ML20198Q879
| ML20198Q879 | |
| Person / Time | |
|---|---|
| Site: | Brunswick  |
| Issue date: | 09/19/1997 |
| From: | Biasca R, Hart G, Nuzman C PERFORMANCE CONTRACTING, INC. |
| To: | |
| Shared Package | |
| ML20198Q855 | List: |
| References | |
| NUDOCS 9711130072 | |
| Download: ML20198Q879 (24) | |
Text
-
R.PCI
.t'F.R.I.RML.^c.?.,%
Summary Report on Performance of Performance Contracting, Inc.'s Sure-Flow
- Suction Strainer with Various Mixes of Simulated Post-LOCA Debris Revision 1 b7111300Z3971106 e* qf('
gm sua0500g4 Revised by Richard A. Biasca September 19,1997 g
PROFotMM NW' w .a
Sussmary Report on the Perforsmance of Performance Contracting, Inc.'s S:re-Flow" Suction Strainer with Various Mines cf Simulated post-LOCA Debris, Rev.1 TABLE OF CONTENTS Page No.
Suamary of tbe Performaacc Evaluation................................................. 2 L Description of the Tested Sure-Flow Strainer Prototype.......................... 5 IL Description of the Test Facility................................... ...................... 6 Ill Tes t Res u lts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
A. Bart St rain er H ead Loss. ... . .. . ... .. . .. . .. . .. . . . . ... .. . .. . .. . .. . ... ... .. . .. . .. . 7 B. Strainer IIcad Loss with Simulated Debris.................................. 9 C. A pparent Filtration Efficieacy................ ...... ......... .... .. ...... ......! 4 i
D. Fibro u s Bed Co m paction. .. . . .. . . . . .. . .. . .. . .. . .. . .. . .. . .. . .. . ... .. ... ... ... . ... ! 4 l
IV. Regression A nalysis of Test Data. .......... ... ......... ......... ..... . ... .. ...... ....I 5 l
l V. Co n cl u s i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . .. 21 i
R e fe r e n c es .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
i l
l 1
L
Summary Report on Ce Perf;rmance cf Performance Contracting,Inc.'s Srre-Flow Suetion Strainer with Various Mixes of Sim: lated post-LOCA Debris, 09/19/97 Rev.!
SUMMARY
OF THE PERFORMANCE EVALUATION ,
The Sure-Flow" Suetion Strainer has been designed and developed specifically for attachment to the Emergency Core Cooling System lines on Boiling Water Reactor nuclear plants. The strainer is intended to reduce the post-LOCA Head Loss across the entrance to the ECCS line with the purpou of maintaining ECCS pump flow at the design value. To accomplish this, these strainers are also designed to be connected to the ECCS lines in suppression pools upstream of the ECCS pumps. High performance strainers se:h as these have been shown to reduce head loss resulting from the collection of LOCA generated debris and other materials that can otherwise block existing small, passive stainers. The Sure-Flow Strainers can alleviate that problem and thereby keep water flow at design values through the ECCS lines.
To evaluate the Sure-Flow Strainer's performance, a prototype was fabricated by PCI and tested at the Electric Power Research Institute by the US Boiling Wster Reactor Owners' Group in December,1995 as part of their strainer testing program (Ref.1).
Those tests used shredded NUKON' fiberglass insulation to simulate post LOCA fibrous debris and they used iron oxide particulate to simulate suppression pool sludge.
However, in these 1995 tests, the quantities of fibrous debris were limited to relatively low quantities. Therefore, to evaluate the perfonusnee of this strainer with larger quantities of fibrous debris, additional tests were conducted at EPRI in October,1996. In addition, another test was conducted with stainless steel feil, shredded to simulate foil debris froni Reflective Metallic Insulation (RMI), and then combined with fibrous debris and particulate.
Combined, these two sets of Head Loss performance tests showed the following behavior of the Sure-Flow Strainer prototype with this debris mixture:
. The bare strainer (i.e., with no debris) showed a low Head Loss behavior and that Head Loss is dependent on the entrance (i.e., at the strainer's nozzle) water velocity; e the debris Head Loss behavior is essentially linearly dependent on both Mass of Fibrous Debris and Water Flow Rate for 25 lbs. of more of fibrous material; e the addition of 100 lbs. of CP particulate increases Head Loss across the strainer significantly; e the Head Loss behavior can be accurately modeled with regression equations, developed from the test data, and applied over the tested range of those variables, namely Mass of Fibrous Debris and Water Flow Rate; e addition of stainless steel foil fragments, which simulate Reflective Metallic Insulation debris, increased the Head Loss across the Sure-Flow Strainer by less than 1.0 fl of water; e thick fibrous debris beds exhibited an effective filtration efliciency greater than 90%;
2
I
! Summary Report on the Performance of Per.'ormance Contracting,Inc.'s Sure-l Flow Suction Strainer with Various Mixes of Simulated post LOCA Debris, Rev.1 09/19/97 e on this strainer prototype, the fibrous debris beds exhibited an apparent bed i compaction of approximately 24% (using the as fabricated insulation density as a reference); ,
e the Sure-Flow Strainer, mounted in a horizontal position, did not vortex, even when the tank was drained so that the strainer was about half exposed above the water level.
The particular strainer prototype was tested on 124 inch NPS line and it had certain l y
j geometric features:
1
- 170 ft2 total surface area ofperforated plate e 24 inch NPS attachment flange and Internal Core Tube
. . 40 inch outer diameter i e 48 inch active length and a 54 inch total length e 56 A' of circumscribed cylindrical surface area, including the ends l
i e thirteen disks with a width of 1.85 inches each l l
- twelve gaps (between the disks)with a 2.00 inch width and a total volume of 8
i about 10.3 A l
- holes in the Internal Core Tube that are smallest at the flange end and largest
- at the opposite end; these are sized with a linear distribution, over the length ;
of the Internal Core Tube, so as to provide equal Water Flow Rate from disk l
l to disk and hence uniform water flow over the strainer's length.
l t1
)
'j i . w , . .
m , _ , !
NT, .-. xA j
,i .s l h ,K.
