ML20214G847
| ML20214G847 | |
| Person / Time | |
|---|---|
| Site: | Columbia  |
| Issue date: | 10/03/1972 |
| From: | Boyd R US ATOMIC ENERGY COMMISSION (AEC) |
| To: | Tedesco R US ATOMIC ENERGY COMMISSION (AEC) |
| References | |
| CON-WNP-0095, CON-WNP-95 NUDOCS 8605220557 | |
| Download: ML20214G847 (25) | |
Text
.'
--_p,._....,.-..
g y.
5y
..}c m>c,,
un 4
OCT 3 B72
'r s, w ws.
' ged.h&%
{ ::;; - : g, s
- e W O G O UPd. c;.-&FT L Ji t&, :er y".,
,geghetsmi.'9BuBptt*
,u aeOO@m6 7c.' @ rc G FS18&iiG6 M $ dP<A Q 6
- ..a @ ka e :D ;
.+ a
.a ;M. wry %w ny a
Q ) f& W iA M9edeces. a.mpp=en 1mseqmmen,m.
gesagiyg gg. 3 p.;c; s
D @ G S C w r.s.:r+.w3940sM(W
O,1;;.>r d O-r Sn:ammesa;sames.+.5mmesina&
+M iw mister of emisi M.
- %ammaine me. a nseery -- -*-=e= -4. wets;
.s.annuen s e
.- y.3 W 34 of ties 904E and to assomsg 's am;mysammame es voyieur. the F
.tes to Amsetuames.,e
.~Q
.Be. 1 and 3.
1kis esapletes o
Beefood Ms. 2 primary===*W====* destga and as propers the 1eyste for the Safety Beelmaties Espect.
The estashed Emmford Be. 1 Safety Bus &setism Buyert Secties 5.0 "n==*a4====* Syeeene and Streeteral Desige gestariel" ouheestiene 5.1 thesugh 5.4 see hosed en Eersehe1's 3ecess. 'It is my under.
etanding that yes med ese 1mises have rettened this port &am of D e Set and that your esammets sure Samorpasseed.
As poet CP rorier items the appidamme hee to providing the fe13meles Auferestiam for year sorters o
A esm =s====* poet IACA hydveges' eenteel sposes that will set-1.
e infy eefety guide me. 7.
2.
A etudy of meditimari esses to mitigste the esseegneeses or ums e-ema the potentia 1 for hypees 1ad myr-The _ ;;'d--- t stated in respamme to a phoos em11 that the bypees leakene study will be eespleted sesumi Her er Jume of 1973, and l
that he espects to diesene a peepened poet lac 1 kydresen sentrel system with se in either Jessary ear Dutweary of 1973. We will
( 3 s.. enhedmie a meetles es esen espsuplieset indgg"geen he siehas y
to meet with us.
2 -- F v u t D D g - ; t.
xm+
9 G G C O t9 O m 3 $QJate MQ
[
~ a ;~
. :s t
.u a G D u:L ~ v'; Mwd
,;a %51 e
, m w s ~ y., y,
,~ p*.
~;. :: ' v QU y; y y 3:zy'..,
- p w-y
$'3Q6i e M er @ G n.2 'J G G $ @ @ QidG#y%m e is. k
-y w:9er:, & c w a ;
r.
+ @ D sanger',sigsge, h 3tsenter-f
,, ;-( c <: g >
~.Ot' e..;&W~
J: M &nGhat M Sassenes E e ? #2 C ) "8 W Edsges0mg 9%/
Q3s e -
T 42 x 5
, n* c q ~ m r 9 m.;m:C eg b]::l c
m@
A%
est A. Oh Distribattes A e-J. Emedrt*
Docket File R. S. Boyd C 1'h GCR Reading H. Goerin' S. Miner i
!/ ' '" 0 Lt2WR
.L.;C)CR Jl,-QCR omer,
P 17tt)
?
pg summe,.!iM.inennb....
...RAC1,a,y
- 9. / #..../. 7. 2...
9./pf/72 9/2f_72 9 ATE >
Forum MC-lig (Rev. 9 33) AECM 02e e u a eovenesseset peuvesea cpect 197A ses -04 8 8605220S57 721003 PDR ADOCK 05000377 E
i
?
~ 5.0 CONTAINMENT SYSTEMS AND STRUCTURAL DESIGN CRITERIA 5.1 General The c'ontainment systems include the primary containment which utilizes the pressure suppression concept and the secon'dary confine-ment which is forued by the low-leakage reactor building that sur-rounds the primary containment. The reactor building has a Standby Gas Treatment System (SBCTS) to filter the primary containment leak-age p'rior to its discharge to the environment. 'Ihe primary contain-ment which consists of a drywell and a wetvell is a steel pressure vessel designed in accordance with Section III of ASME Boiler and Pressure Vessel Code. The drywell (shaped like a frustum of a cone) is separated from the wetwell (a right circular cylinder) by a reinforced concrete floor penetrated by 102 vent pipes. A compari-son of the containment design parameters for the Hanford No. 2 Nuclear Power Plant with those of Bailly and I.a Salle is presented in Table 5.1.
