ML17326A148
| ML17326A148 | |
| Person / Time | |
|---|---|
| Site: | Cook  |
| Issue date: | 09/17/1999 |
| From: | Prelewicz D SCIENTECH, INC. |
| To: | |
| Shared Package | |
| ML17326A143 | List: |
| References | |
| NUDOCS 9909240209 | |
| Download: ML17326A148 (21) | |
Text
Ql JVJD LD'Dff HCI NI 6 SOUTH 616 697 5646 P.83i16 Independent Review of "Control Rod Insertion FoHowing a Cold Leg LBLOCA, D.C. Cook, Units 1 and 2" D. A. Prelewicz
. SCIENTECH, Inc.
Rockville, Maryland Introduction WCAP-15245 (proprietary) and WCAP-15246 (non-proprietary), dated May 28, 1999, present the results ofWestinghouse analysis ofcontrol rod insertion followinga cold leg large break loss ofcoolant accident (LBLOCA)at D. C. Cook, Units 1 and 2. SCIENTECH performed a third pity independent review ofthese Westinghouse reports, which support rod cluster control assembly (RCCA) insertability followingcertain design basis (DB) and leak before break (LBB) large cold leg LOCAevents combined with a seismic event. The intent ofthis third party review is to provide assurance that the conclusions of the report ate correct and in conformance with industry safety practice, as RCCA insertion credit following a cold leg LBLOCAhas not been previously pursued by Westinghouse, The. intended use of the results of the Westinghouse analysis is to credit control rod insertion for purposes ofdetermining whether there is a return to criticalityfollowingswitch to hot leg injection.
Approach SCIENTECH reviewed the technical approach, assumptions, analysis methods and results presented in WCAP-15245 to assure that there is no safety impact due to recriticality followinga cold leg LBLOCAat the D. C. Cook plant. SCIENTECH did not perform a quality assurance verification of the work supporting the subject Westinghouse topical report.
The work scope included a review of the Westinghouse methodology compared to industry practice and pertinent Babcock &Wilcox/ Framatome Technologies Inc, (B&W/FTI)LOCA analysis methodology, which relies upon the insertion of the control rods followingthe blowdown phase of a LBLOCA. B&W/FITreceived NRC approval for their LOCAevaluation model for recirculating steam generator plants in the early 1990s, Review Results SCIENTECH performed the followingreview tasks:
1.
Reviewed the approach used in the Westinghouse Topical Reports (WCAPs -15245 and -15246) and provided an independent assessment of the technical approach used. The emphasis was on technical adequacy ofthe methods used, and the regulatory (licensing/safety) aspects such as leak before break, risk informed implications, etc.
SCIENTECH did not perform an independent verification of the Westinghouse calculations.
2.
Investigated and reported on the Framatome methodology, i.e. how does this methodology compare to that used by Westinghouse.
.o~
99'09240209 9909i7 PDR ADQCK'050003i5 P
PDR I~
>~Nb 6 SUUTQ 616 697 5646 P. 84i16 Internal to the reactor vessel, control rod insertion time can be influenced by d' rt' th y iso iono' gui e tu s or o the fuel assemblies.
Westinghouse performed analysis ofupper internals guide tube loads due to combined LOCA.and seismic forces. They also calculated fuel assembly displacements and grid loads due to combined LOCA and seismic loads. These analyses show that there are no loads which result in displacements sufficient to preclude control rod insertion.
Break Sizes and Locations Considered Two LBB breaks and three DB breaks were considered:
LBBBreaks:
60 in accumulator line break 98 in pressurizer surge line break DB Breaks:
144 in reactor vessel inlet nozzle break 144 in reactor vessel outlet nozzle break 5Q4 in'eactor coolant pump outlet nozzle double~nded guillotine break The break not considered is the double-ended guillotine hot leg break at the steam generalor inlet nozzle. For return to criticalityconsiderations, this break is irrelevant since the injection flow willgo to the break and willnot contribute to decreasing the boron concentration in the core.
Thus, the spectrum of breaks considered includes all of those which could result in blowdown forces being generated sufficient to preclude control rod insertion followinga cold leg LBLOCA.
The 144 in reactor vessel outlet nozzle break is actually a hot leg break, so there was no need to include this case to reach the conclusions drawn in the report.
