ML20077M406
| ML20077M406 | |
| Person / Time | |
|---|---|
| Site: | Limerick  |
| Issue date: | 08/19/1983 |
| From: | BECHTEL GROUP, INC. |
| To: | |
| Shared Package | |
| ML20077M397 | List: |
| References | |
| NUDOCS 8309120324 | |
| Download: ML20077M406 (13) | |
Text
.
't EVALUATION OF DRYWELL INSULATION DEBRIS EFFECTS ON ECCS PUMP PERFORMANCE Limerick Generating Station Units 1 & 2 Philadelphia Electric Company Bechtel Power Corporation San Francisco, California Revision: 2 P-261(b)/7 Date: 8/19/83 8309120324 830901 PDR ADOCK 05000352 A
TABLE OF CONTENTS Page No.
TABLE OF CONTENTS i
1.0 INTRODUCTION
1
2.0 BACKGROUND
1 3.0 METHOD OF ASSESSING DEBRIS HAZARD 2
3.1 Assumptions Used to Determine Maximum LOCA Debris Generation 2
3.2 Amount of Debris Generated 3
3.3 Insulation Debris Transport to the Suppression Pool 4
3.3.1 Short Te.a Transport 4
3.3.2 Long Term Transport 4
3.4 ECCS Suction Strainer Blockage Due to Insulation Debris Carryover Into Suppression Pool 7
3.5 Effect of Suction Strainer Blockage on ECCS Pump NPSH 9
4.0 REFERENCES
10 LIST OF TABLES Table 3.1 Summary of Insulation Debris Trans-port Into the Suppression Pool 8
Table 3.2 Summary of Calculated Debris Head Losses and Effects on ECCS Pump NPSH 11 LIST OF ATTACHMENTS A.1 LDCN FS-331, Revising PSAR Sections 6.2.2.2 and 6.2.9 P-261(b)/7
-i-8/19/83
1.0 INTRODUCTION
1.1 Purpose There is a concern that a high energy pipe break within the primary containment might create insulation debris that would build-up on the Emergency Core Cooling Systems (ECCS) pump suction strainers, impairing operat-ion of these pumps.
Such debris considerations are part of Unresolved Safety Issue A-43, Containment Emergency Sump Performance.
In response to this concern, an evaluation was made to estimate the maximum quantity of insulation debris that might be generated by a LOCA, the amount of such debris that might enter the suppression pool and cover the ECCS strainers,.the attendant pressure losses and resulting effect on the ECCS pump NPSH margins.
In attachment A.1, it is concluded that small pieces of fiberglass insulation are the only type of drywell insula-tion debris that might enter the supression pool and went Lim ECCS strainers.
Therefore, only fiberglass insulation debris is considered in this evaluation.
1.2 Summary of Results:
It has been determined that fibrous insulation debris generated by the LOCA jet impinging on nearby insulating materials will not jeopardize ECCS pump operation at Lime-rick Units 1 & 2.
The head losses due to fibrous debris build-up on the ECCS pump strainers in the suppression pool from a worst case LOCA (in regards to debris generation), concurrent with minimum NPSH conditions at the ECCS pumps, does not cause NPSH available to drop below NPSH required by the ECCS pumps.
2.0 BACKGROUND
2.1 Limerick Units 1 & 2 are 1,100 MWe, BWR's with Mark II containments.
2.2 The ECCS pumps take suction from the suppression pool inventory.
Water released into the drywell from a LOCA accumulates on the drywell floor and enters the suppression pool by overflowing into downcomers that extend 18" above the drywell floor.
Limerick has 87-24" nominal diameter downcomers penetrating the drywell floor.
2.3 Insulation Types Employed in the Drywell:
P-261(b)/7 8/19/83
- -. -.. ~. - - -
.. -.. _.. -. -... - -.. - - -. ~.. - -. -. _ - - _ - - - _. -
a)
Stainless Steel Jacketed Fiberglass Thermal Insulation (Owens-Corning Nukon).
Fiberglass with woven fiberglass covers is used to insulate high temperature piping and the recirculation pumps.
The insulation blankets are protected with 22 gage, minimum., stainless steel jackets having seismically qualified latches.
b)
Metallic Reflective (Diamond Power).
This type is used on the reactor pressure vessel (RPV).
c)
Jacketed Fiberglass Anti-Sweat (Owens-Corning Fiberglass).
This type is used on chilled water lines inside the drywell and constitutes only a small percentage of the insulation used in the drywell.
L d)
Encapsulated Low Conductivity (Johns-Manville Min-k).
This type is used at the whip restraints where the clearance does not allow the use of thicker Nukon insulation, and constitutes a very small percentage of the insulation used in the drywell.