Ei,
! V & - ::,
- 1 ,. \
- \
1 i.' j st!
ii 1
i . .. , ,
\ \
'y' li [*1 - 'l
\
/ j, /
- ; o; r.. ,
q \ \
l .
f
[,' /
j i
x
,5l,I s ,
/ .
l <&, ' .' l o
Figure la - Photograph of the Sure-Flow" Stralner 3
- H T
I S,E L
E A n - R T V .S i
o , EK t .
c NE NH E TN -
OT F O' u . BA F' B' 2 S .
RL7 I8 R^ , u w AP04 T / A' c -
o C S1C" S t D o l
F A.E D N C D e1
. G TA A ""- C 'c r
1O S 1 R 7 E +
- u. v FE s _ n Se S,R LA 3,
R sR
_ 4, E
]
KEOE V n SPHR
~
I
'c s, I A .. E I
N +
c nir DL # . I I
gD n
,b e l LE'8 E
LT / Po-AS1D N
E- bN , -
. i 0
8 1
V D
N A
R TR
+
o o
i tA . ' x N
E SNE I n,
s cC A f aD ) " T R
+
n rL R t -i
- t. E WST ~
n e
C r
Fs p V
N OS 2 - n AS '#g LETO.
e cd (E K S CC 8 F EN e
a n
n at no B C
A PS ECY S e
ro un t 1 T o R .TOT _ J" i T
fi E
R S 2' 1 4 UtLO UR w rS e
U G
I 5 P2 3 SFP -~ f n
P fo F 8 Y - -
r 1 T _ ._ - _
f _- : o N
o se _
O I
r e
S e eix __ :
_ M T cM 2! __. -
V D n P n r.:: ,->
E_ S E
as A __ :
W
+h nu 8' G P.C __
- _ I t ro 4 YA - : V 6 '
f oir 2' (PTL _.
_~
- _ 5 .:
s n
r a -
t
~ _
E D :
D eV _
_~
_ I E
R E : n P h _-
i S E -
- i n
- N w et i _
I G
N n hV t _ \. _
- E 4,
r ne r \ n on i .
ta '
h 1
r r 6 ot NE o pS OT c e , BA e s
R DE .E RL n ER E P S. T AP - i y TOI CAL C f,, .- .
r r
o ACP'P R 4 e L
" E g .y "
n OLE / 2 E -.,
S n
u FAB 1 RNU ERT PE 8
/ T 1
S ;g,7r T
=,
N I _.y. "
l R;;e" 4 ,' 4 l 1l . ,. !
S:mmary Report on Ce Perf:rmance cf Performance Contracting,I:e.'s S:re-Flow Suetion Strainer with Various Mixes cf Sim: lated post-LOCA Debris, 09/19/97 Rev.1 ,
While only one strainer pmtotype was tested at EPRI as part of the reported testing program, the results of the testing program can be used to verify a general predictive model that can then be used to predict the behavior for other sizes, with other Water Flow Rates and debris quantitles. One such model, for predicting Head Loss across a fibrous and particulate debris bed, has been developed by Science and Engineering Associates (SEA) for the United States Nuclear Regulatory Agency. It is based on the one-dimensional, flat plate filtration equations for flow resistance (i.e., for Head Loss) and is developed in Appendix B of NUREG/CR-6224 (Ref. 2). This PCI report suggests how this can be done empirically (i.e. through a regression analysis of some of the test data) to account for the gaps, between the disks, that fill with fibrous debris and thereby change the disks and gaps to a single, large cylinder which increases in diameter and length with increasing quantities of fibrous debris.
I. DESCRIPTION OF THE TESTED SURE-FLOW STRAINER PROTOTYPE The tested Sure-Flow Stralner prototype consists of a series of coaxial stacked disks that are equally spaced and mounted on an Intemal Core Tube. Figure la is a photograph of the tested strainer prototype and Figure Ib is a mechanical drawing. The Intemal Core Tube is a pipe with holes in it. These holes are spaced so as to line up with the disks and l
the gaps (between the disks). They have a varying size: those holes closest to the flange end of the strainer are smallest and those at the far end of the strainer are largest, with those in between sized linearly so as to provide approximately equal water flow from l
disk to disk and from gap to gap. In addition, the Intemal Core Tube is designed to keep highly turbulent water flow (and thereby all high velocities) in the tube itself, away from the disks and the debris. PCI believes that this feature helps prevent compaction of the collected fibrous debris, thereby preventing an even greater increase in Head Loss. ,
For structural reinforcement, each disk on the prototype has six (6) intemal stiffener plates. These plates are radially oriented and are welded to both the core tube and the inside of the disks. The disks are all fabricated on their exterior from perforated metal plate. For this prototype, that material is 11 gauge steel with 1/8 inch diameter holes, spaced so as to give a 40% open area. Fabrication is such that each disk consists of two sheets of perforated plate welded to the Intemal Core Tube and one strip of perforated plate welded to the outside of the two sheets like the edge of a wheel, thereby creating a disk. Each disk is 1.85 inches wide and each gap is 2.00 inches wide. The outside shape i
of the strainer is that of a cylinder that is 48 inches long and 40 inches in diameter, all I mounted around a 24 inch outer diameter Internal Core Tube with % inch thickness. The 24 inch NPS flange is welded onto the end of the Intemal Core Tube, six (6) inches from the first disk. Therefore, the overall strainer length is 54 inches (i.e.,48 inches of active strainer plus 6 inches of attachment pipe). See Figure Ib.
Duke Engineering and Services, Inc. (DE&S) performed a structural evaluation of this I strainer prototype and determined that its stmetural integrity was adequate for the Head 5
S:mmary Report on Ce Performance cf Performance Contracting,1:e.'s S:re- l Flow" Suetion Strainer with Various Mixes cf Sim: lated post LOCA Debris, l Rev.1 09/19/97 Loss tests up to, but not above, about 19 feet of water. When that Head Loss limit was !
reached during a test, in some cases the wa'er flow rate was reduced to keep the pressure ditrerential acrou the strainer to less than or equal to that value and thereby allow us to collect data. It is notewonhy to point out that this pressure differential limit of 19 feet is actually approximately equal to the specified post LOCA maximum hydrodynamic pressure for strainers at several BWR plants. PCI has supplied strainers to several BWR ,
nuclear plants and, for structural robustness, these have been designed with greater thickness steel for the Inteme.1 Core Tube , the addition oflongitudinal stiffeners for the Intemal Core Tube, and the addition of more intemal disk stiffeners. These additional ,
structural reinforcements will not interfere with either the internal water flow or with the extemal debris collection and that has been a very important design consideration is reinforcing the Sure Flow Strainers for these nuclear plants.
it is important to point out that the tested prototype is a bolt-on, cantilever strainer, designed to be supported by a BWR's attaciunent ECCS pipe in a radial orientation. ,
DifTerent plants may require other structural mechanisms, depending on the strainer size.
