ML20199L207
| ML20199L207 | |
| Person / Time | |
|---|---|
| Site: | Perry  |
| Issue date: | 11/24/1997 |
| From: | Elliot R NRC (Affiliation Not Assigned) |
| To: | Berlinger C, Dangelo A NRC (Affiliation Not Assigned) |
| References | |
| NUDOCS 9712010343 | |
| Download: ML20199L207 (12) | |
Text
. . ._- _ _ _ . _ _. . - _ - - - - _ _- --
e fpen%
~
f p 7's UNITED STATES g
,2 j NUCLEAR REGULATORY COMMISSION o WASHINGTON. O.C. 20WM001 s,...../ November 24, 1997 t
MEMORANDUM T0: Carl H. Berlinger. Chief Containment System and Severe Accident Branch '
Division of Systems Safety and Analysis FROM: Robert B. Elliott '/ 6 '
Containment System and Severe Accident Branch Division of Systems Safety and Analysis elo en d AnthonyD'AnkystemandSevereA Containment identkranch ;
Divisic.1 of Systems Safety and Analysis
SUBJECT:
TRIP REPORT T0-PERRY NUCLEAR POWER PLANT A meeting was held on October 6, 1997 at the Perry site to gather information on the new and unique strainer design which was installed at Perry during the most recent outage. The installation was performed using the guidance of 10 CFR 50.59. As a result, this meeting wr the first real opportunity for the staff to understand the 50.59 process for the strainer installation.
The staff members who v1 sited the site were D. Pickett the Project Manager.
R. Elliott, the Program Manager for the strainer program, and T. D'Angelo the Senior Reviewer for the evaluation of loading on the strainer. The delegation for the licensee was led by Mr. Gary Rhoads with the rest of the personnel identified in the attached attendance list. Attachment 1. Attachment 2 contains five questions the staff had asked Perry prior to the October 6 meeting about their September 13. 1997, response to NRC Bulletin 96-03 on ,
strainer plugging. The following is a summary of the this meeting.
Historically. hydrodynamic loads within the suppression pool have been well defined by the General Electric Company (GE) and described in many topical reports produced by GE (NE00 21471. NEDE-21730. NEDE-23817) with detailed 3rocedures specified giving the calculational methods to compute the lydrodynamic loads. These methods were specifically approved by the NRC in SERs (NUREG-0979. NUREG-0661). For the Mark 111 containment designs GE described their calculational methods in the GESSAR 11 document which gave the analytical procedures to be used. GESSAR 11 described each phase of the LOCA events by analyzing discreet time periods during the LOCA. In addition to the LOCA. operational events were also analyzed such as SRV actuation as a result of reactor isolation from the main condenser and, if required, a reactor depressurization event (ADS). These events all produce forces acting on the containment wall, other structures, and submerged piping in the suppression pool.
t EP AM Mjjp NE iEE CEDIG CTV L"# .p! p p.I)ll,IlI.ll D '\
4 2
The calculational method used for determining the submerged structure drag load from LOCA air bubble load was described in the PNPP USAR Appendix 3B, Attachment L. This method is based on applying a uniformly distributed constant pressure to a given surface area and multiplied by a dynamic load factor. This translates to an equation of the form. Air Bubble Pressure x Area Of The Submerged Object x Dynamic Load Factor (referred to here as the PNPP load methodology). The air bubble aressure was obtained from the LOCA air bubble pressure curve given in the GESSAR an'i shown in PNPP USAR Figure 3B-68. Area of the submergoi object is the area of the object under study such as a +he- area on the strainer to which the drag forces are a ting upon, A dynamic load factor is a multiplier used Lo account for the suddenly applied force acting on the object and having an effect on the dynamic response of the object.
