ML20215H327
| ML20215H327 | |
| Person / Time | |
|---|---|
| Site: | Beaver Valley |
| Issue date: | 04/13/1987 |
| From: | Carey J DUQUESNE LIGHT CO. |
| To: | NRC |
| References | |
| 2NRC-7-078, 2NRC-7-78, TAC-62908, NUDOCS 8704200267 | |
| Download: ML20215H327 (11) | |
Text
g 31 @
M 2NRC-7-078 Beaver Val:ey No. 2 Unit Project Organization S.E.G. Building Telecopy (
32 Ext.160 P.O. Box 328 April 13, 1987 Shippingport, PA 15077 United States Nuclear Regulatory Commission ATTN:
Document Control Desk Washington, DC 20555
SUBJECT:
Beaver Valley Power Station Unit No. 2 Docket No. 50-412 SER Confirmatory item #25 Containment Sump 50% Blockage Gentlemen:
Please find attached an evaluation which shows that little insulation or other material would ultimately be deposited in the containnent sump as a result of a failure of a high energy line in the containment.
This provides the required submittal necessary to close SER Confirmatory item 725.
If you should have any questions, please contact Mr. A. N. DiCesaro at (412) 393-7742.
DUQUESNE LIGHT COMPANY By vd. M. Carey 7 Senior Vice President AND/ijr NR/AND/ SUMP /BL K Attachment AR/NAR cc:
Mr. P. Tam, Project Manager
- w/ attachment Mr. J. Beall, NRC Sr. Resident inspector
- w/ attachment Mr. L. Prividy, NRC Resident inspector
- w/ attachment INPO Records Center
- w/ attachment i
9 8704200267 870413 9i PDR ADOCK 05000412 E
United States Nuclear Regulatory Commission SER Confirmatory item #25 Containment Sump 50% Blockage Page 2 COMMONWEALTH OF PENNSYLVANIA )
1 SS:
COUNTY OF BEAVER On this /[N day of
/
, before me, a
/>
Notary Public in and for said Comdnwealth and County, personally appeared J.
J. Carey, who being duly sworn, deposed and said that (1) he is Senior Vice President of Duquesne Light, (2) he is duly authorized to execute and file the foregoing Submittal on behalf of said Company, and (3) the statenents set forth in the Submittal are true and correct to the best of his knowledge.
tb $
$0Wb I
Notary Public g,m roas 4 % %'
$94]P r
i l
l l
l l
ATTACHMENT A.
POTENTIAL FOR PIPE RUPTURE / FORMATION OF INSULATION DEBRIS The comments concerning the performance of the BVPS-2 containnent sump which appear in the original issue of the BVPS-2 Safety Evaluation Report (NUREG-1057) and SER Suppl enent No. 2 do not take into consideration advances in the general understanding of the potential for pipe ruptures in the containnent.
In May,1986, the NRC anended General Design Criter-ion 4, Appendix A to 10CFR50, permitting the use of analyses to renove pipe whip restraints and jet impingement shields for postulated primary coolant loop pipe breaks in presstrized water reactors.
Further rulenak-ing is in process to extend the use of these analyses to other high energy piping systens.
Because the NRC accepts the general notion that a pipe will leak before it breaks, and that a pipe is therefore not likely to break without warning, many restraints are now seen to be unntecessary.
Duquesne Light Company has pursued the implenentation of a leak-before-break (LBB) approach to pipe rupture in the reactor containment for both the primary loop and balance-of-plant stainless steel piping. A schedular exenption has been received from the NRC for the primary loop piping and a final report for most of the balance-of-pl ant piping was submitted in early 1987 and has been accepted by the NRC staff.
The LBB progran is based on the fact that austenitic stainless steel pip-ing will develop a stable through-wall crack and leak in an easily detect-able manner for a long period of time (often years) prior to growing to the point of instability and subsequent rupture.
Tests conducted by EPRI in late 1986 indicate that detectable leaks in pipes (insulated wi th several alternative types of insulation) do not result in a significant loss of insulation.
Instead, the fluid passes through seans in the insulation and exits to the environment with no sig-nificant danage to the insulation.
As a result, it is expected that little insulation or any other fragile plant material would ultimately be deposited in the sump as a direct result of through-wall leakage of stean or water from high energy lines.
i B.
ALDEN RESEARCH LABORATORY TEST OF BVPS-2 CONTAINMENT Sul;P MODEL l
The original issue of the BVPS-2 Saftey Evaluation Report contains the l
following statenent:
i the applicant reports ( Alden Research Laboratory,1983) that initially, as the RWST water is being discharged to the containnent via the quench spray systen and energency core cooling systen, flow velocities on the floor of the containment muld be on the order of 1.7 f t per sec.
l The applicant further states that the velocity distribution would be such that reflective metallic type insulation (used on most of the piping in
{
the containnent) would not be transported to the sump.