Both primary and secondary containments will meet the criteria for Category I seismic design.
5.2 Primary Containment The vapor suppression concept for the reduction of pressure inside primary containment following a IDCA has been used in the Hanford No. 2 design, as in other BWR facilities. The drywell is constructed above the wetwell and together they form a continuous,
TABLE 5.1 COMPARISON OF CONTAINMENT DESIGN PARAMETERS Parameter Hanford No. 2 La Salle Bailly Primary Containment Type ASNE Pressure Vessel-over Prestressed Concrete Prestressed Concrete and under Pressure Steel Lined-over and Steel Lined-over and Suppression under Pressure under Pressure Suppression Suppression Drywell Frustum of Cone upper Frustum of Cone Frustum of Cone upper portion upper portion portion Pressure Suppression Cylindrical lower portion cylindrical lower cylindrical lower portion Chamber (PSC) portion Internal design pressure (psig) 45 45 45 External design pressure (psig) 2 5
2 Internal design temp (*F) 340 340 340 e,
Maximum post-blowdown dryvell pressure (psig) 37.2 35 37 Drywell free volume (ft )
202,242
'209,300 160,800 3
PSC free volume (ft )
144,166 164,000 103,000 3
PSC water volume (f t )
108,387 12'4,000 73,500 3
Break area / total vent area
.0105 0.0105 0.012 Leakage rate % Free vol/ day 0.5%
0.5%
0.5%
Reactor Building Lower level construction Reinforced concrete; Reinforced concrete; Rcinforced concrete; Steel superstructure &
Steel superstructure Steel superstructure siding
& siding
& siding Roof construction Steel decking Steel decking Steel decking Design inleakage rate (I free 100 100 100 volume / day at -1/4" H O Vacuum) 2 p
g w
a h
e
- 61 single a:ructure for the primary containment. The steel primary con-LAjnsent sa well as the air locks and penetrations will be designed to sustain the combination of loads resulting from the loss-of-coolant cecident, the operational basis earthquake, and the conven-tional live and dead loads within the stress limits defined iti Sut.section E of the ASME Section III Nuclear Vessels Code for the norr.a1 and upset operating condition categories.
For time conbinatlott of loadings which includes those calculated to result from the lo9e+c.f-coolant arcident and the design basis earthquoke, the functional integrity of the metal containment system wil*, be assured by design wit'hin the stress limits for the emergency operatsng condition rategory of the specified Code. Stresses in the ccataituent she"1 and pet 0tration assemblies resulting from jet forces associated with the flow from the postulated rupture of piping will be limited to 90 percen; of the yield strength of the materials of construction. We have concluded tha' the design stress limits for the cetal ccntainwnt system are acceptable.
Connecting th'e drywell and wetvtl1 are 102 straight pipe vents.
The venta project a short di.tance aSove the reinforced-concrete dry-well floor and extend into che ruppression pool to provide a flow path for steam $nto tt.e water. Each sent opening is shielded by a steel deflector plate to prevent overloading any single vent by
+
direct flow from a pipe break near that particular vent. The plate
l e
i 2
~
i will also prevent foreign objects from entering the vent. The vents-c, i
are supported at the dryvell floor and in the suppression chamber by a structure between the containment vessel and the reactor pedestal to ' provide resistance against forces which will be developed during the postulated lOCA. Vacuum breakers are provided to equalize the static pressures between the suppression chamber and the _drywell and -
provide a return flow path from the suppression chamker to the drywell..
5.3 Fost-lOCA Containment Pressure The maximum calculated pressures af ter blowdown associated with a postulated double ended break of a recirculation line are 37.2 psig and 28.0 psig for the drywell and suppression chamber, respectively.
flow The pressere calculations were based on the containment vent model described in General *Electrke topical report NEDO-10320, "The General Electric Suppression Containment Analytical Model."
The use of this vent fim[ model has been accepted in many recent BWR applica-tions, starting with the Limerick planr. The AEC regulatory staff has compared the NEDO-10320 predicted peak pressures for the Limerick, Bailly, Zimmer, and Newbold Island plants to the peak pressures calculated using the AEC CONTEMPT-PS Code and found close agreement.
The maximum differential pressure across the drywell deck, using the vent clearing model described in NEDO-10320, is 18.3 psi.
i l 1 This peak deck differential pressure occurs during the first second of the design basis accident as the downcomers are cicated of water.
On the basis of our review we concluded that for design purposes a 15% pressure margin should be added to the peak calculated drywell pressure and a 30% pressure margin should be added to the calculated peak deck differential pressure. These design margins were applied to the Hanford No. 2 containment system and the following containment design p,arameters were found acceptable.
