad 2 no tater than November:3.
has been cancelled or rescheduled, the i
ory Cemitta on %cmr inu!:s for the OCS Pclicy Safeguards Subcommittee un Fluid Chairmar. s ruling on requests for the
- l lcmmittee and the OCS Advisory Board Dynamics; hung opportunity to present oral statements i
and the time allotted therefor can be Menary Session will be svailable for De ACRS Subcomruttee on Fluid obtained by a prepaid telephone call to rublic inspecnon and coevm*: at the Dynamics will hold a meetme on the Desiimated Federal Employee for i
2f! Ice of OCS !%: ram Coned: nation.
November 18,in, at the carrett Hotel
- this me etmtt. Dr. Andrew I. Dates.
~
toom 5t50. De:artment of the Intenor.
501 Post Street. San Franciaco. CA to (telephone 200/034-EG7) butween 8:15 Vashington. U C.
continue its review of tecs related to Minutes t'or the Scientific Committee the DWR h! ark 1 Contai.c ent Lon t.
"""" *"d # '"
P
{
ad thi Repunal Tet.hmcal Workmg Term Program and the hCR Acceptance
)
1 b
lY
~~
~
ERS SUBCCNMITTEE MEETING ON FLUID DYNAMICS NOVI!NBER 16, 1979 SAN FRANCISCO, CALIFORNIA Attendees List ACRS NRC M. Plesset, Chairman C. Brennen, Consultant H. Etherington, Member C. I. Grimes, DCR I. Catton, Consultant S. H. Hanauer
- 2. Zudans, Consultant E. G. Adensam S. Bush, Consultant J. R. Fair, DOR V. Schrock, Consultant R. L. Cudlin A. Bates, Staff
- R. Kosson, Consultant S. Duraiswamy, Staff MIT
- Designated Federal Emcloyee A. A. Sonin (NRC Consultant)
GE-BNL V. S. Tashjian A. MJkherjee J. D. Ranlet R. M. Nelson J. R. Lehner (NRC Consultant)
G. E. Wade PHIIA ELEC CO J. S. Gay L. J. Sobon B. W. Smith R. H. Logue L. D. Steiner T. J. %21 ford
'!OSHIBA CO B. Kohrs J. Torbeck Y. Takizawa t
U. Saxena BECNTEL ACUREX R. Broman W. S. Kennedy (GE Consultant)
DETROIT EDISCN 1
D. F. Lehnert NtfrECH A. F. Deardorff G. R. Ederds ATTAOMENT B
ACAS FLUID DYNAMICS SUBCOMMITTEE BARRETT NOTOR BOTEL - SAN TRANCISCO NOVEMBER 16, 1979 8:30AM - ACRS Opening Comments 8:45AM - Mark I Long Term Program Status (NRC) 0i Wra *5 TA J Cip u 9:00AM - Pool Swell Testing Programs & Load Definition Methodology (CE/Mk I owners) 9:30AM - Net Vertical Pressure Load Data Comparisons (NRC) R a n l e t-10:00AM - Pool Swell Flow Distribution Effects (NRC) kosj e B:n kenne4 10:30AM - Vent Header Deflector Load Definition (GE/Mk I Owner.f) 11:15AM - Deflector Load Data Assessment (NRC)p,gon~3 Da tor 66cn 12:00PM - LUNCH Rdp g ga g,g g UMG H $ttn & N A L
I 1:00PM - FSTF Torus Shell Pressures & Load Definition Bases
.((GE/Mk I h ers)
-}
2:00PM - Torus Shell Condensation Load Assessment (NRC) )r. Br ea a en agow n,4 )
F R4-23 2:45PM - Downconer Condensation Load Definition gl(CE/Mk I Owners) f 3:30PM - Downcomer Condensation Load Assessment (NRC)
G rIe** 3 I(GE/Mk I Owners) 4:15PM - Summary
-)
5:00PM o Adjourn
~
l t
i ATTAC4NdNT C
.. n.. - wma + :...:..,.....
.o
._ _ l _....
1/4 Scale 2D GE 2.5 o 1/12 Scale 30 EP,R:
y.._ _ f _.,.
~
..u.
2.0 r
- 1 M._ _
4[--
g, h
xs 3
+ // -
-A r-f, -
E N
f' 1.5
.MX I P:tw '_-'-i y
_l__-
5 l-RN3E 6.
.e b
J
_/_
s 1-3 u
o 1.0
~~
=
.m.
E
'~
\\
0.5
_._. i 0
20 40 60 80 100
!Y well Pressurf:stien Rata, psi /sec
/
COMFiRISCN OF PEAK UPPORCES BETWED 1/4 SCALE AND 1/12 SCALE VST - 12 11/16/79 A TTMCNN & NT D
- i I
m m
-4>--I
EFFECT OF STRUCTURAL OSCILLATION ON LIVE.V.0RE 3D/2D UPLOAD RATIO LIVEV.0RE 3D
'- LIVE.Y.0RE 2D
ACCELERATION-CORRECTD LIVE.EORE 3D 10 Y
l 5
/
l
/f.