However, the l
Page 1 of 5 t
e Alden Research Laboratory reported in NUREG/CR-3616 that floa velocities well below 1.7 f t/sec. are capable of transporting the various component parts of this type of insulation to the sump structures (See attached Table 4-1 of NUREG/CR-3616).
In light of this, the staff recommends that the applicant provide a debris-generation and tr ansport analysis that describes the transient behavior of the sump as the water level in the containment is rising to justify the acceptability of the 50 percent sump blockage assiription throughout the accident."
The first sentence of this statenent is incorrect and is apparently due to a misunderstanding of the BVPS-2 containment sump model tests.
The BVPS-2 test progran was designed to ensure that no air-entraining vortices muld form, that head losses across the screens and in the inlet would be accep-t ab l e, and the swirI in the pump suction pipes would be minimal.
To achieve this end, test configurations and conditions which would simulate the most severe flow patterns, potential for air entr ainment, and head losses were studied.
it was never the intent of these tests to attenpt to realistically model water flow velocities on the floor of the containnent for the purpose of assesssing the potential to transport debris to the sump.
For exanple, the test which gave rise to the flow velocity cited in the SER excerpt was for a configuration where height of water at the sump screens was almost one foot lower than it would normally be when punping fran the sump is initiated. Again, this extrenely conservative configura-tion was enployed for the explicit purpose of assuring adequate punp performance under the most extrene conditions for potential vortexing and head loss across the screens.
To imply the water flow velocities on the floor of the containment model which were measured during this test are even renotely representative of the flow velocities expected during an actual accident scenario is completely erroneous.
Analysis of the poten-tial for tr anspor t of insulation debris to the containment sump is discussed below.
C.
POTENTIAL FOR TRANSPORT OF INSULATION DEBRIS / SUMP SCREEN BLOCKAGE As noted in Section A, advances in the understanding of the potential for a catastrophic rupture in the primary loop piping indicate that this is not a credible event. Likewise, the potential for containment sump block-age as a result of postulated secondary systen line breaks is not a safety issue because the use of the recirculation systen to mitigate the effects of such breaks is not required.
if one were to postulate the possibility of a catastrophic rupture in the primary loop piping, analysis has determined that the " worse-case" scenar-to would be a break at the stean generator hot leg nozzle in stean genera-tor cubicle "C",
i.e., such a break would have the maximum potential to generate insulatinn debris and would be in the closest proximity to the containnent sump.
Although four basic types of insulation are used in the containment build-ing, practically all of the containnent insulation which could be impacted by jet impingement forces from ther " worse-case" pipe rupture is reflect-ive metallic insulation.
Each individual unit of this insul ation is composed of an outer cover, an inr.er cover, and end covers encapsulating a number of stainless steel foil sheets held in place by crimped separators.
The outer, inner, and end covers are joined together by spot we1ds; the crimped separators av e spot welded to the inner foils.
Design details permit tight interlocking of adjacent sections of the assembled insula-tion.
The individual etenents of the insulation are fabricated fran Type Page 2 of 5
__________________________________-___.___J
304 austenitic stainless steel sheets with the following thicknesses:
Outer Cover:
0.037 inch inner Cover:
0.015 inch Inner Foil:
0.0025 inch Crimped Separators:
0.0025 inch The reflective metallic insulation is used to cover the majority of the reactor loop piping, reactor coolant pump, pressurizer vessel, the lower portion of the steam generator up to the tube sheet circumferential weld, and a portion of the reactor vessel.
This includes piping and equipment inside the reactor coolant pump cubicle between the operating floor eleva-tion 740'-3" and 718'-6",
the prime target area for the jet from the
" worse-case" pipe rupture.
A study of the transport and screen blockage characteristics of reflective metallic insulation materials has been performed for the Nuclear Regula-tory Commission by Alden Research Laboratory.
This study, NUREG/CR-3616, is cited in the BVPS-2 SER as the basis for a concern that more than 50 per cent of the BVPS-2 sump sreens could potentially be blocked by metallic isulation debris in the af termath of a LOCA event, in fact, the results of this study confirm the BVPS-2 position that there is minimal potential for transport and subsequent blocking of the sump screens.
The key find-ings of NUREG/CR-3616 are summarized as follows:
Transport Velocities a.
Single sheets of thin stainless steel foil used in reflective metallic insulation (0.0025 and 0.0040 inch thick) can be transpor-ted by water flow velocities as low as 0.2 to 0.5 f t/sec.