It would seem that the double~nded guillotine break at the reactor coolant pump outlet nozzle willgenerate the maximum blowdown forces. Westinghouse, however, also included the smaller breaks at the reactor vessel inlet and outlet nozzles in their analysis, The maximum break area is smaller at these locations because the pipe motion is geometrically constrained to prevent the broken ends of the pipe from moving a sufficient distance to allow the fulldouble-ended fiow area to open. The reviewer thus coiicludes that Westinghouse has considered an appropriate spectrum ofbreak sizes and locations, comparable to that of the approved FIV/B&Wtopical report.
Calculation of Blowdown Loads Westinghouse used the MULTIFLEX3.0 computer code to calculate the blowdown loads on the reactor vessel and the reactor vessel internals, including the guide tubes and core barrel.
MULTIFLEXhas previously been used by.Westinghouse to calculate blowdown loads. The version ofMULTIFLEXused for this analysis was an improved version that was developed specifically for the WOG Baffle-Barrel-Bolt Program (BBBP). According to Westinghouse, previous BBBP analyses performed using this version of the code were accepted by NRC. A conservative, and previously accepted, 1 ms break opening time was assumed.
Therefore these calculations of LOCAblowdown loads are judged to be conservative.
Also, the blowdown loadings have been determined using a methodology which has been previously accepted by NRC.
G de Tu e valuat on At this point, the guide tubes are handled differently from the fuel assemblies.
For the guide tubes, Westinghouse has experimentally established allowable loads. When the combined loads
d V
,acr-oi-xvaw lb:D0 AEPNG - 3 SOUTH 616 697 5646 P.85r16 remain below these experimentally determined levels. control assembly insertion willnot:.be impaired. Westinghouse determined the total I.OCA loads by combining the inertial acceleration and acoustic loads calculated by MULTIFLEXwith the hydraulic cross fiowloads, i.e. drag oads, which were estimated based upon scale model tests and plant strain measurements, together with information from the MULTIFLEXand other hydraulic calculations. Ad ic c cu ations.
ynamic oad factor was apphed to account for the transient nature of the drag loading. This total 1:.OCA load was added using the square root sum of squares (SRSS) method to the peak safe shutdown earthquake (seismic) load to obtain the total load. The reviewer judges. that the. methodology used by Westinghouse to calculate the combined peak guide tube loading is conservative and consistent with industry practice.
Westinghouse compared the calculated combined peak loads to the allowable values.
Due to the differences in fuel assemblies between Unit 1 (15 x 15) and Unit 2 (1/ x 17), the allowable loads and the peak combined loads differbetween the units. For both units, the calculated peak combined load showed considerable margin to the allowable. Unit 2 showed a greater than 100%
margin for all 5 breaks. Unit 1 has a minimum margin of24% for the doublewnded guillqtine break at the main coolant pump outlet nozzle. Therefore, the maximum guide tube deflection which occurs under the combined LOCAand,seismic loading willnot prevent the control rods from inserting.
uel Assembl Evaluation
~
'uel assembly deflection and fuel assembly grid loading was then determined using a multi-step process.
First. Westinghouse calculated the core barrel, baffie and plate displacements doe to LOCA and seismic loadings using the WECANcomputer code.
Displacements of the fuel assemblies and grid impact forces are then calculated separately forLOCAand seismic loadings using the Westinghouse Commercial Nuclear Fuels Division (CNFD) methodology. Figure 8,2 of Reference 1 shows a schematic ofa typical CNFD model. The model used for D.C. Cook has four individual fuel assembly array models. The effects of seismic and LOCA induced motions are then combined using the SRSS method to obtain maximum grid impact forces.
Because of the different fuel designs in Unit 1 and Unit 2, separate calculations were performed for each unit.
Westinghouse has experimentally established allowable grid loads at the 95 percent confidence level. When peak grid loads are below the allowable then RCCA control rod insertion willnot be impaired. The methodology used is judged to be representative ofthe hest current industry practices for determining fuel assembly response to LOCA and seismic loads.