The' method described above is not the methcd used in the GESSAR 11 Appendix 3B
-for calculation of the LOCA air bubble dran loads on submerged structures but thismethodisclearlydescribedintheUS4RandwasnotdisputedinthePNPP USAR NRC safety evaluation report. NUREG 0887. The basis for acceptance by the NRC is now unknown. However, it appears from the statement in the PNPP USAR Section 3BL.8.1.2 that a comparison calcul6 tion was made by Perry and that calculation had demonstrated that their method yields approximately twice the force acting upon the object as would the GESSAR 11 method was the basis for staff acceptance. Therefore, one could reasonably conclude that their method is more conservative.
The licensee has continued to use this PNPP method for the calculation of the LOCA air bubble drag loads actina on the new strainer. Since the new strainer is very different from previously envisioned objects, the licensee needed to reevaluate parameters in the calculation of the drag loads. The most significant parameter is the strainer area. For this area, the previous
-calculation used the entire exposed area. However, for this new application.
the licensee has assumed only the metal surface area. This neg.ects-the hole open area. The licensee had not performed a comparative calculation for the new strainer design usinj the GESSAR Il methodology.
Further. the licensee's method in the USAR does not account for all physical effects such as loads acting on a body submerged in a moving or accelerating fluid. For example, drag is discussed in detail within a paper by Dr. F.-J, Moody, presented at the Thermal Reactor Safety Meeting, July 31 - August 4 1977. Sun Valley. Idaho. In that paper, drag forces are defined as "a
-combination of skin friction and pressure effects caused by down stream wake formation." This definition of drag most certainly_ describes the force acting on a booy due to a constant velocity fluid passing over a submerged body,
~
~ However, as the paper further discusses. " steady drag does not include the forces _ associated with- fluid acceleration which is referred to as the acceleration force."
. j
- t 3-The neglecting of the more important acceleration drag may not have been ,
significant for the smaller past strainer designs. But for this new design.
the staff indicated that the omission of the acceleration term is the i
- elimination of a significant load acting on the strainer. Therefore, the '
staff indicated that the continued use of the previously accepted methodology was unacceptable for the new strainer design. >
We also indicated that there existed an acceptable approach as described in GE
, topical report NED0 21471. " Analytical Model for Estimating Drag Forces On Rigid Submerged Structures Caused By LOCA and Safety Relief Valve Ramshead Air
! Discharges." This report describes a method for establishing a procedure to -
analyze the drag forces acting on a submerged body. The topical report is e based on a concept of modelling drag force based on work done by Dr. J. R.
Morison et al. . "The Force Exerted by Surface Waves on Piles." Petroleum Transactions, AIME. vol. 189. 1]50. and later su> ported by experimental and analytical work done by Drs. G. H. Keuleo m and .. H. Carpenter. " Forces on
- Cylinders and Plates in an Oscillating Fluid." Journal of Research. NBS. <
vol. 60. no. 5. 1958, pp. 423-440 and further work done by Dr. T. Sarkaya.
" Forces on Cylinders and Spheres in a Sinusoidally Oscillating Fluid." ASME paaer no. 75 APMW-27. Those reports all support the notion of a model with su) merged bodies in an accelerating fluid producing a drag force and i acceleration force. ;
The GESSAR 11 lists ten events which all produce hydrodynamic loads acting on l submerged objects within the suppression pool. The above discussion relates
- only to one of these events. the LOCA air bubble. The other events are
- related to longer term events, that being SRV air bubble. chugging and ;
4 unstable condensation oscillation (CO). Except for the LOCA air bubble. PNPP l
- USAR appendix 3B has followed the GESSAR 11 methodology for calculating hydrodynamic loads. Therefore, the staff indicated that the approach for these other loads is acceptable.
The drag load calculation for the other events utilize analytical procedures documented in GE topical reports NED0-21471. NED0-25153, and NEDE-21606.