3.0 METHOD OF ASSESSING DEBRIS HAZARD 3.1 Assumptions Used to Determine Maximum LOCA Debris Generation:
3.1.1 The criteria for selecting the worst case pipe break with respect to insulation debris generation are the same as the assumptions outlined in Appendix C of Reference 4.1.
These are:
I a)
All insulation inside the Region I jet cone is destroyed.
The Region I jet cone is a 7 pipe diameter long cone, expanding 45 degrees on either side of the centerline of the pipe.
b)
All insulation inside the Region II jet cone is dislodged.
Region II is the extension of the Region I jet cone beyond l
the 7 pipe diameter distance with the same expansion angle.
I c)
No credit for the shadowing effect of l
the broken pipe in the Region I portion jet cone has been accounted for in the l
l P-261(b)/7 8/19/83 t
_ ;.. _.. a
.m=_
a =
._=:
- m. -
i analysis.
In the Region II portion, credit was taken for the shadowing effect of whip restraints and structural steel.
i d)
Insulation on the broken pipe between whip restraints and any piping inter-secting the Region I jet cone is assumed to be shredded.
e)
Insulation in the Region II jet cone is assumed to be dislodged as whole or torn blankets.
f)
The protective effects of the stainless steel jackets on the Nukon insulation were not considered, and the Nukon in-sulation was considered to be unprotected for this evaluation.
It is expected that 4
limited protection against shredding would be provided by the insulation jacketing.
3.1.2 The Region I jet cone shape is considered con-servative in regards to the 90' jet cone ex-
~
oansion angle.
Based on the analytical techni-ques in Reference 4.8, the actual Region I jet cone i
shape would be much narrower and longer.
The same l
obstacles ~ mentioned in paragraph 3.2.1 would break up the actual longer jet cone before it can damage insulation that would be in assumed Region II.
Consequently, if the actual jet cone shape was used in this case the amount of shredded insulation debris generated by the LOCA would be much less.
3.2 Amount of Debris Generated:
3.2.1 A review of the postulated pipe rupture loca-tions (PRL) on large diameter, high energy lines was made to locate where the jet cone would be oriented toward fibrous insulation.
Due to separation of high energy lines, the shadowing effect of the numerous whip restraints on these lines, and surrounding structural steel, there are very few PRL's that have any real potential to generste a significant amount of fibrous debris.
However, following the conser-vative criteria outlined in item 3.1, it was determined that a break of a 26" 7 main steam line would generate the largest amount of shredded i;
debris.
The calculated amount of debris generated is 44 ft3 This is. equal to 19 linear feet of 26" 7 pipe insulation, 3 1/2" thick.
.Other insulation removed by the Region II por-tion of the jet cone was not quantified.
A review of the geometry of the jet cone showed P-261(b)/7 8/19/83
--...-.-.--..;-- :. =
- -... -. ~ -.. -....
.c
=
t that numerous structures, pipe whip restraints and structural steel would provide protection against jet effects in Region II so that only a few whole or torn blankets would be expected to be dislodged.
However, the attached revision to LGS FSAR for Section 6.2.2.2 (Attachment A.1 of this report), discusses the physical barriers that will prevent whole or torn blankets from entering the suppression pool.
As a result, whole or torn blankets were not considered to be a concern.
3.3 Insulation Debris Transport to the Suppression Pool.
3.3.1 Short term transport:
3.3.1.1 Short term transport is caused by the pipe break jet, pipe whip and pipe impact and terminates at the end of blowdown.
3.3.1.2 It was assumed that all shredded debris generated at the PRL would fall onto the drywell floor.
This is considered conservative because some shredded debris (and most dislodged whole blankets) would be trapped by grating floors and other surf aces below the PRL.
3.3.2 Long term transport:
3.3.2.1 Long term transport is caused by operation of ECC systems.
3.3.2.2 Based on data contained in Reference 4.3, the long term transport evalua-tion considered the behavior of three types of shredded fibrous debris after it falls onto the drywell floor:
1)
Debris that immediately sinks upon contact with hot water.
2)
Debris that floats for an indefinite period.
3)
Suspended debris that slowly sinks.
3.3.2.3 It is assumed that the shredded debris P-261(b)/7 8/19/83
is distributed over a large floor area and concentrations adjacent to a few downcomers do not occur.
Con-sequently, the bulk properties of the drywell floor apply to the region where insulation debris is scattered.
3.3.2.4 The shredded fiberglass insulation debris on the drywell floor will initially come into contact with hot water released from the LOCA.