For example, one nuclear utility has ordered two seventeen (17) foot long strainers, each containing three tee pipe connections. These long strainers are designed to be supported like a beam, at each end, by a pair of ring girders. The ECCS pipe connections will connect to the tees such that these pipe are oriented at 90 degrees with the strainer axis rather than in line with it. Nevenheless, the Intemal Core Tube remains the basic structural" backbone" of this long strainer and each disk must be intemally reinforced, just as it would be for a bolt-on, cantilever design strainer.
1 II. DESCRIPTION Olc THE TEST FACILITY The tests at EPRI were all conducted by Continuum Dynamics, Inc. The 1995 testing, sponsored jointly by the BWROG and PCI, is summarized ir theB. WROG's Utility Resolution Guidance, or URG (Ref.1). The EPRI test report (Ref. 3), gives the results of those tests sponsored by PCI and conducted in 1996. Chapter 2 provides a description of the test facility and the test procedures. The same facility was used and the same procedures were followed for both the 1995 and the 1996 performance tests at EPRI.
There was m control over water temperature so that temperature fluctuated and was therefore a little different on different days.
It is significant to point out that the test configuration and strainer mounting at EPRI included a 180' tee, a 90' long radius elbow, and several feet of straight pipe between the two pressure tunsducers used to measure Head Loss across the strainer. Therefore, the Head Loss measurements for the bare (i.e., with no debris) strainer, while relatively low, still included a pressure drop across these piping components. In this paper, a correction is made to determine the Corrected Head Loss across the bare strainer (i.e., the Corrected Head Loss is the pressure drop across the strainer only, without losses associated with these piping components).
6
5:mmary Report on Ce Performance cf Perf:rma:ce Contracting, I:c,'s S:re-Flow ~ ;ction Strainer with Various Mines cf Sim: lated post LOCA Debris, 09/19/97 Rev.1 III. TEST RESULTS The 1995 test results are provided in the PCI report, The Development and Testing of Performance Contracting, Inc. 's Sure-FlowTM Stacked Disk Suction Strainerfor B WR ECCSLines, February 1,1996 (Ref. 4). Those 1995 test results are also contained in Appendix B of the BWR Owners' Groun's Utility Resolution Guidance. Rev. 0 (Ref.1).
The 1996 tests results are provided both in the EPRI test report (Ref. 3) and in the PCI Memo for Record, Qn bdication ofStrainer Testing at EPRI, November 11,1996.
Note that the water temperature for these 1995 tests was 58' to 60' F and that it was 69' to 70' F for the 1996 tests. Because water viscosity is dependent on water temperature, and since viscosity is lower for the higher water temperature, a correction will be made for those licad Loss values across debris beds.
PCI performed a QA dedication of both series of tests and therefore documented the results separately from EPRI. These results and those reported by EPRI were very close in value. For the purposes of this report, PCI has used their own readings rather than those reported by EPRI. All were performed under a nuclear Quality Assurance Program.
Bare strainer testing was performed at the Fairbanks Moise Pump Company. A protocol for testing the bare strainer head loss for the Sure Flow Straine. made by Performance Contracting, Inc., was accepted for the tests conducted at the hydraulic laberatory testing facility of Fairbanks Morse Pump Company located on Fairbanks Drive in Kansas City, Kansas on April 16,1997. Report Number PCI-NPD-CE03," Tests Conducted at the Fairbanks Morse Pump Company To Detennine the Head Loss Performance of PC1's Bare Sure Flow Strainer"(Ref 7), documents the data and testing performed. From the data taken during this testing, a head loss correlation was generated that can be used to detennine head losses on other similar strainers.
A. Bare Strainer IIcad Loss To determine the clean strainer head loss across the PCI Sure-Flow
- Strainer, tests were perfonned on the Prototype 11 test strainer (Ref. 7). These included a second set of tests across the connecting pipe without the strainer. The results were then corrected by subtracting both the head loss due to friction in the straight pipe and the dynamic weter I head at the point of measurement. These corrected strainer head loss values were then used to develop a regression equation and a curve for bare strainer head loss as a function of strainer wnter entrance velocity. This equation contains conservatism as it does not take credit for observed strainer behavior, namely a discontinuih in the test data, which results in lower head losses at entrance velocities in enn of 6.4 fVsec and at least up to the tested maximum entrance velocity of about 9.1 fbec. This regression equation separates variables so that the head loss is expressed as a function of temperature 7
S:mmary Report on Ce Perf:rmance cf Performance Contracting,I:c.'s S:re- ;
Flow" Suetion Strainer with Various Mixes of Sim: lated post-LOCA Debris, Rev.1 09/19/97 dependent viscous, or laminar, flow plus ina.itial, or turbul'ent, flow, and is shown below.
Figure 1, below, shows the head loss test data, regression generated head loss values, connection pipe head loss values and the curve fit connection pipe head losses.
Equ 1 HLc, = A + K, v V,, + K (8 V,,8 / 2g )
where v = water's kinematic viscosity, fl%ec 2
g = the gravitational constant = 32.2 ft/sec V, = water's strainer entrance velocity, ft/sec A = 0.0022 K, = 1,024 K, = 0.8792 ad HLc., = the strainer's bare, corrected head loss, ft of water.
Figure 2 SURE. FLOW B ARE STR AINER HE AD Lo88 TESTS:
CORRECTED D AT A WITH AND WITHoUT A TEST STR AINER, TESTS CONDUCTED AT FAIRB ANKS MORSE PUMP Co.
l l ! ! ! !
l I
l l I [j t l s S T R AIN E R D A T A l l l 88 , j STR AINER CURVE Fit j j j 3.3. - e CO NN(C TING PIPE D AT A - .
f I
3- - --C O N N C E C TIN G PIP t C U R VE Fit
- I g" g 5.T. , ; f !
fi ' ' ! '
j g ,,. .. -
I '
i i 4
i i I i I I l i i i i l 0 l i s i ! 1
! i l -
= _-
7i
' """~ I """ ~ !
i
.3.- >
m L m 4.-- e ** b ** U '
. i : . . . . T . . n si o u i. ..