Those procedures discuss the calculation of a hydrodynamic mass. For the new stacked disk strainer design, a similar procedure should have been used. The areviously approved methods used simple geometries with known tested values of lydrodynamic mass. Because of the complex configuration of the new strainer, this information is not available. Therefore, the licensee has chosen to develop a new methodology for calculating the hydrodynamic mass based on an
-approach which uses the finite square method. This method when summed would in theory be capable of modelling the entire strainer. The staff _ indicated that while in principle we had no basic differences we did recognize that the approach. represented an extension of the technology that was previously used at Perry. Because of this more complex approach, the staff indicated that the
- path for successful resolution would require supporting test data.
s
...,...,.r.,, S .~.-_yw4.,-y... . _ ,, -. ....<y,m, y . ,.,4 # y ._- ,_ ,..-. , _%,,,c,.m,.. .w.....,,,,_w,_..wrm ,m,.. -
4 l
This was particularly important in view of the significant reduction in acceleration drag as calculated by this finite square method. The licensee !
indicated during the meeting that the calculated value would be about 1/100') :
of the previously calculated load using a solid body, in spite of this i significant reduction, the licensee said that the uplift load would still be l 1500 lbf while the weight of toe strainer is about 2000 lbf. ,
The licensee understood the staff position but also indicated that no testing was planned. They said that they needed more time to determine a course of i action and agreed-to get back to the staff in about 30 days.
Another issue discussed during the meeting was the potential for air entrain-ment in the strainer. The issue was first raised because the new strainer is located within the exclusion zone of the quenchers. The exclusion zone was -
initially established to preclude any structures located within the zone being subject to potential forces from the quencher discharge, The concern was the possibility of entrained air from the quencher discharge being ingested into the suction intakes.
The licensee had evaluated the potential effects associated with locating the new strainer within_the quencher exclusion zone and had concluded that the strainer would not be subjected to any damaging effects. Specifically. he calculated that the vertical Dubble movement would be less than four inches which is quite small. In other words, the bubble will not penetrate the pool sufficiently to impact the strainer outer surface, in addition. the surface velocity at the strainer surface is so low as to not overcome the buoyant i forces of the bubble. -
Based on the above analyses, the staff agreed with the licensee that the new strainer design with a separation distance of 11 inches is more than ',
sufficient to preclude ingestion of the a1c bubble. The calculated vertical bubble penetration into the pool is less than 4 inches.
The new design creates a new type of strainer failure. It is the local uplift i potential of a section of the strainer, The design is such that there are no restraints in the vertical direction other than the deadweight of the strainer. The licensee has calculated the uplift forces are about 73 percent of the deadweight. While this wou'd first appear as a reasonable margin. One must keep in mind that the calculated uplift force needs to be carefully
- evaluated in light of this failure mode. ;
The licensee also 3 resented the new calculated stresses within the new >
strainer design. _Jsing ASME Section ill criteria the licensee evaluated the stresses within the frame and perforated plate. The results showed that the ;
minimum margin to allowable stress was about 81 percent for both the frame as wall as the ) late. The staff indicated that if the loadings were found i
acceptable t11s stress analysis would also be found acceptable.
1 r
w ,=- irs *w - w w w-w .-ee .w - em ,zu,w--*e~www " w, .
,w1--.c--v-ww---w - - se , ,ir yv -,v-v gr ,wr y-e- w-e g wvir ww e *= r - We-~ve - w iv,* e-- =vww e'ww 5--w-w=-w ww y- vv -w ~*- v---vws-wvv ww 'e t *em--t wer'---wve- vv- ty--9r"
. i e
5-i I
CONCLUSIONS i
- At the end of the meeting, we summarized the open items that remained, They ;
were justification for the hydrodynamic mass by test and the applicability of
. the pressure times area ap3 roach in lieu of using previously a) proved drag .
i methodology, For both of t1ese items, the licensee indicated tlat they needed time to think about how to respond to the staff's position. As a result, it ;
was agreed that 30 days to properly evaluate staff comments was appropriate. s it shoula be noted that the positions taken during the meeting were based on the presented information. The staff would intend to do a more complete review once the information is docketed. The meeting was then closed. t Docket Nos. 50 440 and 50-441 i
Attachments: As stated 4
1 7
i
.ee-- --. -fm,-> W---- y,r, ,e--a.y.... .we,.<a. -i-.-tr .