0 g,
y 1
2.'
~
T
-5 i
w 5
/
t II
)
e l
10 3
'j
-13 l
-20
-2h L
r 2.9 3.0 3.1 TIME-SECONDS VST - 17 11/16/70
.D-%
LIVERMORE 2D-3D FACILITY INTERACTION s
COUPLED DRWELL ENSURES C0F30N DRIVING CONDITIONS COUPLEJ DRWELL PEPMITS 2D-3D FACILITY INTERACTION e
peywrLL s
t
=
SantL LARst FAclLITY FAc;Liry CONTROL OF INITIAL CONDITIONS EXTREMELY IMPORTANT e
l e
LARGE FACILITY WILL CONTROL DRYWELL PRESSURE e
SMALL FACILITY PHENOMENA CAN BE AFFECTED VST - 16 11/16/79 3 -3
^
m DOHNC0HER TYPES Nundier of Plant J pe Ikunconers
!!rowns ferry 1, 2, 3 IV 96 f.-~~
I'
~
i I 45 Brunswick 1 & 2 II 96 22 30
,/
Cooper Station II 80 30
! /
\\
Dresden 2 & 3 11 96
-p.
7 i
Duane Arnold 111 48 Fenni 2 11 80
~
~
Fi tzpatrick II 96 TYPE - 1 TYPE - Il llatcli 1 & 2 II 80 llope Creek 1 & 2 II 80 Mills tone II 96 l
~
'H
~4 5" F
ikin ticello 11 96 Nine Mile Point 1 1
120
)
x Oyster Creek 1 I
120 Peacli Bottom 2 & 3 II 96 Pil9 rim II 96 TYPE - III TYPE - IV Quad Cities 1 & 2 11 96 Vermot Yankee 11 96 6
t.
pi4
!!jillll[jjj,jlj
.jj;; pj:llj:yj 7f/l i
ir !!!!!Slllllll!i
!jlllllllllllllllllllljji
.i
O LLL 3-0 1/5-Scale Peach Bottom Vent Orf fice
(
9 LLL 2-0 1/5-Scale Peach Bottom Vent Orifice C
GE 2-0 1/4-Scale Peacn Bottom Split Orifice O
EPRI 3-0 1/12-Scale Browns Ferry Solit Orifice O GE 2-0 1/4-Scale Browns Ferry Split Orifice l
AP = 0 Values adjusted to 4' Submergence l a 6
0 12 O
O g
E I
E g4 O.
(
g 10 5
8 O
5 a
O e
i 8 1 2
6 4
t i
i l
I 40 50 60 70 80 90 DRYWELL PRESSURIZATION RATE (PSI /SEC)
FIGURE 4.
Full-Scale Equivalent Download Pressure as a Function of Drywell Pressurization Rate (Zero AP, a ft. Submergence) i D-F
/
a
[
t i
UNCERTAINTIES ASSOCIATED WITH t
DOWNCOMER ORIFICES 1)
FLOW CALIBRATIONS WERE DONE " DRY", WITH UNIFORM i
EXIT PRESSURE AT ALL DOWNCOMERS.
DURING EARLY BUBBLE GROWTH, BUBBLE PRESSURE CAN VARY FRCM ONE DOWNCOMER TO THE NEXT.
2)
DOWNCOMER PAIR #3, WHICH HAS THE LOWEST FLOW RE-SISTANCE, HAS THE SMALLEST POOL AREA AND THE HIGH-EST BUBBLE PRESSURE DURING EARLY BUBBLE GROWTH.
3)
"T" LOSSES WITHIN VENT SYSTEM VARY WITH FLOW SPLIT
(
AMONG DCWNCOMER PAIRS.
4)
ANALYTICAL CALCULATIONS INDICATE MORE UNIFORM FLOW RESISTANCE WHEN INDIVIDUAL DOWNCOMER FLOWS (DUE TO DIFFERENCES IN BUBBLE PRESSURE) ARE MORE UNIFORM.
1 RLK/5 11/16/79 i
BROCKHAVEN NAT'ONAL LABORATORY]);)]
(
A5500ATED UNIVERSIT!E5, INC.O U 3 f
A77Ackuserr E
~
g
4
^%
u N
1 1
O 5
- l'***,asijfy.
e i'!I
[
i f
L,+
/
e 3
i Ybb !
=
k f g>gj s
5
\\h \\
+
m s
1
{ Il i
=w A rra cnuar<r F
F-l
Vent Header Ocwncomer
/
\\
0 I
l/
Vent Header Deflector Typical Vent Header Deflector lvl (g>
a) Pipe (Type 1) b) Pipe with Angles (Type 2) c) Pipe with Tees d) Wedge (Type 4)
(Type 3) l 1
....._....__-__n_.,__.__,_n..e...._.,..,.