E NTR ANC E W ATER VE LoctTY. FTis t C lt should be noted that the Authors believe the use of this straight line extrapolation from the bare, corrected strainer head loss vs. entrance velocity graph is conservative because it bounds the experimentally detemiined values of strainer head loss. Examination of the BWROG test data on other strainers, including PCI's Prototype I and 11 strainers, show a similar behavior pattem.
Total uncertainty (from syrtematic measurement error and randomness) for the test data generated at the Fairbanks Morse Pump Company was detennined from the uncertainty of the data. The tests showed that the data points were very repeatable. Thus, the l
l randomness in the data is small.
8
Sammary Report on Ce Performa:ce cf Performance Contracting,1:e.'s Sure-Flow" Saeties Strainer with Various Mixes cf Sim: lated post LOCA Debris, Rev.1 09/19/9/
B. Strainer Head Loss with Simulated Debris !
J Table i below summarizes all the results for the testing conducted at EPRI on this Sure- ,
l I Flow Straiser prototype. For the ptirposes of distinguishing between the 1995 and the 1996 tests, PCI has used the prefix of"95 " or "96 " to designate the corresponding data.
2 In both cases, the Clean StrrJner Measured Head Losses, referred to in Section A above, were nrst subtracted so that the values reported on Table 1 are for Head Loss across the debris bed only.
TABLE!
Summary of Actual Hoa Loss Test Data from EPRI 1995 and 1996 Measured Data i
All Head Loss Values in Feet of Water INfEIliM2' 95 3 954 95 5 96 2 90 3A 96 3B 96 3C 96 3D 96 3E 96 4 96 5 NIS' FlaERE TT 1 25 3 SO 25 100 150 200 250 300 100 200 (Las.) i MAas CP 85 100 100 100 100 0 0 0 0 0 100 100 (Las.)
AREA 0F 0 0 0 0 800 0 0 0 0 0 0 0 FOIL (FT')
57 58 59 58 69 69 70 71 72 73 69 70 Ep, 2500 gpm 0.58 0.83 0.00 2.29 0.96 4 65 6.15 8.32 9.65 5.40 10.73 3000 gpm' 13.00 3500 gpm 16.67 3750 gpm 1.01 1.46 0.01 3.81 1.60 7.58 10.58 13.66 16.16 8.66 4000 gpm 16.58 19.17 5000 gpm 1.53 2.13 0.16 5.42 2.33 6.08 10.00 13.83 17.75 12.25 6250 gpm 14.75 7500 gpm 1.67 2.42 0.27 8.08 2.95 10000 gpm 1.67 2.58 0.00 10.17 4.34 In most of the tests, the debris was collected on the strainer using a constant 5000 gpm Water Flow Rate, then after its collection, the Water Flow Rate was varied to allow the generation of Head Loss data for other values. However, for Test No. 96 3E (300 lbs of Fibrous Debris and no CP Particulate) and for Test No. 96 5 (200 lbs, of Fibrous Debris and 100 lbs. of CP Particulate), the Water Flow Rate was reduced to 4000 gpm to keep the total Head Loss below about 19 feet of water, judged by PCI's structural consultants (i.e., Duke Engineeririg & Sewices, Inc.) to be the maximum allowable pressure differential across this strainer prototype, in fact, for all of the 1996 tests, this upper li: nit of 19 feci of water was observed by PCI as the maximum allow 2ble Head Loss, thereby preventing the collection of Head Loss data at Ngher than 5000 gpm Water Flow Rates for Al but Test No. 96 2.
9
1 S:mmary Report on Ce Performance cf Performance Contracting,I:c.'s Sure-i Flow
- Suetion Strainer with Various Mixes of Sim: lated post LOCA Debris, 09/19/97 !
Rev.1 l l
l Figure 3 below shows the results of the debris collection, as Head Loss vs. Time, for Tests Nos. 95 3 and 96-2. For the first of these two tests,25 lbs. of Fibrous Debris and l 100 lbs of CP Particulate were added; in the second, those same quantities of Fibrous Debris and CP Particulate were added but also 800 square feet of stainless steel foil, simulating LOCA generated Reflective Metallic Insulation debris, was added. As can be seen, both sets of data followed the same time constant and the Head Loss results, at '
particular times, are close in value, with the addition of the RMI foil increasing Head Loss by about 0.9 feet of water. It can also be seen that for this test configuration, using 5000 gpm and a 50,000 gallon tank, that it took approximately 50 rninutes to reach an equilibrium Head Loss. This pattern was found for all the debris collection tests, for which debris collection was performed at 4000 or 5000 gpm, u is clearly shown in the transient plots included in Reference 4. The 1996 test data, collected using about 70' F water, was corrected for 60* F water temperature in order to compare it to the 1995 test data which was collected with about 60' F water. This same correction, to 60' F, will also be made for other Head Loss data presented in this paper unless noted otherwise.
Figures 4 and 5 (following) show the results taken from the series of Tests 96 3A through -3E. In these tests, no CP Particulate was added to the ten' . Instead, fibrous debris was added to the test tank in increasing quantities, starting out with 100 lbs., then increased in 50 lbs. increments till 300 lbs. was added in total. Figure A-3 from the EPRI report (Ref. 3) shows the sequencing clearly in the transient Head Loss graph.
Figure 4 below shows that Head Loss is roughly linearly dependent on Water Flow Rate, for a given quantity of Fibrous Debris. Figure 5 below shows clearly that H ad Loss is .
approximately linearly dependent on Mass of Fibrous Debris for a given Water Flow Rate through the strainer.
In both cases, there was no CP particulate and the linear behavior can only be considered for the range of the tested variables. However, the essentially linear dependence of Head FIGURE 3:
EPRI TESTS WITH AND WITHOUT SS FOIL:
HEAD LOSS VS. TIME WITH $000 GPM OF ROOM WITH 25 LBS. SHREDDED NUKON & 100 LBS.CP PARTICULATE CORRECTED aFOR 60 DEGREES F WATER 3 :-
M "~
s .. --
~~ ~~
s
~
3 ,, -
/ "
d * , MWrR80,03 1
d
_ s /' I
. . 4. .. .. i. .. .. .. .. .. ..