,.,qy,e me -+mg-gg,,-upp,-y-gpa.m-up-+9.,,-inwe,yer's-T'W%e-N'c'P T*9 -Tm'-- g.q #y19-- w w- y-gT-s'vv TgTW9?F'" -"**f -'T*F- Y-""N'*"MT
l
- At'achment 1 .
4 Wf6F * '\ 9%
lbels,(,sa, C.E.I Ecw A,cu esE 4 Sq ,. sil.
Gcmm Tree Great, y Alue/ed$fdh.M'ws S~ g <.
Smerra, AsA F-mcasy
.DoD6;24 .fxs- Suna. bme.J Sociaat W E ,// E' n y .FewM. Sm <e;- & 4 4 Heaisar ,sks fukaw Cowdv be Ewnm AfTD, Arut & cs c . .,
C A c4< f hryan, $ober/ NA"/cEr CA>ERc0Al l' '
b h y e ils , & y v.sNKc s y in e e e.
p u n ,s ni _ a ea n
- w. ,u,...-
i bbad#1" (
Q 167 ) I*) ('l bN61k C\ M tt T ' bow NM s
Nh f'NN I 'W~ % k'
___ _____ ___ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~
5-CONCLUSIONS At the end of the meeting. we summarized the open items that remained. They were justification for the hydrodynamic mass by test and the applicability of the pressure ttmes area ap) roach in lieu of using previously 9) proved drag methodology. For both of t7ese items the licensee indicated tlat they needed time to think about how to respond to the staff's position. As a result, it was agreed that 30 days to properly evaluate staf f conrnents was appropriate.
It should be noted that the positions taken during the meeting were based on the presented information. The staff would intend to do a more complete review once the information is docketed. The meeting was then closed.
Docket Nos. 50-440 and 50-441 Attachments: As stated DISTR.llIUTION:
Docket Files SCSB r/f (2)
PDR GHolahan SNewberry DLynch DPickett RSavio JHopkins JDonohew
. JClifford DOCUMENT NAME: PERRY.TRP 10 ccesse e copy of this document, Md6cate M the bes: 'C' = Copy without ettschmententiogure *E' = Copy with attachment enclosure 'N' = No copy oi l ILL SCSB:DSSA:NRR (( SCSB:DSSA:NRR l 5 SCSB:DSSA s l lij NAMl AD' Ange:o:bw A///f8Elliott i ,a - JKudric h u\
DAll l l/Z.y/97 V ll/ t/97 11/ fl97 Q / /97 / /97 0FFICIAL RECORD COPY 1
e 0
CONCLUSIONS At the end of the meeting, we summarized the open items that remained. They were justification for the hydrodynamic mass by test and the applicability of the pressure times area ap) roach in lieu of using previously a> proved drag
- methodology. For both of-t1ese items, the licensee indicated tlat they needed time to think about how to respond to the staff's position. _ As a result it '
~
was agreed that 30 days to_ properly evaluate staff comments was appropriate.
c , it should be noted that the positions taken during the meeting were based on -
the presented information. The staff would intend to do a more complete review once the information is docketed. 1ne meeting was then closed. .
4 - - Docket Nos; 50-440 l and 50 441 j : Attachments: As stated . ! -
+
+ 4...