4 RANGE OF PARAMETERS i
INFLUENCING DEFLECTOR LOADS (FULL SCALE VALUES)
DEFLECTOR LOADS MEASURED IN QSTF REMAINING PLANTS FOR WHICH (6 PLANTS - 12 CONFIGURAT. IONS)
DATA IS NOT AVAILABLE (7 PLANTS 1)
CLEARANCE (IN) 0 - 21.05 0 - 14.29 (DISTANCE FROM BOTTOM 0F DEFLECTOR TO WATER SURFACE) 2)
DEFLECTOR WIDTH (IN) 25.3 - 30.0 20.0 - 26.0 3) i'(PSI /SEC) 46.1 - 74.0 54.4 - 74.7 4)
DOWNCOMER SUBMERGENCE (FT) 3.0 -
4.25 3.33-4.4 A ACUREX
?
/C Corporation u
F0
LDR' DEFLECTOR LOAD PREDICTION METHODOLOGY II)
LOAD PREDICTION e
LOAD CONSISTS OF IMPACT, ACCELERATION DRAG, B00YANCY AND " STEADY" DRAG e
IMPACT AND STEADY DRAG CALCULATED BY:
D1=Cg(Y)Ag
.j WHERE CD (Y) = IMPACT & " STEADY" DRAG COEFFICIENT AS A FUNCTION OF DEFLECTOR IMMERSION DEPTH, Y.
A = DEFLECTOR PROJECTED AREA
'./
4=DYNAMICPRESSUREOFWATERSURFACE=4fv"
[j
'9
'!j e
ACCELERATION DRAG & BOUYANCY CALCULATED BY:
l '.
D2 " (M (Y) + Mg (Y))V + MD(Y)p-H WHERE M (Y) =
IlYDR0 DYNAMIC MASS OF DEFLECTOR AS A FUNCTION OF Y g
M (v) =
DISPLACED WATER MASS OF DEFLECTOR AS A FUNCTION OF Y g
ACCELERATION OF WATER SURFACE v
=
A ACUREX n
/C Corporation i
-9 F-4
I 11J e
5-F 00 00 4
"y d
0 0
a e
0 t
1 g s 1
a "
3 r
D d n
\\
d n
a 0
e o
0 t
i t
y-a t
c 0
l a
a )s 6
u r
ia 3
g c
e n r l
l Id a
e
\\*
C c
[
0 c
0 A
l d
0 a
e 4
t r
3 o
u l
s ae H
0 0
d
)
0 c 2
e 3
sm
(
e A,
0 m 0
i
{.
O e,
l 00 3
00 s*
0 A
./
1 12 0
0 96 2
0
/,
0 0
4 2
00 00 0
0 2 0
0 0
0 0
0 0
0 0
0 4
0 6
2 4
0 2
2 1
1
- 5. 2 C 3 m.N
- i l
i l
lI
, i
6
-F 00 06 4
00 0
\\
4 2
4 esa 0
C 0
0 s
"y q
2 d
d g d 4
a e
a a
o t
r e
L a
D t
l S
0 r
u n
o c
o 0
t l
i d
0 c
_\\
a t
n 0
e C
d a
a 4
l r
.e f
\\
r e
t e
l u
l c
D a
s e
ag 0
t a
c pa t
0 n
o e
c mr
)
T A
iD 0:
c e
l I
T-e V
1 e
3 sm f
/ [ [
_e l
(
S e
Q 0 m d
0 i
e 1
0 t
6 a
3 lu c
&y.[-
j l
0 a
C 0
0 d
4 n
3 a
s de 0
r u
/
a 0
s 0
e 2
Yf 3
M 0
0 00 3
00 0
0 0
00
=
0 0
02 0
0 0
0 0
0 4
0 6
2 0
4 0
2 2
1 i
m 5. 2 = E2 Yeh
- l
';f' i'
.il' l
!;ll I-I i
7 F
00 0
3 2
4 e
s g
a a
C 0
r 0
D g
0 s
a "y
0 d
r D
0 a
d 4
o.
a n
1 e
o d
t i
r e
S t
o t
a 0
t 0
a r
c e
l d
e 0
u n
l l
1 f
c a
e 1
l c
3 e
a t
c D
C d
c A
e a
t r
p 0
n i
)
e l
u n
0 c
V a
s I
0 e t
a s
r o
e f-T 3
ii i
l f
i
(
S Q
ii d
y e
0 i e
l
[,-
0 I
ta 0
l 4
u 3
c la C
0
^
0 dn
)
0 a
2 3
d e
r u
0 s
0 a
e lL 0 H
0 3
00 u
0 0
0 0
0 0
l 0
0 0
0 0
02 f
8 4
2 8
4 0
2 2
1 75 bo.
u 3ed
~
COMPARIS0N OF CALCULATED AND MEASURED PEAK DEFLECTOR LOADS i
i 3
F PLANT TEST DEFLECTOR TYPE CALCULATED CLEARANCE / WATER SURFACE TO DEFLECTOR l
MEASURED (INCllES)
.t A
5 PIPE W/Ts 1.50 0.0 17A PIPE W/Ts 1.00 1.635 21 PIPE W/Ts 1.28 3.585 B
8 PIPE W/ ANGLES 1.10 5.645 0
12 PIPE W/ ANGLES 1.08 5.645 i.)l!
C 8A PIPE W/Ts -
1.31 0.54 10 PIPE W/Ts 1.09 0.54
)!