Results of the comparison ofcalculated peak combined loads to allowable values shows that there is greater than 50% safety margin. The minimum margin occurred for the double-ended guillotine break at the main coolant pump outlet nozzle for the Unit 2 intermediate flow rrdxer (IFM)grid. Allother cases had greater than 100% margin. These results demonstrate that for a design basis or LBB cold leg break, control rod insertion willnot be impaired.
Comparison.to Babcock 4 Wilcox/Framatome Technologies Methodology The Babcock 8: Wilcox/Framatome Technologies(B&W/FH) LBLOCAMethodology for recirculating steam generators, described in Reference 3, clearly takes credit for rod insertion followingall LBLOCAs. Section 43.2 4 of the Topical Report states that "For the reflood'phase.
the void moderator coefficient willbecome ineffective and both safety rod injection and borated water injection are used to maintain the reactor subcritical." B&W/FTIreceived NRC approval for their LOCA evaluation model for recirculating steam generator plants (see Reference 4).
Note 2 to Table 2.L2-1 of this SER (Reference 4) notes that fission heat during reflood is
vs s J42 JO Her%; - '6 SOUTH 616 697 5646 P.8&'16 calculated by assuming that the fission power to total power ratio is constant at the value determined at the end ofblowdown. Therefore, a kinetics calculation is not performed so;no calculated return to power is possible.
B&W/FITscalculation ofLOCA loads is given in Reference 5, which addresses only B&W designed plants with once through steam generators.
The B&Wdesigned plants also include "vent valves" which ate check valves between the upper plenum and the downcomer.
Their purpose is to allow steam to vent from above the core to a cold leg break without having to flow t
ough the loop; thus alleviating steam binding. Incidentally, they may also reduce the forces on the core barrel followinga cold leg bteak by relieving differential pressure loadings. Hoover, t is is not t eir primary purpose.
Crediting these valves during a blowdown loading calculation is questionable since the valve opening time is of the same order as the loading transient. BRs methodology credits scram insertion even for Westinghouse plants without vent valves.
Therefore, they obviously do not play an important role of the B&%approach.
Westinghouse has demonstrated that cold log LOCAloadings are not sufficient to prevent control rod insertion, so the presence of vent valves is clearly not needed to conclude that control'rod insettability is acceptable in this case.
The B&Wmethodology of Reference 5, i.e. calculation ofblowdown loads with subsequent comparison to allowable loads is essentially the same approach currently being used by Westinghouse. The purpose of the B&Wreport was to document the calculation ofasymmetric LOCA loadings on the components and equipment in the reactor vessel comparunent in response to an NRC request.
Loads on fuel assemblies and reactor internals were included. Calculated peak LOCA loads are compared to allowable loads forthe guide tube assemblies and spacer grids (Section 10.6) and shown to have significant safety margins.
Comparison of the B&% Fuel Bundle with westinghouse The B&W17x17 fuel assembly is shown in Figs.
1 and 2 while the Westinghouse 17x17 fuel bundle is presented in Figs. 3 and 4. As demonstrated in the figures, both bundle designs are identical. Both bundle designs ate 12 feet in length, contain eight spacer grids, and utilize a 24 finger control rod cluster. As such, the mechanical behavior of both bundles is expected to be similar so that the Westinghouse load evaluation can be concluded to also reflect the expected behavior of the B&%bundle design.
Given that References 3 and 4 document an NRC approved LBLOCAmethodology for recirculating steam generator, i.e. Westinghouse and CS designs. plants which credits rod insertion followingall LBLOCAs, there is no relaxation of safety margin in crediting rod insertion followingcold leg LBLOCAs as requested for D. C. Cook.
Conclusions SCIENTECH has performed an independent third-party review of the Westinghouse Topical Reports, WCAP-15245 and%CAP-l 5246. Westinghouse used a version of the MULTIFLEX 3.0 computer code previously accepted by NRC for the WOG BBBP to calculate the blowdown loads on the reactor vessel and the reactor vessel internals. A conservative, and previously
- accepted, 1 ms break opening time was assumed and seismic loads were combined with the LOCA loads. For the guide tube acceptance criteria, Westinghouse used experimentally established allowable loads. A dynanuc load factor was applied to account for the transien;
~l cled J,VVV LO'D Ht=l NL: - 6 SGUTrl 616 697 5646 P.87/16 nature ofthe drag loading. The reviewer judges that the methodology used by Westinghouse to calculate the combined 'peak guide tube loading contains considerable conservatism.