Calcul'ated Peak Design Test Containment Pressure peig 37.2 45 52 Peak Deck Differential 18.3 25 25 Preesure psig 5.4 Drywell to Suppression Chamber 14akage The drywell' deck will be designed to reduce the likelihood of bypass of blowdown steam from the drywell directly into the air chamber above the suppression pool. Short circuiting of the pool could affect the pressure suppression capability of the containment system. Several potential bypass areas have been identified. These are the peripheral joint of the drywell floor to containment wall, joints between the downcomers and the drywell floor, and cracks in the reinforced concrete floor. In Amendment No. 6 the applicant i
stated that he is presently developing his seal design for the peri-l It consists l
pheral joint between the drywell floor to the containment.
l l
l l
L
i 4
. of a circular segnent of steel pipe, formed to an ancular diameter of approximately 85 feet which is capable of accommodating the maximum anticipateil vertical and radial differential thermal movements and resisting the maximum anticipated differential pressure in an elastic manner. Since this seal is sensitive to differential gross lateral and torsional movements, sheer lugs between the drywell floor and the containment vessel will be provided to assure that they move in unison'at' the time of a seismic occurrence. These shear lugs permit radial and vertical differential movements, but restrict torsional and horizontally lateral displacements between the floor and the containment vessel. The downcomer vents that penetrate the drywell floor will be welded to a plate that will be embedded in the center of the drywell floor slab. The applicant stated in Amendment No.12 that the drywell floor will be sealed with a plastic coating that has elastic. properties. On the basis of our review of the dry-well deck design, we have concluded that the potential for deck bypass leakage has been significantly reduced.
In order to preclude excessive upward differential pressure on j
the drywell deck as a result of dryvell cooling following a blowdown, vacuum relief valves are provided on the downcomers. During a blow-down the vacuum relief valves prevent direct bypass flow from the drywell to the suppression pool vapor space. To minimize the poten-tial for a vacuum valve being stuck open and thereby providing l
]
. drywell to vetwell leak paths we require, and in Amendment No. 9 the applicant has agreed to provide, testable vacuum valves with redun-dant position indicators for each valve. These valves will be. tested at a frequency equivalent to the testing frequency for ECCS valves.
To detect other possible leak paths the applicant has agreed to conduct drywell deck leakage tests at each operating cycle. In i
addition, prior to startup, the applicant will test the drywell floor to its design differential pressure of 25 psi by sealing the down-comers and pressurizing. We will review the detailed test program during the operating license review to assure that an acceptable program is developed.
We have concluded that as a result of the proposed design modifi-cations to the vacuum breakers and the surveillance program proposed for the drywell deck and the vacuum breakers the potential for bypass leakage has been reduced. However, in Amendment No. 12 the applicant has agreed to study additional means (i.e., smaller vacuum breakers or additional containment spray) to mitigate the consequence or minimize the potential for bypass leakage.
5.5 Primary Containment Penetrations Penetrations of the primary containment arc in accordance with current design criteria. The applicant's instrument line isolation system is designed in accordance with AEC Safety Guide No. II,
" Instrument Lines Penetrating Primary Reactor Containment" and is therefore acceptable.
1
A
. The steci containment structure is designed to accommodate negative pressures of 2 psi less than atmospheric pressure. To pre-vent the negative pressure from exceeding this value the applicant is providing vacuum relief valves between the inside of the primary containment and the reactor building atmosphere. We have concluded that this is acceptable.
5.6 Primary Containment Leakage Testing The applicant has stated that the primary reactor containment and Its components have been designed so that periodic integrated Icakage rate testing can be performed at the calculated peak pressure. Pene-trations, including personnel and equipment hatches and airlocks, and isolation valves have been designed with the capability of being individually leak tested at calculated peak pressure. We have con-cluded that the design of the containment system will perndt contain-ment leakage rate testing in compliance with proposed Appendix J
" Reactor Containment Leakage Testing for Water Cooled Power Reactors,"
to 10 CFR Part 50 and is acceptable.
5.7 Secondary Confinement The secondary confinement structure (Reactor Building) is designed to ILmit release of airborne radioactive materials and provides for a controlled release of building atmosphere so that offsite doses from the postulated design basis accidents will be less than 10 CFR Part 100
J
' guideline values. The Reactor Building will enclose the reactor and its primary cot.tainment. The Reactor Building exterior walls and superstructures up to the refueling floor are ' constructed of ' rein-forced concrete. Above the level of the refueling floor, the build-ing structure is steel frame with insulated metal siding and roof.
Joints in the superstructures' panelling will be scaled during instal-lation to assure leak tightness. Penetrations of the Reactor Building are designed to have leakage characteristics consistent with leakage requirements of the entire building. ' The applicant's design criterion is to limit inleakage to 100% of the building volume per day at a negative pressure of 1/4 indi of water while the Standby Cas Treat-ment System is operating.
S.8 Standby Gas Treatment Systems The Standby Cas Treatment System (SBGTS) consists of two parallel full capacity systems designed to meet seismic Category 1 requirements.
Each has a demister for removing excess moisture, a prefilter capabic of removing 80 to 85 percent of particulates, electric heating coils to reduce the relative humidity of the gas entering the absorber beds to less than 70 percent, a high efficiency particulate filter (HEPA) is 0.3 capable of removing 99.97 percent of particulate matter that micron or larger in size, two iodine filters (impregnated, activated carbon bed), and an additional HEPA filter identical to the one
l
)
. described above. Both standby gas treatment trains will start automatically following receipt of an appropriate signal (high drywell pressure, high radiation in the reactor building ventilation exhaust duct, reactor vessel low water level or manual activation).