13 PIPE N/Ts 1.00 3.83 D
6B PIPE W/ ANGLES 1.93 0.575 1i E
10 PIPE N/ ANGLES 1.50 1.13 5,
15 PIPE W/ ANGLES 1.60 1.13 F
10 PIPE W/ ANGLES 1.54 1.15 AVE 1.33 A ACUREX
/C Corporation F-8 J
l l
i ii i
l i
I I
i 1
i i
n 7
T T
\\
QWNEQS 8
\\
NEC l
l e
- s
\\
k 5
'1 4;
i fa'd v
\\
'i l
l
\\.
c n ( r-)
i l
1.a.
y es 4
s O
f 05 10 S
ve.
_ I.- -
d.
/MPACT *MMSIEurf SMoy CRAG foot CnNog I
t i
1 i
I I
F~9 I
W l
^
.;/
l
- l'l...
d= 0.707 ur 5
4.95 Y
7 /,- / ) < O y / / < >
^
o 4
55 V
E
$!i ifi' %'.}
_ 3,/4 3
i l
2 l'
.1'
/\\
/
/
i
/
07
/
I
\\
g
/
j i
\\
I
\\
/
i
\\.
'\\/t
(__
o O
O.5 tO g
y o
m Q
N N
f f
.D
~
l l
i lMf:Chl$Ad ctrag frce Cc rr2lA kC'-r Sgttre. 3:
y
.N UP'E d'7b-N e
l.
F f v?.W 2
l
/,,n-a C-~
9 l)X cl.=0.4l4 V N
'l e!
/// <-,
,,//,,
8:
3.If A
2 90 i
V
~~
g l
1 I
2 1
I I
l 1
I s
o7 i
i
.\\
/
t s-r yp 1
t s
l I
\\
.2 d
e D
i k.
M.
m e
N O
7 o
I D
4 i
I I
/mpac/-lslead clmg free cor,e-fa/ic-:,,
Egure 4 :
y G
7jpe 3 a'eflec.4e-yr l
F-il
ar3, MYy1.'nw~ r*J 1 :.+ + < ~ +m s e
~~
I f
O 8,
e a
6 e
1r
-(
D
/ ef //[/*
a',
4 l
l fY $
j\\
-=
's.g V
i.
s.
~;.
F
~
dj y-f:
84' 1
3 2
~
I s
~
07 i
\\
/
\\!
2(/f.
i g
7 o
/
tc>
2 W~
G e
alp E
Kgua.s:
/nyd shad 43 pa corre&dh y
& Jype. 4 de(kche.
I 1
i
/.
=
F-t2E x
e e
i I,
l I
I I
l 5
8.
5 k
~
- 1 Al 3
- E C SRc1;Gcahory
-.Ge I
e-PWre w2dge
~
1
,pg 2-G5!!-
simuladbn of Duane Arnofd G ~O' I
, AD D,
~
c.7 GT
'G,
6 f,
D-e
'%,^
f i
O 6
0 l
g l
l Vf i,
l b
1 I
I 4.S' k'.a!-ang/e.
lmpac!
b~.2r;sien! fr.pcg. ccmpard(wiK da f
weage.in an.p<.m i
frm Nark I cy7;r.
i h
l t.
F-t3
L a
j
- j. 4
/
\\
(v l
, Dam
~.
NW
[
i g
DOweeCCteER #
% SUPPR4530*s poet af ELAVATICes vetW t new Ane onwrnt e
r 0
NAd qaja c 3
4-I dcil: CL c
en eastteet PLass viaw
, -- m b 7 T M C H M E t4 T" d l b-I
}
- s
..w M6 MMS 6 me,,
4 9
e s
h 3
b=
I, 1s==-
a 4 s 5
I,
=
IE '!
/ D g
8, 5
'Q e o C
g w_
- p ge a
9--m3 3
i
.n,l i
f'1, x
s Is.,
se s
- jH j!"e
(,3
=
g
=
w
I
- i y t $5 I
I 7 :-
~
CD I
a a
egg J
i, I
=
=
s-n.
a E
i.i.