TIME FRoM START oF TEST (MINUTES) 10
S:mmary Report on Ce Performance cf Performance Contracting,Inc.'s S:re-Flow" Suetlos Straiser wl3 Vcrio:s Mixes of Sin *: lated post LOCA Debris, '
Rey.1 09/19/97 Loss on Water Flow Rate is expected, per the USNRC NUREG/CR 6224 Equations, for predominantly laminar flow through a fibrous bed collected on the strainer. For Head Loss dependence on Mass of Fibrous Debris, the dependence can be expected to be essentially logarithmic with very large quantities of Fibrous Debris. This logarithmic behavior will be explained later in this report. For the range over which these tests were conducted (i.e., up to 300 lbs. of Fibrous Debris on this particular strainer prototype), the ;
dependence of Head Loss on Mars of Fibrous Debris is, as expected from the Head Loss equations in Reference 2, also essentially linear.
For those tests with 100 lbs of CP Particulate, graphs similar to Figures 4 and 5 can be plotted. The four Tests 95v.h 95 5,96-4, and 96 5 all used the same quantity of CP particulate, namely 100 lbs., and 25, 50,100, and 200 lbs. of Fibrous Debris, respectively.
The Head Lots dependence on Water Flow Rate, for the four different quantities of Fibrous Debris, is shown graphically in Figure 6. Again, the dependence on Water Flow Rate is essentially linear. The Head Loss dependence on Mass of Fibrous Debris, for several different Water Flow Rates, is shown in Figure 7. Again in this case, the dependence is essentially linear, as was the case without any CP Particulate.
Other than Test No. 95 2, only two di.Terent quantities of CP Particulate were used for all of these te ts: either 0 lbs. or 100 lbs.
l l
l l
l l
11 i
--- - --,w, - - -. ,.. v . - , - . . - , -
Summary Report on the Performance cf Perf;rma:ce Contracting,Inc.'s S:re-Flow" Section Strainer with Various Mises of Sim: lated post-LOCA Debris, ;
09/19/97 ;
Rev.1 i
FIGURE 4:
MAD LOSS VS. FLOW RATE FOR TM SUREfLOW STRANER WITH FBROUS DEBRIS AND NO CP PARTICULATE MAD LOSS VALUES CORRECTED FOR 80 DEGREES WATER
$ bl***
, //
l* ,, ;
//! A . , - -
// / ;
E ,,,;, // / i x. ,.e-g,; // / /
s //// ; w .... ...
l, // V-f
//x y/ ;
o
- p' ,
t 1000 2000 3000 4000 8000 6000 0
WATER FLOW RATE, GPM N
FIGURE 5:
HEAD LOSS VS, MASS FIBROUS DEBRIS FOF THE SURE FLOW STRAINER WITH FlBROUS DEBRIS AND NO CP PARTICULATE HEAD LOSS VALUES CORRECTED FOR 60 DEGREES WATER to
,,. I i / """ J
, ,, ; i i / I/* *"
! i /t fi
- i V y ! !
1
! ,,,,; I A /I iA-g, > i/ / I4
. g, i A/ I / i I
/A /
g,;_ /M
' \
I i I o l 5 o 50 100 ,50 200 250 300 350 MASS OF FIBROUS DEBRIS, LBS, 12
1 1
S:mmary Report on Ce Performance cf Performance Contracting.1:c.'s Stre- j Flow" S:ction Strainer with Various Mixes of Sim:Isted post-LOCA Debris, l Rev.1 09/19/97 1 FIGUfE 6:
TEST DATA ON PCPS SUREfLOW STRAllER AT EPft:
LEAD LOSS VS. FLOW RATE FOR SEVERAL QUANTITIES OF FIBROUS DEBRIS,WTH 100 LES. OF CP APO $0 F WATER n . . . _
n,_ - - -
,gg&, & V~ ~ % nG ~
- n. _ . so tas.FBR0ts DEBRS
/
M y, _ s 100 LDS.FBROUS DEEM 8
/
g 200L88.FBROUS DEBRE y gg_ AsTested Lets /i h it -
/' '
I
- it 14 -
/" l -
12 f ' I' ' 'LA~
10 / -
- fl s / 7 i
'- _ ;cuemm e / l ./ i
~
Ph' 4 / / I -
menu
- f-f /
Wf 1
s 0- ,
- ' I
'I . .
1 1
"m 0 1000 2000 3000 4000 3000 soon 7000 WATER FLOWRATE,GPM FIGURE 7:
TEST DATA ON PCrS SURE FLOW STRAINER TESTED AT EPRl:
HEAD LOSS VS. M ASS OF FIBROUS DEBRIS,100 LBS, OF C.P., AND 60 F WATER 30 .-
n- -- :it8? W /'
n 24 . .
- !!M .
o 6000 GPM !
!i
/ i e estee Deia Parus ( Regressen Care Woo GPuf
< jo _
22 l l / f [ ' ~ ~op 7 "
' i
/ //~ tc~ "" '~
,s I I / // J _ ____
a ,,
- / // / . x-E 12
/ // / /
",o=
=
A // / /**YX'"
s-
//[// !
a:
I I 2 ~b's o- 8 0 50 100 150 200 250 l
' MASS OF FIBROUS DEBRIS, LBS.
13
S:mmary Report on Ce Performai.ce cf Perforna.1ce Contracting, Ize.'s S:re-Flow" Section Strainer with Various Mixes of Sim: lated post-LOCA Debris, 09/19/97 Rev.1 C. Apparent Filtration Effielency After the completion of all the tests, PCI took water samples to be analyzed for per cent solids. A comparison of these were used to calculate the Apparent Filtration Efficiency of the fiber bed on the strainer. The per cent solid at the beginning of each test can be detennined by taking the total quantity of CP particulate and assume it is uniformly distributed throughout the approximately 50,000 gallons of water in the EPRI tank and i piping system. Table 5 gives the results of tests for Apparent Filtration Efficiency. It is worth noting that these values ranged from a low of 41%, for a very thin layer test, Test No. 95-4, up to over 90% for the thickest layers in Tests No. 96-4 And 96 *.