The applicant has indicated that the iodine filter efficiency of each train will be 99%. To achieve this he has provided 8" of acti-vated charcoal (2-4" beds) and a gasketless, welded seam type design that eliminates the bypass of air around the charcoal bed.
The gasketless 8" bed depth (2-4" ' beds) filter design has been evaluated and while a filter efficiency of 99% is possible, we have assumed for the purposes of calculating doses,a filter efficiency of 95%. The calculated doses are reported in Section 11 below. We have concluded that the design of the SBGTS is acceptable.
l l
l 1
t f *.Ef.10.".0UTE SLIP l_j 5a a"a "'"*-
._Ir.,ere.-.ne..
' +' am"* aa-l a$a-4 ln,........<.
,,,,, ai c-7 e i.n7.. u. i.>., m si ono T o cu....<.a.a.o v i.au at ms Attached are Iferschel Soccter's co-ones nn enn R. A. Clark, Chief Gas Cooled Reactors Branch - L 11anford 2 contain ent design.
circ TO {N em. ar.a e.n.y ahiT;AJ.$
RLwAA K5 The standard c o format was used however. rince Specter's work was not conducted under try aegis I oar have not concurred. Specter's com ents represent t o (u......a niu ntwaao his own views.
'f -,
on;t
,.; 2
.*f F h0M (N.me and uret) ntuanas D. B. Vassa11o', Chief Pressurized Water Reactors Brane'- No. 1-L 4@h[ 961
- f. A1 E 7411 8/10/72 b5f 01sta scs #@ ap0 TIC 48L RtMAPn3 cro tett o - est 454
?
g e
I e
s e
a e
I se e
e Y
e o
e 3
e d 1 o
O O
e
s.
?
uc'r,%
e e
UbHTED STATES y
- [d
-k*
ATOMIC ENERGY COMMISSION W ASHINGTON. O C.
20545 s,,
hun e' Robert A. Clark, Chief, Gas Cooled Reactors Branch, L I have reviewed various aspects of the Hanford 2 primary containnent design. This review uas based on information given in sections 5 and 14 of the PSAR and answers to questions 5.2 through 5.11 in amendments
.2 and 3.
For your convenience I have divided this writeup into three sections:
, general infornation, acceptable responses given in amendments 2 and 3, and those responses that =ay need additional review.
I.
General Infornation The Hanford 2 primary containment is an,"over-under" pressure suppression design. The drywell, which has a free volute of 202,242 ft3, is connected to the suppression pool by 102 douncemers. The superession chamber has a pool water volume of 108,387 ft3 and a free volute of 144.166 ft3,
\\
Both the drywell and the suppressien chamber have internal design pres-sure of 45.0 psig. The naxinun calculated pressures after blowdown are 37.2 psig and 28.0 psig for the dryuell and supprcssion chamber, respectively. The pressure calculations were based on the G.E. con-tainacnt vent flew model described in NEDO-10320. The use of the vent flow model deneribed in NEDO-10320 has been accepted in many recent 51R applications, starting eith the Limerick plant. The AEC staff has conpared the NEDO-10320 predicted peak pressures of the Limerick, Bailly, Zimmer, and Newbold Island plants to the peak pressures cal-culated by the CONTEMPT - PS Code and found close agreement.
Based on the use of the NEDO-10320 vent flow model and a pressure margin
~
in excess of 15%, the d.rywell and suppression chamber design pressures are acceptable.
The drywell design temperature is 340*F.
This design temperature is high enough to accosmedate the temperatures that night occur if there was an isenthalpic expansion of steam in the drywell because of a steam Icak. The drywell design te=perature is therefore acceptable.
Based on figure 14.6-7 of the PSAR 'the peak post-bicudown suppression chanber temperature will be approximately 195*F. This peak tenperature is for below the suppression chamber's design temperature of 281*F and therefore the suppression chamber's design temperature is acceptable.
-~
v.
_ = _
t Robert A. Clark,
Using the vent clearing model described in NEDO-10320, the maximum differential pressure across the drywell deck is 18.3 psid. This peak deck differential pressure occurs during the first second of the design basis accident as the downcomers are cleared of water. The Hanford-2 deck differential design pressure is 25 psid and therefore
]
there is a differential pressure margin of over 36% on the deck. The AEC has required a minimum deck differential pressure margin of 30%
on all "over-under" E!R's since the Limerick plant and therefore the 2
.Hanford-2 deck design is consistent with present practice.
II.
Acceptable Responses 4
Questions 5.2, 5.5, 5.9, and 5.11 called for additional infornation.
The requested inforention has been adequately supplied and therefore these responses are acceptabic.