i..!.F
!'=
s2 x
L.-
~
i.g
=
J CL u n y
5
/
\\
ml rrt
=
c=
,, uj 2
gL b
ene
[./
I t
o
--a 5
i I
I 3:
en
=
8:
!,==
. [
3 v
g gj gim z
E
.g g
E y
i oi
=
(
bu d
im
'M h
d e
uun hd
SYSTEM ~ INSTRUMENTATION DATA RECORDING CAPABILITY 0
256 CHANNELS 8
EACH CHANNEL SAMPLED AT 1000 SAMPLES /SEC PRIMARY MEASUREMENT GROUPS 8
TORUSSHELLRESPONSE(E,X,h 8
TORUS SUPPORTS STRAINS 8
DOWNCOMER BENDING MOMENTS 8
RING HEADER STRAINS AT DOWNCOMER ATTACHMENT 8
TORUS WALL PRESSURES 8
RING HEADER AND VENT PRESSURES 4
DOWNCOMER PRESSURE 0
DRYWELL PRESSURE i
8 DOWNCOMER AND RING HEADER LEVEL PROBES 8
POOL TEMPERATURE DISTRIBUTION 1
16/79 8
SYSTEM FLOW RATES g
TEST INSTRUMENT
SUMMARY
DIFFERENTIAL PRESSURE STRAIN DISPLACEMENT TEMPERATURE LEVEL ACCELERATION PRESSURE TOTAL WETWELL SHELL 26 122 16 54 6
14 222 4
4 HEADS VENT HEADER 1
28 4
33 HEADER SUPPORTS 16 16 DOWNCOMERS 13 16 16 9
-54 40 WW SUPPORTS 40 VENT DUCTS 4
4 2
10 6-INCil BLONDOWN 3
1 4
18-INCH BLOWDOWN 3
1 4
2 9
1 12 DRYWELL 3
6 STEAM VESSEL 1
2 6
6 BASEMAT TOTAL 53 222 16 71 26 33 6
427 9
bp
FSTF TEST MATRIX
SUMMARY
TEST BREAK FARAMETER NUciLER*
C0liFIGURATI0d IliVESTIGATED M1 SMALL STEAM REFERENCE IEST 112 MEDIUM STEAM BREAK SIZE INCREASED (STEAM)
R3 SMALL LIQUID 3REAK TYPE CHANGED TO LIQUID.
M4 SMALL STEAM FREESPACE PRESSURE INCREASED.
M5 SMALL STEAM Pool TEMP. INCREASED M6 SMALL STEAM SUBMERGENCE DECREASED AND Pool TEMP. INCREASED.
M9 SMALL STEAM SUBMERGENCE INCREASED.
n10 SMALL STEN 1 VENT AIR CONTENT DECREASED.
M7 LARGE STEAM EREAK SIZE INCREASED (STEAM).
M8 LARGE Licu!D BREAK SIZE INCREASED (LIQUID).
i
- IN ORDER OF PERFORMANCE G-5
}
NEDE-36539-P GE COMPANY PROPRIETARY Class III Table 6.2.1-1
SUMMARY
OF CHUGGING DATA 3ASE Test Number M1 M4 M9 M10 Ini:ial Conditions nominal 5 psig free 4.5 feet no vacuum space press.
submergence breaker
- Approximate Chugging 30-)30 26-116 25-305 20-120 Periods, Seconds 250-305 Seconds of Chugging 300 90 280 155 Data Recorded Approx 1= ace Number 670 110 480 200 of Dovncocer Chugs
)
(.
- Time = 0 is the start of data recording l
O s
l l
i
~
6.2-18 6,-b
,-w-w-
ww- -,-
m-w-
w-
w
--*--v
1 II l
)
gm4a34
[,
' h=:
M N.
08 1
I 6
0 I
7 8
gn igg 0
I 6
nu 1
lC r
o f
0 5
s I
1 no l
i 2
f t
"(
i S
d H
n 0
I 4
E o
M C
1 O
4 C
e N
r QI W
u O
t D
a o
0 r
3 E
i
.,}
8 p
1 8
e 6
T m
6 F
e O
T A
M y
2 O
l A
\\
3 0
I o
V I
2 T
o 1
0 P
8 s
E d
1,V\\ \\
q li u
T s.
\\
s u
y T
G 4
o A x
t n
N I
l u
i A
l o
2 SI l
Go H
t u
F GGn
- g Guu l
N g *-
A s
ul o a
H s
SCs s
N E
a A
lA T
0 r M
0 M N
H I
8 E
1,~ ~..
i L
3 n
A g
O 8
N s
0 O
$g\\
1
\\
m O
P 0
\\
2 A
0 0
I s
GGE 9
6 o
NmT l
I A
O GGG C e
H
\\ \\H mGGi AI r
uuo u
I G Hi N g
A e
G CC.
i
\\-
s 0
u F
.2i:
0 0
b2 t
I
)
8 U
C.EEAO l
0 PrnN GYv oT
_01 uT1MS 9
oE N T 4
3(
0 EG t x 0
1 L N
. t l 1
1 7
H BI
. A T A G G
uA I
Ou(J sM 5
4
. a HH PC 0
,6 S
y 6
6 3
2 i
v 7
$j I2 3 3=3"I ag" <
e d
PYM sin l<
I
,l
,;: j i
NEDE-26539-P Gw Glass III i
Table 6.3.1-2 DYNAMIC STRESSES DURING CONDENSATION OSCILLATION AND CHUGGING Condecsation Osci' ation Chugging (M8)
(M1)'
(osi)
(psi)
Wetvell Shell*
Wetvell Shell 3,300 2,500 Wetvell Shell/ Ring Girder 14,800 2,900 Intersection Wetvell Suceor Columns Radial Bending 1,500 300 500 300 Lensitudinal Bending Tensile /Cc=p ras sive 1,600 500 Vent Header Shell Dcuncemer/ Vent Header Intersection e " Tied" Downcemers**
14,000 e
"Ftee" Downcomers 46,000 25,000 Maxinum surface stress intensity.