Fuel assembly deflection and fuel assembly grid loading using the WECANcomputer code and the CNFD methodology. This methodology used is judged to be representative ofthe best current industry practices for determining fuel assembly response to LOCA and seismic loads. Results of the Westinghouse analysis clearly and conservatively demonstrate that for a design basis or LBB cold leg break, control rod insertion willnot be impaired.
It is also observed that there is a large margin ofconservatism in the evaluation of whether recriticality can occur followingthe switch to hot leg injection Following a LBLOCA,due to the followingfactors:
1.
The event itself, i.e. LBLOCA,has a frequency of occurrence which is very smail, especially for the design basis accident.
2.
LBB should be credited for these LOCAevents since it has been well established that:the design basis guillotine break is unphysical. Ho~ever, the Westinghouse analysis demonstrates that control rod insertion willnot be impaired, even for the unphysical guillotine cold leg break.
4 3.
The calculation ofboron concentration is very conservative.
On a best estimate basis, itis.
judged that the boron concentration willremain suRiciently high to preclude a return to criticality, even ifthe control rods do not insert. The Auachment addresses the conservatism in the boron concentration.
4.
Westinghouse has demonstrated that the combined seismic and LOCA loading forces willnot result in deformations which impair control rod insertion followinga design basis or LBB cold leg break.
With these multiple layers of safety margin, it is apparent that the risk associated with return to criticalityfollowinga switch to hot leg injection is insigniGcant.
References 1.
J. A. Barsic, et. al., "Control Rod Insertion Following a Cold Leg LBLOCA,D. C. Cook, Units 1 and 2", WCAP-15245, May 28, 1999 (proprietary).
2.
J. A. Barsic, et. al"Control Rod Insertion Following a Cold Leg LBLOCA,D. C. Cook, Units 1 and 2", WCAP-15245, May 28, 1999 (non-proprietary}.
3.
BOW Fuel Company, "RSG Loss-of<oolant Accident Evaluation Model for Recirculating Steam Generator Plants, Volume'1'-" Large Break", BAW-10168-A (non proprietary) 4.
Letter from A. C. Thakani, USNRC to J. H. Taylor, BkWNuclear Technologies, "Acceptance for Referencing ofLicensing Topical Report, BAW-10168P, Revision 1 'RSG LOCABOW Loss-of-Coolant Accident Evaluation Model for Recirculating Steam Generator Plants", dated January 22, 1991.
5.
Babcock 2 Wilcox Co., "Effects of Asymmetric LOCA Loadings - Phase IIAnalysis".
BAW-1621, July 1980.
c
Ht=t'NLi 3 SOUTH 616 697 5646 P.BS/16 ATTAC
-- B ron oncentratio Anal ses D.C. Cook Nuclear Plants Units 1 and 2 BA KGROUND Westinghouse has identified the possibility forre~riticality for DC Cook Units l and 2 fo)lowing a 1arge break LOCAfor locations only in the cold leg. The potential re~titicality is a result of the melted ice in the containment followinga cold leg break that dilutes the concentration of s'oiubie boron m the containment sump. During the recirculation mode ofECC injection, the ECC pumps take suction from the containment sump. Atsome later time during post-LOCA long term cooling. the ECC high-pressure injection is switched from all cold side injection to simultyleous hot and cold leg injection. Atthis time, typically ten hours or more after the opening of th
- break, the injection of the diluted'sump water into the core from the hot leg with the break in the:old leg creates the potential forre~riticality. Hot leg breaks are not a concern since the high concentrate boric acid that accumulates in the core is continuously expelled by the cold side injection through the break in a flushing action to supply sufficient borated water to the sump to preclude dilution with the melted ice. For cold leg breaks, because the cold side injection does not provide a flushing action to expel the high concentrate mixture from the core into the containment, the ice coupled with the accumulation ofboric acid in the vessel allows the sump liquidboric acid content to become dilute.
Itis important to note that the recriticality is a result ofthe assumption that the dilute sump water, with a concentration of 1500 ppm, does not.mix with the high concentrate boron mixture in the core. This no mixing assumption is considered very conservative, as there are natural convection and diffusion processes that willpromote the mixing of the
~
dilute mixtures with the high concentrate boric acid mixtures in the core.