Responses to questions '5.3, 5.6, 5.7, and 5.10 are acceptable with the follouing explanctions. Question 5.3 discusses the design basis that establishes the height of the downcomers above the drywell floor. For Hanford-2, this height is two inches. The concern here was that come downconers near the break location might. experience flooding during a LOCA. This flooding could cause choked flou in these downcomers and raise the peak dryuc11 pressure. The applicant argues that experiments indicate that the liquid droplets produced by the bloudoun process would be mixed with the blowdown steam thereby minimizing water accucula-tion. The applicant also argues that any accumulated water on the dry-well floor that rises to the height of the downcocer entrance would soon be mixed with the steam entering the downconer because of the high steam velocities.
In addition to the arguments presented by the applicant I have utilized figure Q.5.11-1 to estimate how many downcomers could be blocked before f
the design pressure is exceeded. If 20 of the present 102 douncemers
~
were completely blocked off throughout the whole LOCA, the Hanford-2 vent arca/ break area ratio would be 76.5.
This is the same vent area /
break area ratio as the 4.0 ft2 break in figure Q.5.11-1 which has a peak 'drywell pressure of 40.0 psig. It would take a blockage of over 20% of the downcomers to raise the peak dryuell pressure up to C.c design pressure of 45.0 psig. Since choked flow ir far less se;ete than completely blocked flow, no localized flooding will cause the Hanford ? design pressure to be exceeded during a LOCA. The large vent area / break area ratio of the Hanford-2 design makes this plant 4.
rather insensitive to reasonable losses in vent area and therefore the I
present two inch downcomer height is acceptable.
i 1
-. ~, - - - - - -. - - - ~, -
Robert A. Clark
- Question 5.6 relates to the limits placed on downconer submergence.
The Hanford-2 downcorer submergence is 11'8" and is within the tested range of submergences chich included depths to 12'5".
The Hanford-2 submergence is acceptable.
The response to question 5.7 is acceptable. The purpose of this question was to more completely understand the utilization of the suppression pool water during small breaks. Based on the applicant's response and my own analysis of the data, I see no basis for D. R.
Miller's statement in his ASME paper,* "The twelve foot submergence limit should be approached uith caution giving due consideration to pool mixing for the case of a small break in the reactor system."
Recent conversations with General Electric were held on this subject.
Cencral Electric is not aware of any adverse effects that occur with small breaks and deep submergences.
The applicant has presented data that show that suppression pool temperatures af ter a small break were n6t significantly stratified, indicating good mixing. I have examined the data on pages A-48 to A-53 in KEDE-10182, "Additienal Informa, tion-Pressure Suppression Concept-Test Data Report" (G. E. Proprietary) and find additional reasons to believe that good mixing occurred in the suppression pool during small breaks.
The pool temperature in the Humboldt Bay test series was conitored in four locations; the top, the bottom, one middle position near the downcomer, and one middle position oppcsite the downcorer. The following table li'ts the maximum pool tc7perature difference s
(a 'F) between any two recording locations before and after the blowdouns:
Break Area to Maximum Pool Temperature Dif. A 'F.
~Run Vent Arca Ratio Initial Final 2
.0060' Unkn.
3 22
.0060 37 12 31
.0001 3
8 34
.0020 30 13 35
.0060 11
- 5 43
.0005 32 3
The above information indicates that, in general, the pool was signif-icantly less stratified after the blowdown than before.
- ASME 68-WA/NE-1, " Pressure Suppression Containment Design - Current State of the Art," August 1968
)
o Robert A. Clark
,4 -
Another indicator of good pool mixing is the temperature increase at the recording locations. Naning the top, right middle, lef t middle, and bottom monitoring locations as T, R, L, and B, respectively, the recorded increases in pool temperature were:
T Final - T Initial Run T
R, I
B t
22 5
44 43 50 31 27 18 18 17 34 2
29 30 20 35 17 23 26 25
'l
+1
-1 33 43 Once again good utilization of all the pool is indicated.
Because of the applicant's response and the additional data I have analyzed, the answer to question 5.7 is acceptable.
The purpose of question 5.10 was to examine the effects of different analytical models en the drywell deck differential pressure. The deck dif ferential pressure in an "over-under" BUR containment is dependent upon the transport of dryuell air into the suppression chamber air space. The longer this air transport is delayed, the larger the deck differential pressure. Figures Q.5.10-1 and Q.5.10-2 show that even under the limiting condition of no air transport, the Hanford-2 deck differential pressure would not exceed the design value of 25 psid.
This desirable situation is a consequence of Hanford-2's favorable vent area / break area ratio of 95.5.
Other "over-under" Bh'R contain-ments with smaller vent arca/ break area ratios are much more sensitive to the rate of air transport. The sensitivity of the deck differen-tial pressure to the vent area to break area ratio may be more apparent in the attached figures Here Hanford 2 with its area ratio of 95.5 is compared to an over-under design with an area ratio of 51.5 (Limerick Class).
The response to question 5.10 is acceptable. The Hanford-2 deck dif-ferential pressure design is sufficient to withstand a LOCA differentia.1
~
pressure, regardless of the manner of air transport.
Areas for Additional Review In my opinion the respcases to questions 5.4 and 5.8 are not acceptable and require additional review, i
3
l 1
i Robert A. Clark,
Question 5.4 The response to question 5.4 demonstrates the acute sensitivity of BWR containments to small sized short circuits, or bypass areas, between the drywell and the suppression chamber air space.