l Menticella prototypical tie-straps.
l l
t cn.s 6.3-12
)
,4- >
e-T, i
3;+
M, E
1 P
P TORUS SHELL FREE SURFACE SOURCE (INPUT) r--FREE SURFACE
$\\v j
=
v v'l L
J L
LJ U
P TYPICAL RIGID DOWNCOMER
\\
(OUTPUT) h F
TORUS alcio y
(INPUT) w=0 w i r ) (OUTPUT) w (t ) (OUTPUT) l l
I FLEXIBLE TORUS RIGID TORUS FLEXIBLE TORUS FOHCING SOURCE FORCING SOURCE RIGIO WALL PRESSURE d
I APPLIED AT THE APPLIED AT Tile INPUT AT Tite WETTED
$UHFACE OF TiiE SilELL <
DOWrJCOMER DOWNCOMER 6
6 I
I 4
/
(c '\\
TORUS ANALYSIS FOR CONDENSATION OSCILLATION LOADING e
OVERALL PROCEDURE
- BASIS FOR LOAD DEFINITION IS DATA MEASURED IN FSTF
- PERIODIC LOADING.
FOURIER EXPANSION OF LOADING AND FREQUENCY BY FREQUENCY SOLUTION
- CORRECT MEASURED PRESSURES FOR FSI EFFECTS.
DEVELOP RIGID WALL LOAD DEFINITION
- APPLY RIGID WALL LOADING IN PLANT UNIQUE ANALYSIS.
INCORPORATE PLANT UNIQUE FSI IN SOLUTION
(
e DEVELOPMENT OF FSI CORRECTION CURVE
- NASTRAN MODEL OF FSTF AND CONTAINED flu!D
- ANALYSES FOR UNIT HARMONIC SOURCES AT-DOWNCOMERS.
REPEAT ANALYSIS WITH SOURCE FREQUENCY VARIED IN (APPROX 1 HZ) INCREMENTS OVER RANGE OF INTEREST
- Two SERIES OF ANALYSES.
FIRST IS FOR flu!D AND ACTUAL (FLExIsLE) STRUCTURE, AND SECOND IS FOR flu!D WITH RIGID BOUNDARY
- OUTPUT IS WALL PRESSURES.
INTEGRATE WALL PRESSURES To GET NET VERTICAL LOAD
- FSI CORRECTION CURVE IS RATIO 0F FLEX!sLE To RIGID NET VERTICAL LOAD, AS A FUNCTION OF FREQUENCY Cn -IO l
FSTF ANALYTICAL MODEL (DEVELOPED USING NASTRAN COMPUTER PROGRAM) e STRUCTURAL MODEL
- ONE HALF 0F FSTF (SYMMETRY SEGMENT)
- APPR0x 500 ELEMENTS, 500 NODES
- SHELL MODELED USING QUADRILATERAL SHELL ELEMENTS STIFFENERS-AND COLUMNS MODELED WITH BEAM ELEMENTS l
e FLUID MODEL
- CONSISTANT MASS MATRIX METHOD
- flu!D MODELED USING HEXAGONAL SOLID ELEMENTS
- flu!D ASSUMED INCOMPRESSIBLE e
LOAD APPLICATION
- SOURCE FORCING FUNCTION AT DOWNCOMERS OR
- WALL PRESSURE FORCING FUNCTION (5-Il I
VERIFICATION OF ANALYTICAL MODEL a
STATIC CHECK CASES e
COMPARISCN AGAINST SHAKE TEST RESULTS e
ABILITY TO PREDICT FSTF STRUCTURAL RESPONSE TO CONDENSATION OSCILLATION LOADING
- DATA FRcM TEST M-8, PERIce FRcM 24 To 25 SECcNos, Usan FoR VERIFICATION
- CONVERT MEASURED FLExIsLE WALL pressures To RIGID WALL pressures UsING FSI CORRECTION CURVE
- DYNAMIC STRUCTURAL ANALYSIS 3AsED ON RIsto WALL LOADING
- COMPARE PREDICTED STRUCTURAL RESPONSE QuANTITIss WITH MEASURED DATA bh -12, 9A
FJ i
l l
i ampus c.anam '. uns e! w? c
\\
i I
I h
i"i
-i i
i 6
i i
i I
I i
4 s.
i l
I f
i l
I l
l i
u l
l l
l i
l l
i i
1 i
f i
l 4
t g
I I
l l
I u
5 l
t i
l 4
i i
1 i
[
+
i i
l l
i
~
I 2
i l
3 i
l I
(
) u-
[
l l
3 l/
l l
I I
i
[
I i
I I
I l
I I
l l
u I
i l
I I
l l
l g
i
'I l
f
(
l f
u --
I 1
i A
f i
I o.
sneeumer mm 1
i RESULTS CF ANALYSIS assPLIFICAT1000 PACTCA POA TUTAL vtRTICAL PCdICS VS 74tQutMCY t "19eGIO" POACE = *PLEXISLE* 80 ACE / AssPL17tCAT10ss 7 ACTCA i 0 -13 D
j D
3 93 O
d 3-d.s _ _J_.
x
~.
x -
l f.SilLIS OF VEltlFIC.U10!l RIlllS CtlHtiLATIVE AXI Al. HEHHilANE STRESS Ai lluh lull illD-SPAll (SotiltCE AllALYSIS) a
=..-
i X
5 y--a------_-------=_-=
e 1855 DistM y n
l 5.