Single sheets of thicker foil (0.008 inch) required higher velocities for transport, about 0.4 to 0.8 f t/sec.
b.
Crunpled foils tend to transport at lower velocities than uncrumpled foils, c.
Tr ansport velocity tends to increase with material thickness, except for easy flexible foils where the thickness dependence is
- smaller, d.
In all cases, the velocity of motion of the sample is much lower than that of the flow.
Transport Modes a.
Transport at the lower velocities occurs when the foil sheet is flexible enough that a corner or edge can be bent up by the flow thereby increasing the frontal area and therefore the drag.
The resulting motion is one of intermittent folding, tumbling and roll-ing.
b.
Rigid pieces tend to be transported by sliding along the bottom.
Rigidity can result from greater thickness (0.008 inch and above) or small size (less than about 12 inch by 12 inch for 0.0025 inch foil).
Higher flow velocities are typically needed for transport of rigid pieces than for transport of flexible pieces.
Page 3 of 5
c.
Even with high flow velocities (about 2 f t/sec) and large water depths (60 inches but velocity only 1.6 f t/sec.), the sanples were never observed to become " water borne," i.e, to loose contact with the bottom.
d.
The vertical side walls of the test flume were observed to hinder the transport of sanples.
Sanples entering in contact with a wall were of ten pushed and folded against it, needing higher flow velo-cities to be dislodged.
e.
When several pieces of foil were released simultaneously, their interaction during the transport process often caused janming and immobilization of the pieces.
High flow velocities, up to 1.8 f t/sec., were then required to break up the jams and resume the tr anspor t.
Blockage Modes a.
Most foils readily flip vertically against the screen upon arriving there.
Whether originally crumpled or not, the foils become flattened against the screen by the water force except the thicker foils (0.008 inch) which renain crumpled.
The more flexible foils become folded on the screen, blocking less than their surf ace area.
The large 0.0080 inch thick foils which exhibited rigidity relative to their transport mode but whose dimensions were larger than the water depth of ten fo lded on the screen, a portion being caught under the trashrack.
b.
Because insul ation specimens never became " water borne," they never blocked the screen above their width or length. Blockage up to the diagonal height was never observed but this may be due to the fact that the water depth was less than that height.
c.
When several foil pieces were released simultaneously, significant overlap was observed on the screen so that even if the total foil area was larger than the screen area (1.6 to 2.2 times), the screen was never fully blocked (only up to about 80 percent of area blocked).
With these test results in mind, it is instructive to review the sequence of events which would occur fo l lowing the " worse-case" ruptur e.
In order to become a potential sump blockage concern, metallic insulation dislodged by the break jet must first reach the floor of the containment structure.
NUREG/CR-3616 notes that the danage modes of reflective metallic insulation assenblies exposed to a high energy break jet are not easily determined and therefore, debris could vary from undanaged insulation assenblies to individual components of the assemblies either relatively undanaged or crumpled.
Insulation assen-blies which are essentially undamaged when dislodged or which maintain their structural integrity to any significant degree are extrenely unlikely to ever reach the floor of the containment.
Structur al features and various gratings between the target area and the floor level effectively function as " trash collectors" since debris must traverse an extrenely tortuous path to elude these impediments.
Clearly, the larger the individual pieces of debris, the less likely they are to avoid restraints; the debris most likely to reach the containnent floor would be small pieces which could fit through gratings.
Page 4 of 5
Any metallic insulation debris which does reach the containment floor would generally be deposited in an area approximately 36 feet from the containment sump screens. As indicated in BVPS-2 FSAR Section 6.2.2.2.1, the containment's quench spray pumps are activated upon receipt of a CIB signal and are effective 87.6 seconds after a design basis LOCA.
The recirculation spray pumps are started automatically approximately 625 seconds af ter receipt of a CIB signal (FSAR Section 6.2.2.2.2).
Note that for the period of nine minutes between initiation of quench spray and recirculation spray, water flow with any signif-icant velocity would generally be in directions away from the sump screens due to the direction in which the containment floor is pitched.
The exception to this genral rule is the two foot wide trench running from the center of the containment to the sump.
Although the analysis of potential sump blockage did not attempt to take credit for this effect, the general flow away from the sunp as water from the break and the quench spray initially collects on the contain-ment floor and the expectation that such flow would tend to force individual pieces of debris together couldn't help but to decrease the potential for subsequent debris transport to the sump screens when recirculation flow is initiated.
With initiaton of recirculation flow, the maximum water flow velocity in the general area where insulation debris could f all to the containnent floor has been calculated to be 0.26 feet per second.
This maximum flow velocity would be attained with both 100 percent capacity recirculation trains in operation.