I consider the bypass probica most serious because:
Theprimarysystembreaksizesthatgroducetheworstcontainment a.
pressures are in the range of 0.5 ft or less. Such break sizes have a much higher probability of occurring than the traditional double-ended break of a recirculation line. For some BUR's the worst primary system break area is about the size of one of the safety relief valves. Since numerous safety relief valve mal-functions have occurred in present operating plants, this repre-sents a very practical cencern.
b.
The size of the bypass area that could result in excessive con-tainment pressures (in conjunction with a primary sys tem break) is also small, often smaller than an open uctwell-to-drywell vacuum breaker, c.
There is a possibility that a small primary system break vill it-scif produce a bypass area.
d.
Once the critical combination of a small primary system break and a bypass area are in effect, ther.e is little the plant operator can do to prevent cycr-pressurization of the containment, short of blouing down the whole primary system. Turning a small break into a major blowdown see=s inconsistent with AEC philosophy.
e.
Various recent BUR experiences have had many of the elements of the bypass problem. A recent inspection of the Dresden-2 plant found 9 out of 12 of its wetwell-to-drywell vacuum breakers in intermediate positions after mycu n3
.u...
wou o.
v.
.nc vacuum breakers in Dresden-3 were found to be in intermediate positions after opening. The Millstone 1 plant has experienced sticky vacuum breakers. The Nine (file Point-1 Plant had two vacuum breakers that did not close and required operator adjust-ments. The Monticello Plant reports that the vacuum breaker position indicator microswitches have corroded and have become sticky in the humid torus environment. The seats on a vacuum breaker in one plant vere found to be badly worn and cracked.
The seats were made of Buna-N rubber, rated for operation at temperatures up to approximately 190'F. These seats have now been replaced with a more suitable material.
1 n
L Robert A. Clark.
Failures of safety relief valves have also occurred in some Bt:Rs pressurizing the drywells. As stated before, safety valves are in the worst size range for bypass accidents..These failures of ten were initiated by normal operating transients.
J Another important consideration is the possibility that small sized breaks 2
(.5 ft or less) may create their own bypass areas. As discussed in my comments on question 5.8, it is not clear that the downcomer water oscil-lations and vacuum breaker action observed in the Bodega Bay tests would not occur in a full scale design. If a small break causes uater oscilla-tions and vacuum breaker action, it is possible that the rapid opening and closing of the vacuum breakers will cause one of them to stick open.
At that moment you uould have a small break and a bypass area and the pressure would begin to rise rapidly.
There are other ways by thich a small break might create a bypass area.
Assume that there is a f aulty safety relief valve that opens at pressures considerab1'y below its pressure set point..mni assume the following sequence of events:
1.
A normal primary system pressure transient occurs, such as one due to a turbine trip.
2.
The faulty safety valve opens, and begins to blow down steam into the drywell.
3.
This steam pushes the drywell air over into the suppression chamber and compresses it.
4.
The primary system pressure now decreases and the faulty safety valve i
rescats.
5.
The steam in the drywell begins to condense on the cold drywell sur-faces with a corresponding drop in drywell pressure.
6.
When the drywell pressure drops 0.5 psi below the preccure of the compressed air in the suppression chamber, all the wetwell-to-drywell vacuum breakers open.
7.
Assume one of these vacuum breakers sticks open.
8.
The faulty safety valve then pops a second time and we then have a bypass problem.
It should be pointed out that any primary system break will create com-3 pressed air in the suppression chamber and subsequent opening of the i
a Robert A. Clark.
~
vacuum breakers when the blowdown stops. Any restart of the blowdown requires that none of the opened vacuum breakers remain open.
Another possibic sequence of events is:
2 1.
A small primary system break about.05 to.10 ft in size occurs.
2.
This break causes the dryuell air to be compressed in the suppression
. chamber.
3.
Some time later the dryuell sprays are initiated. Since the drywell sprays can handle break sizes up to about 0.1 f t2, the sprays begin to reduce the drywell pressure.
4.
The vacuum breakers open when,the drywell pressure drops 0.5 psi below.the suppression chamber pressure.
5.
One of the vacuum breakers sticks open.
i 6.
Now assume that the geall break opens ty) further, to a size somewhat greater than 0.10 ft 7.
The increased size is more than the capacity of the drywell sprays and we now have the bypass problem.
Generalizing, any time there is a varying or intermittent energy addition rate to the drywell the vacuum breakers may open. If this varying source is ever combined with faulty vaccum breaker, the bypass problem may occur.
Referring to the Hanford-2 response, figurc Q.5.4-1 shows that a combina-tion of a 0.1 f t2 break and an orifice about 5 inches in diameter would result in obtaining the 45 psig design pressure in about 15 minutes.
The applicenc does not define the specific nature of the corrective cetion that terminttes this pressure rise. He suggests the use of sprays, eliminating t'eck leakage, or primary system depressurization.