_.g____-
' ' ' %/
I a
3gg c
~
1 I
U s
-J--
i.
.0 n
l 8fitOutf.CV IHit Aa50tutt
- - -. t...c i
m L.
T+
,g
w LUllDI.:'I3iti 1Uit UOt eili IUlf LUnui'lu
^
FSTF RESPONSE.
.I TEST FSTF ANALYSIS (2) ton togg33 (3)
MARGIN OllANTilY DATA (1)
ALGEIWlAIC Alls 0LilIE CASE 1 CASE 2 CASE 3 EIXl CASE 2 HEST IwTA AXIAL firiusR-I ulE 1.9'l 1.80 2.22 3.55
- 11. 5 7 2.20 2.36 (KSI)
Il00P NEllilt-2 m:E 2.06 1.80 2.35 3.90 li.00 2.15 2.23 1 'd (xSi) 9. C2 ItADIAL h 'iYa'[~ ~ 0.086 0.101 0.129 .230 .275 .1'11 3.20 ~ B (irlacS) II:t; Eft CG ";8 93.3 .101 136 25'l - 290 172 3.11 4 AXl AL Fal-CE (EIPS) Otlltit '" $ rat- ]11.5 116 152 278 370 386 2.37 i CE (nips) g b I'OTES: (l) l}ATA FOR IEST Il-0 Ilt1E PERIOD 2'i.8 TO 25.9 SE Onps g (2) 1.0AD APPLIED AT NUL11Pl.ES OF.91 Itz. FRE00EliCIES 0-39 iiz C0riSIDERED. G) 1.0AD APPL. LED AT STittlCTtlRE flATURAL FREouEllCIES .'a F:tsousi:CiES 0-30 iiz cons:DERED. ABSOLUIE'SilH. ,i{ i-N.Ij
e MARK I CONDENSATION OSCILLATION APPRCACH e ENTIRE FSTF C0 DATA WAS EXAMINED e MAXIMUM PRESSURE AMPLITUDE DATA SEGMENTS WERE SELECTED AS DATA BASE e WALL PRESSURES (24 SENSORS) WERE SPATIALLY AVERAGED - AVERAGE VERTICAL PRESSURE LOADING ON THE TORUS SHELL e PSD ANALYSES WERE PERFORMED e FSTF FSI EFFECTS WERE ACCOUNTED FOR e RIGID WALL PRESSURES AS A FUNCTION OF l FREQUENCY WERE SPECIFIED AS LOAD DEFINITION / G 16 UCS - 04 11/16/79
MARK I CONDENSATION OSCILLATION DATA BASE THREE DATA SEGMENTS SELECTED ARE: TEST RUN DURATION POWER MAXIMUM M8 29-33 sec 4-5 Hz i MAXIMud M8 24-28 sEc 5-6 Hz fiAXIMUM M7 21-25 sEc 6-7 Hz ( G-I 7 UCS - 07 11/16/79
MARK I CONDENSATION OSCILLATION DATA REDUCTION / ANALYSIS FOR EACH OF THE SELECTED THREE DATA SEGMENTS... e WALL PRESSURES INTEGRATED MEASURED WALL PRESSURES (24 SENSORS) WERE SPATIALLY INTEGRATED INTEGRATED VERTICAL PRESSURE TIME HISTORY GENERATED OBTAINED TIME HISTORY REPRESENT OVERALL LOADING ON THE TORUS SHELL e POWER SPECTRAL DENSITY (PSD) CALCULATED PSD OF EACH 1-SECOND SEGMENT WAS GENERATED PSD VALUES WERE AVERAGED OVER THE FOUR SECONDS AMPLITUDE VS. FREQUENCY VALUES WERE COMPILED e FSTF FSI ACCOUNTED FOR FSI FACTOR AS A FUNCTION OF FREQUENCY OBTAINED COMPILED AMPLITUDE MULTIPLIED WITH FSI FACTOR - G -,g RIGID WALL PRESSURES UCS - 08 11/16/79
MARK I CONDENSATION OSCILLATION LOAD DEFINITION e TORUS LOADING DEFINED AS RIGID WALL PRESSURE VS, FREQUENCY e THREE ALTERNATE FREQUENCY SPECTRA, 4 TO 16 Hz, SPECIFIED e ALTERNATE SPECTRA BOUND VARIATION OF DOMINANT FREQUEt!CY WITH TIME OBSERVED DURING THE TESTS e LOAD DEFINITIONi AMPLITUDE VS. FREQUENCY A 0 - 50 HZ RANGE A INCLUDING ONE SPECTRUM 4 - 16 Hz SPATIAL DISTRIBUTION A UNIFORM AXIALLY A LINEAR ATTENUATION WITH SUBMERGENCE PLANT UNIQUE ADJUSTMENT FOR POOL-TO-VENT AREA RATIO DEFINET AMPLITUDE COMPONENTS SPECIFIED AS STEADY STATE 0 -19 LOADING UCS - 09 11/16/79