Table 2: Estimated Values of Apparent Filtration Efficiency for Tests with Combined NUKON Fibrous Debris and CP Particulate 95-3 95-4 95 5 96 2 96-4 96-5 Test No. 95-2 180* 180* 180* 180* 170 160 180 Mass CP at t=0 mg/l mg/l mg/l mg/l mg/l mg/l mg/l Mass CP at test end 51 mg/l 41 m g/l 107 mg/l 35 mg/l 30 mg/l 5 mg/l 12 mg/l Mass CP Removed 129 mg/l 139 mg/l 73 mg/l 145 mg/l 140 mg/l 155 mg/l 168 mg/l Apparent Filtration 72 % 77 % 41 % 81 % 82 % 97 % 93 %
Efficiency
- Indicates a value calculated by using measured weight of CP and estimated volume of water in the tank and piping system.
! D. Fibrous Bed Compaction
(
l A factor which affects head loss across a fibrous bed is the effective bed compaction.
This compaction probably results from the viscous effects of water flowing past the fibers collected on the surface of the strainer. For the purposes of most of the calculations, we use the "as fabricated" insulation density which is 2.4 lbs./ft' for NUKON Base Wool, the i insulant used in the fabrication of NUKON Insulation blankets. To determine the effective bed compaction, PCI measured the bed thickness following Te.ct No. 96 05.
This was done using a long pole and inserting it into the wet bed, following the draining of the tank. Figure 10 is a photograph of this being done. Trds measurement tecimique gave an rpproximate thickness, from the outer edge of a disk to the outer edge of the bed, of 8 % inches. A hand calculation can show that the bed volume was approximately 56 ft' whereas it would have been approximately 73 ft' were there no compaction. This allows for 10.3 ft' of fibers to collect first in the gaps between the disks. Taking the quotient of the actual and the theoretical gives a ratio of 0.76, or a compaction of (1 - 0.76 ) = 24%.
14
S:mmary Report on Ce Perf;rmance cf Pertrma$te Contracting,1:e.'s Stre-Flow" Suction Strainer with Various Mixes cf Sim: lated post-LOCA Debris, Rev.1 09/19/97 IV. REGRESSION ANALYSIS OF TEST DATA Head Loss across the debris bed: Calculations for Head Loss across the debris (i.e.,
combination of NUKON fibers and corrosion product particulate) on the strainers are performed by first developing a regression equation for some of the data given in Table 1.
For the purposes of predicting behavior of this strainer for different water flow rates, different diameters and lengths and hence different surface areas, PCI developed a couple of MS Excel spreadsheet programs. To do this, several assumptions were made:
- 1. Results from Tei,ts Numbers 95 3,95 5,96-4, and 96 5 can be analyzed by regression to determine the Head Loss dependence on Water Flow Rate and Strainer Surface Area. The prototype strainer's tested behavior can then be accurately scaled by dividing its water flow rate and its mass of fibrous (NUKON) debris by its circumscribed surface area, namely 56 ft' where this strainer had a 40" diameter, a 48" active length, and a 24" Intemal Core Tube (N.qjs: for Test No. 96 5, a circular disk was bolted onto the test strainer's end disk, reducing the circumscribed surface cea to 53 ft'). This assumption is conservative since as more debris buildup occurs,in; 2
strainer's surface area actually grows, thereby not remaining at 56 ft .
- 2. A given strainer's behavior can be accurately predicted by treating the strainer as a large cylinder that has similar behavior, based on its cylindrical surface area, as the ,
\'
tested prototype. As shown in the EMI tests, the NUKON fibrous debris will collect all over the strainer and fill all the voids, gaps, etc. This should be valid when the strainer has disks that are 1.85" wide and are separated by 2.00" wide gaps, the same as the tested strainrr prototype.
- 3. The tested corrosion product particulate used in the EPRI Head Loss tests accurately simulates all the specified particulate in a particular plant's specification.
- 4. Calculated Head Losses at 60*F can be recalculated for other, higher temperatures by simply multiplying by the ratio of the kinematic viscosities at each of the two temperatures. This is based on the derivation of Equ. B 32a in NUREG/CR 6224 (Ref.1). This also assumes that the Head Loss across a fibrous debris bed is dominated by the laminar, viscosity dependent portion of that equation; this assumption can be validated by calculations which show that the turbulc.it, non-viscosity portion of the equation contributes little to the calculated Head Loss at some other water flow rate and some other quantity of fibrous debris.
- 5. The debris build up on the new strainers is uniform over its length and the Head Loss is uniform across any part of the strainer.
- 6. To calculate Head Losses at other values of mass ratio values than those encompassed by the tests, namely 4:1 to 1:2, one can use the portion of Equ. B-32a (Ref.1) that includes the mass ratio:
Mass Ratio Correction Factor = ((1 + 0.54 (new ratio))/(1 + 0.54 (ref. ratio)))".
15
S:mmary Report on the Performance cf Perf:rmance Contracting, Ice.'s S re-Flow Suetion Strainer wi3 Vcrious Mises cf Sim: lated post-LOCA Debris, 09/19/97 Rev.1 Table 3 (following) shows the set up data for a regression analysis of the Head Loss tc ting data as a function of Flow Rate and Mass of NUKON Fibrous Debris. As explained in the Assumption 1 above, each of these two independent variables is first divided by the test strainer's surface area, then a new independent variable that is the product cf th.,se first two, and then use the Excel regression program to generate coefficier.ts for the following equation:
Equ.2 HL = A + B * (Q/A,) + C * (M/A. , + D * (Q/A,) (M/A,)
where Q = strainer flow rate, gptr.
A, = strainer's cylindrical surface area, sq. ft.
M, = mass of NUKON fibers,Ibs.