Prior to the switch to hot and cold side injection, borated water enters the core, which is cooled during the long term by pool nucleate boiling. Withborated water entering the core, the steam in a bubbly flowmixture. rises to the surface ofthe two-phase region in the upper plenum where the steam disengages from the two-phase mixture. The boron remains behind in the two-phase mixture continually increasing the boron concentration.
Aftermany hours, the concentration in the vessel willincrease until the solubility limitis reached causing precipitation ofthe boric acid.
Boron mixing experiments have demonstrated that as the boron builds up in the core. the high concentrations that develop in the core produce concentration gradients that promote diffusion of the born into the lower plenum'from the core. The mixing volume in the vessel therefore includes the portion ofthe upper plenum containing two-phase, the core, and the lower plenum. The natural convection currents that exist in the core during this natural circulation, pool nucleate boilingprocess willpromote the mixing of the highly borated water in the core with the dilute hot side injection mixture.
~i f ClJ J.VVP LD'D H~Nli - 3 SOUTH 616 697 5646 P.89r16 Experiments addressing boric acid mixing in the vessel further demonstrate that the diffusion and natural convection processes that govern the Quid behavior in the vessel during long term cooling provide the physical mechanisms to promote uniform mixing of the boric acid throughout the upper plenum, core and lower plenum regions. During the switch to simultaneous injection, the diluted hot side injection water from the sump is expected to mix with the high concentrate boric acid mixture in the core. sufficieruly to preclude re-criticality followingall large cold leg breaks.
Adiscussion ofthe analysis of the transient boron concentration in the vessel folio'wing a large break LOCAis discussed below.
DISCUSSION A calculation of the boron concentration in the vessel for the DC Cook nuclear steam supply system was performed to assess performance of the ECCS to control boron acid content in the vessel and preclude re-criticality.
The attached figure presents the boron concentration versus time followinga large break LOCAwhere reactor coolant system pressure remains near atmospheric pressure. The figure shows the boric acid content as a function oftime with no flushing How assumed and indicates that about 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br /> is available for the operators to initiate simultaneous injection. The solubility limitat 14.7 psia indicated in the figure is 27.53 wt%. Since the safety injection system provides sufficient flow to maintain the core covered with two-phase followingrefiood of the core, the rate at which the boron builds up is a function of the decay heat generation rate. Any additional injection into the cold side is merely'pilled through the break. The equation solved to compute the boron concentration under these conditions is (please see the nomenclature section for identification of the syinbols):
d
- TVp, C
=
C dr pV The mixing volume, V~, is taken to be the volume in the lower plenum, core, bypass, and upper plenum, During the long term, the operators willneed only a single high pressure pump to cool and control the boric acid build-up in the core. To show the effect ofthe net core or flushing flow, it was.assumed that, the operators switch to simultaneous injection at 12 bours into the event. The effect of several flushing flow rates is demonstrated in the attached figure where injection flows of 467, 233, and 10 gpm were assumed.
These flushing flow rates maintain the long term boron concentrations at the values of 1.15, 1.73, and 12.9 wt% (i.e. 2001, 3017, and 22,680 ppm), respectively.
Flushing flows associated with the 467 gpm injection rate therefore produce boron concentrations above the sump or source concentration at the time simultaneous injection is initiated.
However, these analyses show that for a wide range ofECC injection rates, the concentration can be maintained well above the re-criticality condition of 1500ppin.
AEPNG - 3 SOUTH 616 697 56<6 P.18i16
. The equation that. is solved to compute the boron concentration in the vessel following the switch to simultaneous injection is given below.
C
= C, - C d
W(WJi~g dr Vz Vy The assumption for this analysis are summarized below:
- 1) The mixing volume consists of the lower plenum, core, bypass, and upper plenum.
- 2) The RCS is at 14.7 psia (solubilitylimit= 27.53 wt%).
- 3) The operators do not initiate simultaneous injection until 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> after the break opening. Atswitch, i/iofthe total injection fiowis split equally between the hot and
. cold legs.
- 4) The operators can reduce the ECC fiowby terminating pumps and/or throttling injection.
- 5) Core power is 1.02% of 3411Mwt.