EVALUATION OF DOWNCOMER LOADS DURING CONDENSATION OSCILLATTON e STATIC VERIFICATION RUNS - JACKING BETWEEN DOWNCOMERS #5 & 6 (IEST #7) - JACKING BETWEEN DOWNCOMERS #6 & 8 (IEST #6) - JACKING BETWEEN DOWNCOMERS #7 & 8 (TEST #8) - CORRELATE ON LOAD - DEFLECTION CURVE - CORRELATE ON STRAIN GUAGES ON DOWNCOMERS AND ADJACENT HEADER (S5911-S5918, S5921-S5928) e DYNAMIC VERIFICATION RUNS - MODAL ANALYSIS TO CALCULATE DOWNCOMER " SWING" FREcuENCY - COMPARE WITH RESULTS OF DOWNCOMER "$ NAP" TEST - POSSIsLE ADJUSTMENT OF EFFECTIVE WATER MASS IN DOWNCOMER e STATIC PRESSURE RUNS - UNIT PRESSURE IN DOWNCOMER AND HEADER "Tw0 TO ONE" PRESSURE IN 00WNCOMERS AND HEADER e DYNAMIC ANALYSIS - HARMONIC ANALYSIS (5.5 Hz LOADING) "TWO TO ONE" PRESSURE IN DOWNCOMERS AND HEADER - CORRELATION WITH M-8 TEST DATA (STRAINS IN 00WNCOMER AND ADJACENT HEADER) CLOSURE POSTULATED LOAD DEFINITION EXPLAINS MEASURED STRAINS ? OR LOOK AT PHASING BETWEEN PRESSURES IN ADJACENT DOWNCOMERS AND FINALLY LOOK AT OTHER TE1TS.AND IIME PERIODS nrru nueur M H-l
TYPICAL GENERIC MODIFICATIONS TO PLANTS e T/ QUENCHERS e VENT DEFLECTORS e TORUS SADDLES e COLUMN REINFORCEMENTS e ANCHOR BOLTS e DOWNCdMERTRUNCATION AND CONTINUED USE OF DRYWELL/WETWELL AP Amcueur r 11/16/79
W SCHEDULE FOR COMPLETION OF PLANT MODIFICATIONS
- OWNER PLANT COMPLETION DATE TENNESSEE VALLEY AUTHORITY
- BROWNS FERRY 1,2,3 JUNE 1983 lAROLINA POWER & LIGHT
- BRUNSWICK 1,2 J U N E ' 1'9'81 (EBRASKA PUBLIC POWER DIST.
COOPER MAY 1980
- 0MMONWEALTH EDISON CO.
- DRESDEN 2,3 MAY 1982 OMMONWEALTH EDISON CO.
- QUAD CITIES 1, 2 FEB.1982
[0WA ELECTRIC LIGHT & POWER DUANE ARNOLD APRIL 1981 20WER AUTHORITY STATE OF N.Y. .FITZPATRICK JAN. 1983 3EORGIA POWER COMPANY
- HATCH 1,2 JAN.1983 10RTHEAST UTILITIES SERVICE CO.
MILLSTONE APRIL 1982 10RTHERN STATES POWER MONTICELLO FEB. 1980 IIAGARA MOHAWK POWER CO. NINE MILE PT. JUNE 1981 JERSEY CENTRAL POWER & LIGHT OYSTER CREEK DEC.1980 )HILADELPHIA ELECTRIC CO.
- PEACH BOTTOM 2,3 NOV. 1981 30STON EDISON CO.
PILGRIM MARCH 1981 (ANKEE ATOMIC ELECTRIC CO. VERMONT YANKEE NOV. 1981
- AS OF MARCH 197'
- MULTI-UNIT I 05 i
i 11/16/79 JE -
SUMMARY
OF MARK I OWNER'S POSITION e CONTAINMENT LOADS MORE COMPLEX THAN ORIGINALLY ANTICIPATED e FURTHER INTERACTION ON LOADS AND STRUCTURAL METHODS REQUIRED - FUNDED THROUGH 1980 e UTILITIES PROCEEDING WITH MODIFICATIONS ON " RISK" BASIS e EXPECT INTERACTION WITH NRC ON EITHER GENERIC OR PLANT UNIQUE BASIS e 0WNERS BELIEVE CURRENT LDR GIVES PRACTICAL ENGINEERING SOLUTION e 0WNERS REQUEST CONTINUING ACRS/NRC DIALOGUE TO ASSURE BALANCED PROGRAM CLOSURE
- t-Jb 11/16/79}}