HL = strainer head loss, feet of water Table 4 shows the results of the regression analysis using the test data:
A = 0.7696 B = -0.02292 C = -0.5406 D = 0.08916 where value of R' = 0.9828 for this analysis, indicating an excellent correlation fit over the range of these tests.
l l
l L
S:mmtry Report c3 Ce Performance cf Performance Contracting, I:c.'s S:re-Flow" S:ction Strainer with Vcrious Mixes cf Sim:nated post-LOCA Debris, Rev.1 09/19/97 Table 3: EPRI Head Loss Test Data Used for a Regression Analysis EPR; HEAD LOSS DATA WITH 100 LBS. CP AND 25,50,100, '
AND 200 LBS. NUKON, WATER FLOW RATES 0 TO 10,000 GPM
, HEAD LOSS VALUES CORRECTED FOR 60 DEGREES F WATER FLOW MASS MASS MASS HEAD FLOW MASS CP NUKON CP/ MASSX LOSS RATE NUKON MASS / AREA X
, RATE MASS Fl.OW PER PER FLOW NUK RATE AREA AREA, RATE / AREA GPM LBS LBS LBS/LBS GPM LBS FEET GPM/SQ LBS/SQ GPM-LBS/FT(4)
WATER FT FT TEST NO.
95 3 0 100 25 4 0 0.00 0 0.446 0 2500 100 25 4 62500 0.83 44.6 0.446 19.93 3750 100 25 4 93750 1.46 67.0 0.446 29.89 5000 100 25 4 125000 2.13 89.3 0.446 39.86 7500 100 25 4 187500 2.42 133.9 0.446 59.79 10000 100 25 4 250000 2.58 178.6 0.446 79.72 95 5 0 100 50 2 0 0.00 0.0 0.893 0.00 2500 100 50 2 125000 2.29 44.6 0.893 39.86 3750 100 50 2 187500 3.81 67.0 0.893 59.79 5000 100 50 2 250000 5.42 89.3 0.893 79.72 7500 100 50 2 375000 8.08 133.9 0.893 119.58 10000 100 50 2 500000 10.17 178.6 0.893 159.44 96-4 0 100 100 1 0 0.00 0.0 1.786 0.00 2500 100 100 1 250000 6.20 44.6 1.786 79.72_
3750 100 100 1 375000 9.95 67.0 1.786 119.58 i5000 100 100 1 500000 14.08 89.3 1.786 159.44 I 6250 100 100 1 625000 16.97 111.6 1.786 199.30 96- 51 0 100 200 0.5 0 0.00 0.0 3.774 0.00_
2500 100 200 0.5 500000 12.33 47.2 3.774 178.00 3000 100 200 0.5 600000 14.94 56.6 3.774 213.60 3500 100 200 0.5 700000 19.15 66.0 3.774 249.20 4000 100 200 0.5 800000 22.03 75.5 3.774 284.80 17 l
l S:mmary Report ca the Perf;rmrce cf Performrce Centracti:g,1:c.'s Stre-Flow S:cti:3 S' rainer with Varl=s Mixes cf Simil:ted post-LOCA Debris, Rev.1 09/19/97 Table 4: Results of Regression Analysis of Data in Table 3 and Using Equation 2 for a Regression Fit
SUMMARY
OUTPUT Regression Statistics Multiple R 0.991381 R Square 0.982837 Adjusted R 0.979976 Square Standard 0.975788 Error Ob 3rvations 22 ANOVA di SS MS F Signibcance F Regression 3 981.4552 327.1517 343.5886 4.5185E-16 Residual 18 17.1389 0.952161 Total 21 938.5941 Co. Standard tStat P-value Lower 95% Upper Lower Upper
. etlicienis Error 95% 95.0% 95.0%
\ intercept 0.769632 0.005264 1.271563 0.219713 -0.5019821 2.041246 -0.50198 2.041246 X Variable 1 -0.02292 0.006621 -3.46179 0.002783 -0.0368284 -0.00901 -0.0;'S83 -0.00901 X Variable 2 -0.54061 0.317771 1.70126 0.106104 -1.208226 0.127001 -1.20823 0.127001 X Variable 3 0.089155 0.00491318.14504 5.13E-13 0.07883243 0.099478 0.078832 0.099478 Using this Equation 2. Head Loss values, for a particular plant's strainers, can be generated for a case where the suppression pool has water temperature T, the strainers a constant Water Flow Ra e Q, a total of M, pounds of shredded NUKON Fibers and Mci, pounds of CP Particalate reach the strainers, the strainers have diameters D and lengths L, and an infinite period of ECCS pump operation. The following example problem shows hcw this can be accomplished:
Example Head Loss Problem Using the Regression Equation 2:
Suppose a plant has four equal sized strainers on a cormnon ring header which draws a total of 20,000 gpm. Each strainer is a Sure-Flow stacked disk strainer, with all disks of the same size, each strainer is 4 feet long and 45 inches in diameter, and each is mounted on a 20 inch NPS pipe. Assume, for the purposes of design conditions, that the water temperature is 180" F, that 800 lbs. of CP particulate and other particulate are in the suppression pool, and that 233 cubic feet of shredded NUKON insulation are transported to the pool. Ignore any particulate or fiber sedimentation and assume 100% filtration 18
Summ:ry Report ca the Perf:rma ce cf Perf:rmance Cc:tracting,Inc.'s S:re-Flm* S:cti:2 Stral:er with Vcri::s Mixes cf Sim:1:ted post-LOCA Debris, Rev.1 09/19/97 efficiency of the paniculate by the fibers on the strainers. Also, assume that NUKON insulant has an as fabricated density of 2.4 lbs./ cubic foot.
Solution:
First, calculate the cylindrical surface area of each strainer, including the end disks, this works out to be about 70 ft2 . Therefore, Q/A = (20,000 gpm total / 4 strainers) / 70 ft' cyl area = 71.4 gpm/ft' and M/A = ((333 ft' fibers / 4 strainers ) * (2.4 lbs./ft') / 70 ft' = 2.85 lbs/ ft' Putting these into Equation 2, the Head Loss can be calculated for ambient water:
HL. = 15.7 feet of wmer at 60* F water, which has a kinematic viscosity of 1.217 x 10 ft'/sec.
This Head Loss can now be corrected for 180 F water, which has a kinematic viscosity of 3.85 x 10" ft 2/sec:
4 HL oi = 15.7
- 3.85 x 10 /1.217 x 10 = 5.0 feet of water.
This can now be corrected for the Mass Ratio of CP to fibers:
Mass Ratio CP : Fibers = 800 lbs. / (333 ft'
- 2.4 lbs/ft') = 1.00 This compares to the range of the tested values. For EPRI Test No. 96-4, the Mass Ratio was also 1.00. Therefore, no correction will be made for ma' ratio. Hence, the predicted Head Loss across each strainer, which is 45 inches in diameter and 4 feet long, is 5.0 feet of water at 180* F. This problem is solved.