2
- ; DISTRIBUTION:
Docket files
- SCSB r/f (2)
- PDR
, GHolahan SNewberry DLynch DPickett RSavio JHopkins JDonohew JClifford l DOCUMENT NAME: PERRY.TRP
- to esce4,e e copy of this docunwnt. Indicate bi the ben: 'C' = Copy without attachment 4e(Tohre *f* = Copy with attachmentienclosure 'N' = No copy Of F ICL : SCSB:DSSA:NRR 16 SC5B:DSSA:NRR]is SCSB:DSSA \1 I L' ] I NAME AD' Angelo:bw N/PFElliott W17 JKudricWR\\ \
DATE- 11/Z.y/97 Ull/t1/97 11f\W97 \\Mi \ / /97 / /97 0FFICIAL RECORD COPY '\ Y a
l
.,. Attachment 2 ECCS Sucilon Strciner Issues Attachment 2. nace 11 of 36 issue:
The strainer divisional separation is discussed in the submittal but please explain how mechanical separation is maintained between trains. The hydraulic separation is clear via the baffle plates whMn the strains r body but the 360 degree design of the torus shape ofstrainer does mechanicalsy connect all divisions together, including the RCIC system.. Therefore, if a crediblefailure mode exists that can transmit a mechanicalloadfrom the stralr.cr to the piping, how is divisional separation maintainedforpreventing common modefa"are? The postulated event would be movement of the strainerfrom a hydrodynamic event which could produce either vertical movement or azimuthal rotation of the strainer.
Response
The strainer has been analyzed and designed to withstand all credible loading and load combinations, including asymmetric loading, such that movement of the strainer as the result of any license basis event in either the vertical direction or azimuthal rotation of the ctrainer is not possible. A review of these analyses evidences that the strainer does not experience any significant azimuthal loads (9 kips 5 percent of the total strainer weight). The calculated azimuthal loads are included in the strainer design. The design also includes all possible combinations of uplift loads including buoyancy, seismic, SRV actuation and LOCA. The '
mean downward load per strainer scetion is approximatt.ly 525 lbs. (26.6 percent of strainer weight), under the most adverse loading condition. The analysis shows that the strainer will not uplift under any design loading condition.
Attachment 2 nace 16 of 36 1ssue:
The discussion on potential air entrainment into the strainerfrom a quencher initial discharge not being credible appears to have been based on the low approach fluid velocity to the strainer. Although thefluid velocity is expected to be very low near the strainer perimeter, the airjet exiting the quencher has a significant horizontal velocity component. I recall in a past meeting with Perry that there are afew locations around the strainer where the quencher is within one or twofeet of the strainer, if that is correct, then it appears that the strainer may lie within the quencherjet path and some amount of non-condensibles may enter the perforated plate. Given thatpostulate, please explain whatprevents the non-condensiblesfrom entering a ECCSsuctionflowpath?
Response
Perry's calculation T219 determines the sin of the largest SRV bubble generated, using Perry specific SRV discharge piping and GESSAR 11 methods. The diameter of the largest bubble, and its center location are such that the bubble does not touch the s'mee of the strainer.
Calculations show that an air jet directed downward would need to extend a minimum of 48 inches to reach the elevation of the suction pipe (see attached sketch).
ECCSm-Sv m ardnu Issues m
Attachraent 2 ;we 18 of 36 1ssue:
The paragraph discussing Containment Loads Considerations discusses the LOCA air bubble load calculation procedure which were described in the PNPF USAR. The USAR in section 3BL2.3 does mention the LOCA air bubble pressurefigure comained in the GESSAR Hfigure q 3B-68 is used for LOCA air bubble drag loads. However, the USAR then presents the calculated LOCA air bubbleforces using PNPP methodsfor a sample problem and compares that calculated number with the GESSAR H method. No description or discussion wasfoundof 5 the specific method used by PNPP for the calculation of LOCA air bubble forces. Please describe you.~ calculationalprocedurefor the LOCA air bubble loud calculation andpresent equations used Further, a:! A pg. 8 of 17, paragraph 3.1.8 describes the equation used by PNPP for alculat;.. '
! 4 ?agforcefrom a LOCA air bubble. That equation would normally be usedfor h I'te calculatio.: of a umformly distributedpressure ac:ing against a surface. However, since
[ the even: in quest %n is quite dynamic and the submerged bodies are contained within un acceleratingflowfield, a dynamic loadfactor was applied by PNPP. Please explain how the dynamic locifactor was calculated? Also, how large ofa d,fference msts between the PNPP method and the GESSAR H method in calculating the dragforces actint cn flee strainer.