Test data contained in Reference 4.2 shows that shredded fiberglass sinks in 20 to 30 seconds in 120*F water.
Per Reference 4.7, Owens-Corning tests show that the same sinking charac-teristics apply to fibrous Nukon fragments.
The initial water build-up on the drywell floor will be reactor inventory that is much hotter than 120*F.
There is sufficient volume below the downecmer overflow level (i.e.,
18" above drywell floor) to contain the maximum water release by a design basis LOCA (which would be from a recirculation line break), so there is no overflow into the downcomers until after the ECCS pumps are operat-ing.
There will be substantially more than 30 seconds' exposure to hot water before the start of overflow into the downcomers.
Consequently, it can be assumed that most debris is quickly wetted and is sinking or already on the bottom of the drywell floor pool when overflow starts.
3.3.2.5 The maximum possible ECCS flow condition was used in order to maximize the flow velocities into the drywell downcomers and to minimize the residence time available for slow-sinking debris to settle.
Specifically, the evaluation is based on simultaneous operation of the following ECCS pumps all operating at their runout flowrates.
4 RHR Pumps 4 x 10,750 GPM = 43,000 gpm 4 Core Spray Pumps 4 x 3,950 GPM = 15,800 gpm Total ECCS Flow -------------------- 58,800 gpm P-261(b)/7 8/19/83
3.3.2.6 Transport test data contained in Refer-ence 4.2 indicates that a flow velocity exceeding 0.2 ft/sec is required to entrain fibrous insulation shreds lying on the bottom of the drywell floor pool.
The calculation indicates tho ma;:imum ECCS flow (58,800 gpm) into the 87 downcomers has virtually no effect on the bottom of the pool on the drywell floor.
The maxin.um flow velocity at the bottom of the drywell floor pool is less than 0.06 ft/sec.
Thus, sunken insulation debris will not be re-entrained by the flow into the downcomers when the ECCS pumps are operating.
Based on Item 3.3.2.4 and the foregoing, it is conservative to assume that 40% of the original shredded insulation on the drywell floor immediately sinks and remains on the drywell floor throughout the period of ECCS operation.
This assumption is consistent with test i
data stated in Section 4.7.2 of Reference 4.3.
3.3.2.7 Based on the general recommendations for floating fibrous insulation con-tained in Section 4.7.2 of Reference 4.3, it is assumed that 30% of the original shredded volume remains afloat during ECCS operation. It is also assumed that all floating debris enters the downcomers and supression pool, and coats the ECCS pump strainers.
This is conservative because fiberglass buoyancy test data in References 4.1, 4.2 and 4.5 show shredded fiberglass clumps tend to rapidly sink in hot water.
3.3.2.8 The remaining 30% of the original shredded insulation volume is considered to be slow-sinking debris disperseu in the water on the drywell floor.
A calculation was performed to determine the flow zone around each downcomer P-261(b)/7 8/19/83
)
e rim that could draw the slow-sinking debris into the downcomer before it settles to the bottom of the floor pool.
This calculation was based on a settling velocity of 0.012 ft/sec which is the median settling rate between individual fibers and very small clumps of Nukon insulation based on data contained in Reference 4.5.
Calculation results indicate that a carryover factor of 50% for the slow-sinking debris would be conservative.
In terms of the orignal shredded insulation volume, this results in approximately 15% entering the downcomers due to the slow-sinking phenomena, and coating the ECCS strainers.
3.3.2.9 Based on the forgoing transport analysis) 19.8 ft3 (45%) out of the original 44 ft of insulation shredded by the LOCA enters the suppression pool.
A summary of the fibrous insulation transport phenomena is shown on Table 3.1.
3.4 ECCS Suction Strainer Blockage Due to Insulation Debris Carryover into tha Suppression Pool:
3.4.1 It is assumed that all pieces of fibrous insula-4 tion that enter the suppression pool eventually deposit on the 8 sets of active ECCS strainers.
No credit was taken for quiet zones in the ai.ippression pool where insulation could settle-out without being-drawn onto one of the strainers.
The operating pumps identified in Item 3.3.2.5, have conical type suction l
strainers in the suppression pool.
The total ECCS strainer surface area is 160.5 fte.
3.4.2 Per Reference 4.1, calculation of strainer screen head loss due to fibrous insulation debris blockage is based on the equivalent insulation blockage thickness, ti, as defined below:
e
. Transported debris volume ava11aole screen area i
l P-261(b)/7 8/19/83
TABLE 3.1
SUMMARY
OF INSULATION DEBRIS TRANSPORT INTO THE SUPPRESSION POOL Amount Transported original to Amount Supp Pool ft3 ft 3 a) Short term transport:
Debris generated by LOCA 44.0*
N.A.