- 6) Decay heat is based on the ANS 1971 standard with a multiplierof 1.0..
- 7) The concentration of the RWST is 2400 ppm.
CONCLUSION An analysis of the boron concentration in the reactor vessel following a large break LOCA, where the RCS does not refillwith ECC injection, demonstrates that the core can be maintained in a cooled condition while controlling boric acid precipitation and precluding re-criticality. The analyses contained herein demonstrate that the boric 'acid content in the reactor vessel can be maintained at concentrations that are well above re-criticality. Injection floe in the range 10 to 467 gpm can maintain core cooling (i.e. core remains covered with a two-phase mixture) while controlling the boric acid content well belo'w the precipitation limitof27.53 wt% and above the 1500 ppm value for re-
~ criticality.
0
616 697 S646 P. iii16 NOMENCLATURE C
= boron concentration C,
source boron concentration
= core boil-offrate orinjection rate prior to fiush Wflggg = core flushing fiowrate
= injection fiowrate in hot or cold side V =vessel mixing volume p
= liquid density t
= time
~ om M ~~ID 616 697 S646 P. 12i16 40 30 Boron Concentration vs Time DC Cook - Simultaneous Injection at 12.5 Hrs
NoFlush
'-'-'8467 gpm Flush 8
8233 gpm Flush
~~10 gpm Flush x
O 8 20 O
GO solubilitylimit=27.53%
10 0
10 20 Time (hrs) 40 10
b 0
va
~ ui saJJ iO>> JO
~ariz 6 bUUlH 616 697 5646 P.13i16 Ottl tN littler
~ >>
'og tItttltow ogpu go(
COIIEtl IOI gas ~ tttl (oj)oP lmr~ >)) tlili1
. I]oIIoIIo ff1'Jf
'!,IIM)
Ci L4 le]
JL)j.)!!l.lyt l ')l 0
~
I
>> ~
! Q t4~ I.g I
ttmt tto IIIAg~
g>>
',L 0
~ 0 I
~
~
I Ittt ICI
+IIIIOIIIIIOIINC (II IIICOtll(!III
- ItIIIIOtlIII'
~t1CI
~ ~
era a J4J LO~ 40
'I
~erron - a bUUIH 616 697 5646 P. 14i16 COUILIIID 4t!OK+
TOP VIKO NKUTRON AQOORQN4 NATKRIAL AXIAL'OCICNeHAPIII4'ROD.
Figure 2 8 8 %24 Efement CRA
O.CI1 OIC OCI O.CIS OIL Ltt Ol,+
t.(tt Ctt 4.4OI 'Sf Ctt I
4.4II Ctt
- 4. ISO Ctt I
P OO~OOQOD CCC44L 444
~
CLOCttt CC41LSL 41$ CÃ4LT Nt 'at&
ICCL OCO L14I ~ IL (Ct III~ CICCCSLt ICI 44 (Ct
~41ISa IOltttq COt CCCCLC
~.g at~
rt~atS~
I44.O CCI~
CC.Ostt~
tt.~Ctt~
I~
ICO.O ICI 44.4 att~
L 4.a Ctt Figure 3 Itestinghouse 17 X )T Fuel Assembly LongilLIdinat View
ohio o~e ~oqo
- r. ioiso 0.2U 4 U26 REF l6 SPACES'T O.U06~7.936 0 520
- 0. I22 RKL ASSBILY WITH ROO CLUSTER CIÃRQL GUIOE THIH8LE I U44 0.064 0 b OOO 0 0 0 CeITROL CLIISTEII ELSIEIT RIEL ROOS 26U RE4>0 00, I 0.37%
CLAD THICNKSS ~ 0,0225 CLAD IIATERIAL-2IRC-U
//
Jy cF 4ij~W'ys>
OO O
O O
O O
6 0 O
O O Oo
- OOO, P
00 0 0 0 0 0 Oo 0 0 0 0 0 0
0 0 00 I l.Ã3 RIEL AN' Alo CONTROL 'ROD PITCH 6A66 O.UO IRSTRIIENTATINl SHEATH FUEL ASSOILY tfITIOLV ROO CLUSTER CONTROL Figure 4 Nestlnghouse 47 X 17 Fuel Assembly, Cross-Section Vtee
llll )
- aii "4