Drywell spraya are normally locked out of operation.pnd it is usually assumed that it takes ten minutes for manual operation to initiate the sprays. The drywell spray capacity is quite limited and would only be effective in break sizes Icss than about 0.1 ft2 Larger primary break sizes than about 0.1 ft2 would be essentially unaffected by drywell spray operation. Consequently, drywell sprays do not solve the bypass problem.
The applicant suggests that one alternative to limit this accident would be to clininate the source of deck Icakage. How this is done in the
A Robert A. Clark
- course of an accident is not spelled out.. It has been suggested elsewhere that positive closure devices be used on the vacuum breakers. This assumes that the bypass area is in the vacuum breaker and provides no allowance for other paths. The other alternative of depressurization converts a small break into a large one.
Figure Q.5.4-1 is an optimistic presentation. The Hanford-2 bypass analysis was based upon having a 6 psi pressure dif ference between the drywell and the suppression chamber air space. This pressure difference is enough to expell the water f rom the 11'8" deep douncomers. Conse-quently, mos t of the blowdown steam went down the vents and was condensed in the pool.
If one of Hanford-2's 20 inch diameter vacuum breakers was fully opened 2
and a primary system break of 0.072 f t occurred (doubled ended break of a 2.5 inch line), the do'rncomers would never be cleared of its water.
Under these conditions it would qnly take 163 seconds for Hanford-2 to obtain its design pressure of 45 psig. The pressure vould continue to rise at a rate of 16.5 psi per minute. -If the drywell sprays were initiated ten minutes after the start of this accident, the drywell pressure would be up to 165 psig.
Various EUR applicants have taken steps to reduce the likelihood of the occurrence of the bypass probica. The Hanford-2 design routes the safety valve bloudouns directly to the pool, rather than to the dryacll. This clininates one type of primary system break and is a design improvement.
Other applicants have agreed to place indicator lights on the vacuum breakers to monitor their position. If a vacuum breaker is observed to be open, the plant operator vill exercise the valve in an attempt to close it.
Failing this, he will shut the plant down according to the tech specs. Additionally, the vacuum breakers will be exercised on a frequent basis and leak tests will be made on the entire vent system at the end of each operating cycle.
All of the above are valuable means of reducing the likelihood of a bypass problea. I ru;oinmend that they oc considered for !!anford-2. !!:vever, in my opinion, these precautionary efforts are insufficient, for once a small break-bypass area combination is in ef fect,there are no desirable recourses for the plant operator. Based on present operating experiences, the fact that both the critical primary sys tem break area and the bypass area are small, that small breaks themselves may create their own bypass areas, 1 consider this accident more likely than the present design basis accident and as serious. The ACC requires engineered safety features to mitigate the consequences of the highly improbabic design basis accident. The AEC chould, therefore, require commensurate engineered safety features to mitigate the consequences of this more probable occurrence.
.J Robert A. Clark,
I recommend that the applicant describe What systems, such as an automatic suppression chamber spray system, would be required to give adequate pro-tection; based on the assunption that bypass areas will exist.
Question 5.8 Question 5.8 is concerned with the sizing basis and the use of the wetuell to-drywell vacuum breakers. There are two parts to the applicant's response to question 5.8.
The first part describes the analysis of the upward dif-
.ferential pressure on the drywell deck that would occur if a steam filled drywell were rapidly depressurized by initiating the drywell sprays.
Details of the analysis are absent. The applicant calculates that a 53%
margin exists between the 4.2 psi maximum upward dif ferential pressure and the design value of 6.4 psi. Although the assumptions of 32'F spray water temperature, 100% spray effectiveness, and utilization of both con-tainment sprays, all tend to maxipize,this upward force, the analysis is based on full utilization of all four vacuum breakers. What margin, if any, would be availabic if one of the vacuum breakers was inoperative?
An initial estinate of the upward dif ferential pressure with three vacuum breakers is 7.5 psi, which exceeds the design value.
I recommend that this question be pursued further to determine the sen-sitivity of the upward differential pressure to various parameters includ-ing the vacuum breaker flow area. I further recommend that the AEC staff develop a position on the method of analysis and margin requirements for the upward deck dif ferential pressure.
Another problem related to the sizing of the wetwell-to-drywell vacuum breakers is the prevention of buckling of the primary containment due to excessive external loadings. The drywell has an external design pressure of 2 psig. If the total pressure (the sum of the air partial pressure and steam partial pressure) in the drywell drops below 13 psia, the external loading vill exceed the 2 psi design limit.
p If a small atcan leah purge the dryecil of its cir, the dryuell and wetwell will have a total pressure of about 45 psia. Under these condi-tions, the drywell total pressure will be almost entirely due to the steam i
partial pressure. The reverse situation is, true in the suppression cham-ber air space. The total pressure in the suppression chamber air space is almost entirely due to the air partial pressure.
Initiation of the drywell sprays at the rated 7950 CPM will rapidly reduce the dryvell steam partial pressure. In about 123 seconds the steam par-tial pressure in the drywell vill reduced from 45 psia to about 1 psia.
During this transient the vacuum breakers will open and air will flow l
i t
t 9
J Robert A. Clark back from 'the suppression chamber to the drywell. Enough air must flow back such that at all times the sum of the steam and air partial pres-sures in the dryuell equals at least 13 psia.