The last paragraph last sentence of this page states that the strainer is constructed using 4 perfuratedplate and that the input to the GESSAR H equations were adjusted to appropriately calculate the resultant strainer loads. Please dtscribe what CESSAR H values were adjusted and the basisfor those selected values
Response
The LOCA air 'oubble loads calculated in the US AR for the sample pcoblem are based on the
" pressure x area" using the wall pressure:,, based on GESSAR II, Fig. 33-68. Tie stiainer design utilizes this method with application of a dynamic load factor (DLF) to account for the dynamic nature of the loading. Per the PNPP USAR this methodology provides loads which bound those calculated using the GESSAR 11 methodology. The DLF for the LOCA air bubble loading is determined for a triangular load profile as der.cribed in the GESSAR (Fig. 3B-11) and the structural frequency of the ECCS straine .
No separate GESSAR calculations have been performed for this loading on the strainer; however, according to the sample problem, the WP method oroduced a load that was 2 times the magnitude of the GESSAR method (247 L6 :bf GESSAR II; 4810 lbf USAR). A comparison of the SRV loads on the strainer using the PNPP US AR and GESSAR 11 metho6 shows that the GESSAR method produces loads which are app oximatel, 2 percei.t of the loads based on wall pressures. A similar compatison of the GESSAR .: ample problem shows the GESSAR loads to be 4 percent of those based an pressure (USAR ction 3BL.2.3). Therefore, the LOCA air bubble loads calculated for the strainer are expected to be conservative.
To account for the variation in the flow accelerations, the SRV acceleration drag loads on the strainer perforated section are calculated at 15 locations (see attached sketch). No reduction in the area due to the holes i: considered. However, ce holes in the perforated plates do pro,ide a conduit for propagation of the pressure through the strainer sections under acceleration drag loading. Additionally, the standard and acceleration drag loads are conservatively combined by absolute summation, disregarding the out-of-phase relationship between the two loadings.
[ ECCS Suction Strciner Issues Attachment 2 page 19 of 36 Issue:
PNPP report states that no specific margins had been depned in GESSAR 11 for the hydrodynamic loads but PNPP has met the original licensing basis. Please describe the margins which PNPP has maintained with the current strair.cr design For example please state the calculated stresses within the strainc; frame compared to a:lowable and the up-hft load compared to gravity Respor.se: '
The design margins (calculated / allowable stresses) are specified below:
Comnonent Interaction Ratio Strainer 0.81 Strainer Radial supports 0.79 Perforated Plate 0.81 The above interaction ratios are based on comparison of Level 'D' stresses to Level 'C' allowables. Use of Level 'C' allowables is required only for certain components where lower deformations (strain) limits are required to ensure strainer functionality under accident conditions.
'S j
All strainer sections experience a net downward load under design soading condition, including ji the efTects of buoyancy, seismic SRV actuation and LOCA. Each strainer section maintains a 7
mean downward load of 525 lbs. (26.6 percent of strainer weight) under the most adverse uplift loading.
Telephone conversation dated October 2,1997:
Issue:
PNPP stated that they had utilised a computer code in the calculation on the dragforces acting on the strainer. Please describe the methodology used in the code. Picase describe how the code was validated.
Response
A computer code utilizing GESSAR 11 prescribed methodology was used to determine the PNPP specific SRV air bubble properties. The code 'itilized standard integration routines and was validated by duplicating the bubble properties in the GESSAR sample problem.
.. - _ _ _ _ _ _ - - _ _