1 19
Summary Report on the Performance of Performance Contracting,Inc.'s Sure-Flow Suction Strainer with Various Mixes of Simulated post-LOCA Debris, l Rev.1 09/19/97
. . 'e .
p- .
.~
. -~ 4
,; .;;::- ~
1 . .- -
s,
^([
. - - I ., j- ; , , ( ', l H' y .
i~ .
,s
?- as .
4 .
, \
- . o Figurc 8 - Photograph of the tested strainer following Test No. 95-3 (with 10.3 ft' of shredded fibrous debris and 100 lbs. of CP particulate)
^
N h*?
L
.s -
x .. n 'g
- - . 2.
%~, M _- ,3 # c~','- ,
- S y
,- - . w*
s. ' 9, f
,;)
^,.v. '
' ' . .a .
{ _' :,]f:. ,
- q. .
- , / . .
_ k.
r n .
. / .
s , . . # -
8 Figure 9 - Photograph of the tested strainer following Test No. 95-5 (with 20.7 ft' of shredded fibrous debris and 100 lbs. of CP particulate) 20
Summar.s Report on the Performance of Performance Contracting,Inc,*s Sure-l Flow" Suction Strainer with Various hilxes of Simulated post-LOCA Debris, l 09/19/97 i Rev,1 l
i
. -.t n. m 5
_' !g:if
~
h : %Q,h.5k ,
g.
- j. ,; ; !
l
'N ,-l
.V l
/. - -
.. ; ./ ,
..,9 ss,
,/ ;
- 3. -
i
.l _ - , _
V- f f[ ~ .:f[![?! <$} .).$xf-;.s . .;
'i
? -
^
~
.,y 6, _
.X .. . _ . . ' . . . - ., .-
.. " . a. .
. . . ...:~
Figure 10 - Photograph of the tested strainer following Test No. 96-5 (with 83.3 ft' of shredded fibrous debris and 100 lbs. of CP particulate) 1 1
V , CONCLL'SIONS i i
l A stacked disk strainer prototype was tested to determine its Head Loss performance over a wide range of Water Flow Rates and Niass of Fibrous Debris and for one quantity of Corrosion Product particulate. NIass of Fibrous Debris was varied from zero to 300 lbs.,
h! ass of Corrosion Product Particulate was varied between 0 and 100 lbs., and Water l Flow Rate was varied between zero and 10,000 gpm for this single strainer prototype.
From these tests, the following conclusions cou!d be reached about the behavior of the Sure-Flow Strainer prototype tested:
1
. the bare strainer (i.e., with no debris) showed a very low Head Loss behavior and that j Head Loss is dependent on the entrance (i.e., at the strainer's nozzle) water velocity;
- its Head Loss behavior is essentially linearly dependent on both hiass of Fibrous Debns and Water Flow Rate; l l
l l
i 21 l
. = _-- - .-- _ _.
S:mmary Repot t ca the Perf rmrce cf Perf:rm=ce Cc tracti::g,1:c.'s Sire-Flow S:ction Stral:er with Vcrious Mixes cf Sim:1:ted post-LOCA Debris, Rev.1 09/19/97 e the addition of 100 lbs of CP particulate increases Head Loss across the strainer significantly, e the Head Loss behavior can be accurately modeled with regression equations, developed from the test data, and applied over the tested range of those variables, namely Mass of Fibrous Debris and Water Flow Rate;-
e addition of stainless steel foil fragments, which simulate LOCA generated Reflective Metallic Insulation debris, increases the Head Loss across the Sure-Flow Strainer by about 0.9 feet ofwater at 5000 gpm; e thick fibrous debris beds exhibited an effective filtration efficiency that exceeded 90%;
e on this strainer prototype, the fibrous debris beds exhibited an apparent bed compaction of appmximately 24% (using the as-fabricated insulation density as a reference);
e the Sure-Flow Strainer, mounted in a horizontal position, did not exhibit development of vortices, even when the tank was drained so that the strainer was about half exposed above the water level.
22
S:mm:ry Report ca the P;rf;rmrce cf Perf:rmrce Ccatracting,I c.'s S2re-Flow S:etion Stral:er with Vcrl:cs Mixes cf Sim: lated post-LOCA Debris, Rev.1 09/19/97
References:
- 1. Utility Resolution Guidance for ECCS Suction Strainer Blockane, General Electric Nuclear Energy Co., Report No. NEDO-32686, Rev. O, Class 1, November,1996.
- 2. Parametric Study of the Potential for BWR ECCS Strainer Blockane Due to LOCA Generated Debris, prepared by G. Zigler, J. Brideau, D.V. Rao, C. Shaffer, F. Souto, W. 'Ihomas of Science and Engineering Associates, Inc., prepared for U.S. Nuclear Regulatory Commission, Report Number NUREG/CR-6224, September,1995.
- 3. Results of ECCS Sure Flow Strainer Testine for Performance Contractine. Inc.. Rev.
O, December 1996, Prepared by Electric Power Research Institute, by Kaufman, Andrew E., Dienl, Robert W., Louderback, Richard G,
- 4. The Develonment and Testine ofPerformance Contractine. Inc.'s Sure Flow Stacked Disk Suction Strainer for BWR ECCS Lines, by Gordon H. Hart, Febmary 1, 1996.
- 5. Performance Contracting, Inc. Memofor Record, "QA Dedication of Strainer testing at EPRI", by Gordon H. Hart, November 11,1996.
- 6. Merritt, Frederick F., Editor, Standard Handbook for Civil Encineers, Third Edition, McGraw Hill Book Company,1983, p. 21-27.
- 7. Tests Conducted at the Fairbanks Morse Pumn Comnany To Determine the Head Loss Performance of PCI's Bare Sure-Flow Strainer. Report Number PCI-NPD-CE03, Rev. O, by Gordon H. Hart, June 5,1997.
l Prepared by: Prepared by:
- w. W Gordon H. Hart, P.E.
(LiA k t%L l Richard A. Biasca Date: N, If 97 Date: bdL / 9 [N 7 v ( <
Reviewed and Approved by:
l
&f Carl E. Nuzryin, P /
Hydraulic Consu arft Date: ./f y /Wh .,
l
.