As in the case of the analysis of the upward deck differential pressure, the relationship between the external loading on the drywell and the vacuum breaker flow area should be examined.
Question 5.8 also included an analysis of water level oscillations that might occur in the Hanford-2 downconers due to small breaks. These oscillations are closely related to an important safety matter - the operation of the wetwell-to-drywell vacuum breakers.
(See the discussion of question 5.4.)
Downcomer water level oscillations were observed in certain small break tests in the Bodega Bay pressure suppression experi-ments. These oscillations were accompanied by a frequent opening and closing of the vacuum breaker. If small breaks produce a frequent open-ing and closing of the vacuum breakers in a full scale plant such as Hanford-2, the potential for a creation of a bypar.s area between the drywell and uctuell is greatly increased. It is therefore important to determine if the events observed in the test series would be applicable to a full scale plant.
I have examined the applicant's method of analysis used to determine the maximum height to which vater might rise in a downconer during a small break.
I disagree with the mathetatical model used and its conclusiuns.
The applicant has used a steady state model to determine the height to which water night rise in a downcomer during a small break. The appli-cant's model infers that this maximum steady water height would exceed the water heights that occur during the observed transient oscillations.
The applicant has not compared this model to experimental measurements and has not calculated whether or not the vacuum breakers would open at the maximum water height.
I have utilized data on Bodega test #35 found in NEDE-10182 to check the applicar.t's nadel..Dodega Lcst "35 was selected because it was a small break [ BREAK AnrA
=.000156] and water oscillations were observed in
[ VENT AREA this test.
Bodega test #35 had a submergence of four feet and water surges to 5 felt above the bottom of the downcomer were recorded. Using measured values whenever possible, substitution into the applicant's equations results in a heat transfer coefficient of 600000 2,.7 required to match the test surge level.' Actual measurements of condensing heat transfer coeffi-cient for pure steam and water are in the range of 10,000 to 300,000 BTU's
- F.
HR-FT 1
i
...j Robert A. Clark The applicant's model therefore, appears to be incorrect since it requires condensing heat transfer rates f ar in excess of known values. It appears that the applicant's steady state model does not adequately describe this transient event.
I believe that the observed oscillations were caused by cold pool water splashed up into the downcomer.
Only small amounts of the pool wat r are required to rapidly reduce the pressure in the vent. For the Bodega configuration, 1 css than five pounds of pool water are needed to drop the pressure by 3 psi in the vent volume between the pool and the vacuum breaker. A 3 psi pressure drop in this part of the vent uould be enough to open the vacuum breaker.
U If a condensing heat transfer coefficient of 50,000 2,.7 H
is assumed, this 3 psi pressure drop could be accomplished in about one scConde The introduction of cold water into the bottom of the downcomer would rapidly reduce the stcan pressure in the downcomer at the steam-pool water interface to about 1-2 psia. This vould create large pressure disturbances up the vent, vould soon accelerate dryuell steam down the vent, and would cause additional pool water to be forced up into the vent by the suppression chamber back pressure.
Based on the above, it appears possibic to bring about significant pres-sure surges in the downcomers after the occurrence of a small primary system break. These surges may be accompanied by frequent operation of the wetwell-to-drywell vacuum breakers.
I recommend that the applicant further review the likelihood of small breaks which could produce frequent vacuum breaker operation in Hanford-2.
I further recommend that there be maximum utilization of the test data to verify analytical models.
Questions 5.4 and 5'8 are generic in nature and may require an extensive effort to resolve them. As such, I recommend that they be pursued by the Containment Systems Branch.
p*ek 9
Herschel Specter Pressurized Uater Reactors Branch #1 Directorate of Licensing ec:
(See next page) i i
f
---n-n.~
_,,,-w-.nn,
r e
12 -
Robert A. Clark cc:
S. Minct, L G. lainas, L R. Tedesco, L R.'Boyd, L J. Glynn, L S
=
4 S
S S
e*
t e
i i
e 0
[.
.i
- ~~
~
e V0tl T 61*E 8\\
/
= 51. 5
%ttss sees yo -
30
\\
i
. D '!CN CA$l$ LO510F CC3LAhT ACCf Di!IT 4-CEllGN VALUE N514 T Ace n
- q5,5 g
N
/v.a.c.w, twe n e.
uimu._ 2 PEAK OCCUR $ DUAthG VENTFLCMr *a PIAM OCOUR$ DUm*4G VENT C'.EAati G y
e b.
o h
to Ox<
M He cc at00s AIM JTICit l
i
_ J_ J,. t
)___
,d
' ' ' t ' j,'
--J A t 4
e'a l C.02
,3.1 1.0 it.
AIR CA.*.AYOVER FRACTIO.'s (X)
WASHII;Gic.';.et,31.!C PO'. !R SUPPLY SIST 3; FEAX 0*m.'EL. DICs D:T ES t.TIAL PRESSURE iS A FR*CTIO.*1 0F A!R 0
HNif0a0110. 2
. Ct.RRYO';ER.eRACT!G.*i
.-