If only one 100 percent capacity train were in operation, the flow velocity in this area would be 0.18 f eet per second.
Based on the data reported in NUREG/CR-3616 (see TabIe 4-1 attached), no metal 1ic insul ation in any form could be transported when one of the 100 percent capacity recirculation trains is in operation.
If both 100 percent capacity trains are in operation, - the only type of metallic insulation debris capable of being transported are single sheets of the 0.0025 inch thick inner foils.
Clearly, a single sheet of the inner foil material is incapable of producing any significant blockage of the sump screens.
If one postulates the presence of multiple sheets of the inner foils on the containment floor, the following quote from NUREG/CR-3616 indi-cates that there is not credible potential for transport to the BVPS-2 sump screens:
" Interaction of the individual foils, however, tended to create jams in which foils would interlock and stop moving.
A high flow velocity, up to 1.8 f t per second, was then needed to dislodge the jam and for transport to the screen to resume".
In summary, there is no credible accident scenario requiring use of the recirc-ulation spray system that could produce and transport any significant anount of insulation debris to the BVPS-2 containment sump screens.
Page 5 of 5
TABLE 4-1 TRANSPORT AND BLOCKAGE TEST RESULTS For 10 inch Pipe Insulation Assembly Velocity to Velocity to initiate transport Sample motion to screen Comment s Description (ft/sec)
(f t/ sec)
Undamaged unit (half assenbly) normal to flow concave side up 1.0 1.0 Either flipped on screen-(see Figures 8-a and 9) or got stuck partially flipped (Figure 8-b).
concave side down above 2.2 Never moved.
For more details, see See Section 4.1.1..
Outside Cover (0.037" thick diameter - 19")
concave side up 0.7 0.8 Same blockage mode as undamaged units.
concave side down above 1.8 For more details, see Section 4.1.2 Inside Cover (0.015" thick Di aneter = 13")
i concave side up 0.7 0.8 With both initial concave side down 1.1 1.6 positions, covers flipped the screen on arrival and j
got flattened against it by the flow force.
For more details, see Section 4.1.3 l
l End Covers above 2 Never moved.
Page 4-1.1 l-L
Velocity to Velocity to initiate tr ansport Sample motion to screen Comments Description (ft/sec)
(ft/sec)
Single sheet 0.35 0.5 Moves in folding and tunbling inner Foil mode (Figure 10).
Flips (0.0025" thick) vertically against screen as uncrumpled soon as it reaches it, with and without (Figure 11) May be folded on separating crimp
- screen, i.e., not cover full sheet area.
Never covered screen higher than maximum sheet dimension, even for flow velocity of 2 ft/sec, and water depth of 60 inches.
For more details, see Section 4.1.5-a.
Single sheet 0.20 0.25 Moves in folding and tumbling mode.
Flips against screen as soon as it reaches.
Gets flattened on screen by current.
For more details, see Section 4.1.5-b.
Four sheets 0.25 0.4 to 1.8 When numerous foil sheets inner foil are used they tend to jam up (0.0025" thick) in piles that may need two crumpled high velocity to unjan, two uncru.r91ed Significant overlapping on screen.
i For more details, see Section 4.1.5-c.
i Page 4-1.2
Velocity to Velocity to initiate tr ansport Sample motion to screen Comments Description (ft/sec)
(ft/sec)
Single cut-up sheet inner foil (0.0025" thick 24" x 21") uncrumpled 0.20 0.25 Folding and tumbling transport mode.
Flip vertically on screen upon arrival, sometimes folded.
Flip vertically on screen Crumpied 0.20 0.25 upon arriyal, sometimes folded.
For more details, see Section 4.1.5-d.
Several cut-up sheets inner foil (0.0025" thick 8" x 8")
uncrumpled 0.5 1.2 Pieces not folded by flow as larger ones.
Sliding transport mode.
One piece reached screen at 0.5 f t/sec - all flipped vertically on arrival to screen.
crumpled 0.5 l 1. 2 One piece reached screen at 0.9 f t/sec - all flipped vertically on arrival to screen.
For more details, see Section 4.1.5-d.
Several cut-up sheets inner foil (0.0025" thick 3" x 3")
uncrumpled 0.8 2.0 Pieces not folded by flow as larger ones.
Sliding transport mode Page 4-1.3
Velocity to Velocity to initiate tr anspo rt Sample motion to screen Comments Description
,(ft/sec)
(f t/ sec)
Several cut-up sheets inner foil (0.0025" thick 3" X 3")
(continued) crumpled 0.6 1.0 Pieces flip vertically on screen unless a corner gets trapped under screen bottom, in which case the piece stays flat on bottom
.For more details, see Section 4.1.5-d.
Page 4-1.4
. - -..........