Debris falling onto drywell floor 44.0*
N.A.
b) Long term transport:
Breakdown of debris types on drywell floor:
- Immediate-sinking (40%)
17.6 0.0
- Floating (30%)
13.2 13.2
- Slow-sinking (30%)
13.2 6.6 l
44.0 19.8
- All insulation debris in shredded form.
t i
l l
1 P-261(b)/7 8/19/83
. ~,.
The total transported debris volume is distributed between the RHR and Core Spray strainers based on the ratio of individual pump flowrates to total ECCS flow.
3.4.3 During the Reference 4.7 telecon with Owens-Corning, Bechtel was informed of a head loss formula developed specifically for fibrous Nukon insulation.
The Owens-Corning formula was approved for use in this review during the Reference 4.6 telecon.
The approved Owens-Corning Nukon formula is given below:
AH =
68.3 (ti)l.07 gy)1.79 where:
AH = screen loss, ft of H O 2
ti = equivalent blockage thickness, ft V
= screen approach velocity, ft/sec 3.4.4 Using the above head loss formula, the maximum strainer head loss caused by the deposition of LOCA generated fibrous insulation materials are shown for the RHR strainers, and the Core Spray strainers in Table 3.2.
3.5 Effect of Suction Strainer Blockage on ECCS Pump NPSH:
3.5.1 Minimum NPSH is available for the ECCS pumps when the suppression pool temperature is 212*F, the drywell pressure is 14.7 psia, the suppression pool is at minimum level and the strainers are coated with debris.
The calculated minimum NPSH available at the pump suction was compared to the required NPSH.
3.5.2 Table 3.2 shows that the worst case strainer blockage, concurrent with the minimum available NPSH operating condition, only consumes about 10% to 25% of the NPSH margin for the ECCS pumps based on the NPSH required values provided l
by the pump manufacturer.
(The NPSH required values stated by G.E.
in the FSAR Process Flow Diagrams contain a large safety margin over the NPSH required values given by the pump l
manufacturers).
l l
I l
P-261(b)/7 8/19/83 l
l
4.0 REFERENCES
l l
4.1 Serki:,
A.W., " Containment Emergency Sump Performance,"
NUREG-0897 (for comment), NRC, April, 1983.
I 4.2
- Brocard, D.N., "Buoyuancy, Transport, and Head Loss of Fibrous' Reactor Insulation," NUREG/ CR-2982, SAND 82-7205, l
Alden Research Laboratory, November, 1982.
l 4.3 Wysocki, J.; Kolbe, R.,
" Methodology for Evaluation of Insulation Debris Ef fects," NUREG/ CR-2791, SAND 82-7067, Burns & Roe, Inc., September, 1982.
4.4 Final Safety Analysis Reports, Limerick Generating Station, Units 1&2 FSAR Vol.4 and Vol.7.
4.5 Owens-Corning Fiberglass Corporation, " Topical Report OCF-1, Nuclear Containment Insulation System, NU'K'ON,"
January, 1979.
4.6 Telecon between Messrs. Serkiz (NRC); Tutton, Phillabaum (PECo); Schlueter, Lewis, Blakely, Klein, and Bielanowski (Bechtel); July 21, 1983.
4.7 Telecon between Messrs. Pinsky (Owens-Corning) and 1
Bielanowski (Bechtel); July 14, 1983.
{
}
4.8 Bechtel, " Design for Pipe Break Ef fects," Topical
)
Report BN-TOP-2, Revision 2, May, 1974.
(
l l
l P-261(b)/8 8/19/83
i l
f 11
!ll' TABLE 3.2 - SUP9tARY OF CAICUIA'IED DEBRIS HEAD IDSSES AND EFFECTS ON ECCS PLMP NPSH l
RUPMING
'IUTAL TDtP STRAINER b'IRAINER MIN.NPSHA HEAD IDSS NPSHA NPSH I
PLMPS FIIM SURFACE BIDCKAGE -
WI'IH CLEAN BIDCKED WI'IN BIDCKED REQUIRED AREA
'INICKNESt.
STRAINER 5 TRAINER STRAINER GPM
'F ET2 IN FT FT FT P1' r
g l
4xRHR PUMPS
}
4x10,750 = 43,000 212'F 105.7 1.7 27.0 6.9 20.1 4.0 (1) i 4xCORE SPRAY PUMPS 4x3,950 15,800 212*F 54.8 1.2 28.0 2.5 25.5 10.0 (1)
=
i NOIES:
(1) Value frca certified ptmp performance curve I
U i
P-261(b)/7 8/19/83 l
I s
-. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _