ML17213A554
| ML17213A554 | |
| Person / Time | |
|---|---|
| Site: | Saint Lucie  |
| Issue date: | 09/30/1982 |
| From: | Durgin W ALDEN RESEARCH LABORATORY |
| To: | |
| Shared Package | |
| ML17213A553 | List: |
| References | |
| 106-82-M7KF, NUDOCS 8210130287 | |
| Download: ML17213A554 (41) | |
Text
PERFORMANCE EVALUATION OF A CONTAINMENT SUMP AT FULL SCALE ST. LUCIE NUCLEAR POWER STATION NO. 2 by William W. Durgin Prepared for Ebasco Services, Inc.
George E. Hecker, Director ALDEN RESEARCH LABORATORY WORCESTER POLYTECHNIC INSTITUTE HOLDEN, MASSACHUSETTS September 1982
>>0280? 821006 EPDR ADOCK 05000g89 PDR I
Printed at ARL September 1982 Model tests at full scale were conducted in order to-evaluate the hydraulic performance of the containment sump (Emergency Safety Feature Sump) of the St. Lucie Nuclear Power Station, Unit No. 2. A replica of the sump was con-structed within a test basin. The model sump was operated with various com-binations of water level and two flowrates in each outlet line. Measurements and observations were made to identify surface vortex types, outlet line, swirl angles, and combined screen/pipe inlet loss coefficients.
By testing at fixed outlet flowrates and at two water levels, various com-binations of vertical screen blockage (50%), horizontal screen blockage (50%),
and approach flow distribution were utilized to arrive at two configurations for testing over an appropriate range of surface elevations and flowrates.
The results indicated that the loss coefficients were dependent on submergence and blockage over the range tested but essentially independent of Reynolds num-ber. The average vortex types observed were between 1 and'2 for all tests, and, therefore, practically insignificant inasmuch as these types do not entrain air or debris. ,Outlet line swirl angles were found to be between 0 and 5.74'nd depended on blockage but showed no systematic dependence on Reynolds number.
The observed and measured loss coefficients, vortex types, and swirl angles were compared to similar findings in the literature and found to be consis-tent.
3.3.
TABLE OF CONTENTS P~aa No.
ABSTRACT TABLE OF CONTENTS INTRODUCTION PROTOTYPE DESCRXPTXON SIMXLITUDE MODEL DESCRIPTION INSTRUMENTATION AND OBSERVATION TECHNIQUES 15 TEST PROCEDURE 20 RESULTS AND DISCUSSION 24 SUl1l1ARY AND CONCLUSIONS 35 REFERENCES
INTRODUCTION The containment building of the St. Lucie Nuclear Power Station, Unit No. 2, is equipped with an Emergency Safety Feature (ESF) Sump which must function satisfactorily under certain conditions.
Following an accident that would release a significant amount of energy in-side containment (eg, Loss of Coolant Accident or Main Steam Line Break) the subsequent rise in pressure and temperature would trigger the operation of certain safety systems. At a predetermined level of containment pressure, operation of the Containment Spray (CS), High Pressure Safety Injection (HPSI),
and Low Pressure Safety Injection (LPSI) Pumps would be initiated. During the initial phase of the accident, these six pumps (two of each) draw water from the Refueling Water Tank and deliver it into containment. This estab-lishes a minimum water level inside containment at EL 21.0 ft.
Low tank level would initiate a Recirculation Actuation Signal (RAS) which would deactivate both LPSI Pumps and transfer the suctions of the CS and HPSI Pumps to the ESF Sump. During this mode of operation, total pump flow in each sump outlet pipe will fall within the range 300 gpm to 7300 gpm.
The ESF Sump was designed to provide pump suction in the recirculation mode for an indefinite period.
The ESF Sump is a collection reservoir located in the annulus between the secondary shield wall and containment, and functions to provide an adequate supply of water to the Containment Spray and High Pressure Safety Injection Pumps during the recirculation mode. Two redundant suction lines, each han-dling one CS and HPSI Pump, are located at opposite ends of the sump. In this location, the sump is protected from the direct effects of high energy line break, such as jet impingement and pipe whip.
Following an accident, the entire cross-section of the containment would be filled with water to an elevation within the range EL 21.0 ft and EL 26.0 ft.
Water drawn from the sump would be returned to the containment via the con-tainment spray headers located high in the containment and possibly through a break in the Reactor Coolant System. As such, the majority of the water would be returned inside the secondary shield wall and would reach the sump via shield wall drain openings. These drains are large rectangular openings located at various points along the perimeter of the wall. Each opening is equipped with a grating to prevent large debris from reaching the sump.
The ESF Sump is contiguous with the pipe trench around the perimeter of the secondary shield. Some of the filter screening is, in fact, located within the trench. Pipes of various sizes penetrate the screens and a drain collec-tion tank is located within the sump.
A full scale replica of the FSF Sump and nearby features was constructed in a test facility at the Alden Research Laboratory (ARL) of Worcester Polytechnic Institute (WPI). A test program was devised in order to investigate free sur-face vortex formation, swirl in the inlet piping, inlet losses, or any other flow conditions that could adversely affect the performance of the decay heat removal pumps and the reactor building spray pumps in the recirculation mode.
Operating conditions involving a wide range of possible approach flow dis-tributions, flowrates, water levels, screen blockages, and combinations there-of were tested in the model.
It is of primary importance that Containment Spray Pumps and High Pressure Safety Ingestion pumps function properly after a Loss of Coolant Accident or Main Steam Line Break when pump suction is switched to the Emergency Safety Feature Sump. In particular, it is necessary to evaluate the flow through the sump in terms of head loss, air-entrainment, and outlet line swirl. Be-cause the approaches to the sump may be partially blocked by debris, the re-sulting flow patterns can affect these parameters. The head loss through the (partially blocked) sump and outlets can be a principal determinant of avail-able NPSH. The tolerable levels of air content and swirl are dependent on pump
design so that air is not entrained and to minimize swirl so that it has negligible contribution to pump inlet swirl given the remainder of the
,suction piping.'his report presents the findings of the study including a description of the prototype and the model, and summarizes conditions investigated, simil-itude considerations, test procedures, instrumentation, and interpretation 3
of results.
PROTOTYPE DESCRXPTION The ESF Sump is located between the shield w'all and containment contiguous with the pipe trench, Figure 1. The pipe trench is nominally 5 ft wide with the bottom at EL 12.00 ft. The sump recess is nominally 9 ft wide with bot-tom at EL 7'.58 ft centered about 180 , azimuth. A floor at EL 23.00 ft oc-cupies the space between the pipe trench/sump recess and containment. There are three smaller pipe trenches in the floor in the vicinity of the sump.
The shield wall is provided with periodic drain openings nominally 4 ft wide.
Three of these are located in the vicinity of the sump. Two 24 inch outlet lines, Figure 2, lead from the lower corners of the sump through containment at EL 9.00 ft. These lines project into the sump and are provided with sleeves forming re-entrant inlets with exterior steps.
All shield wall drain openings are provided with heavy bar racks to prevent ingestion of large debris. The sump is completely enclosed by a fine mesh filter screen. This screen is made up of .047 inch OD stainless steel wire spaced to provide an open area of approximately 90 mils square. Vertical screen sections are arranged in a sawtooth pattern, forming 60 degree angles, to increase the available surface area. Horizontal sections of the sump screens are attached to floor gratings at EL 23.00 ft. These gratings are made of 3/16 inch wide, l-l/4 inch deep rectangular bars spaced 1-3/16 inches apart with cross members every 4 inches and cover all pipe trenches. Screen,
.however, is only provided on those portions inside the vertical screen panels.
A small horizontal grating with screen is provided at EL 11.00 ft between the outlet pipelines. A flat vertical panel of screen, reinforced with bars, di-vides 'the sump into two, nominally at 180', azimuth, with a small sawtooth panel protruding into the portal at 180', azimuth.
The reactor drain tank, 12 ft long by 5 ft diameter, is located in the sump, inside the screens, and between the outlets at EL 12.00 ft. Numerous pipes of various sizes run in the pipe trenches, through the sump, and penetrate the screens and,gratings. Associated with these pipes are pipe supports and seismic restraints.
CD I
ca CV CD I
I 0Y I z I X Z I ca I I
UJ UJ O I-UJ a=
I z I I
CD I
o ca 0
LL: K~'0~
0 0 O
Y O IJ:
I o <<,I z0 0U-U I Z0 CD 0
OU LU J J 0I- LU I
I v 0 LU I .Ct:
0 f I C4: LU o Lll 'O O I U-
.O ~OI I I ca I z .'P I I- I-x0 U
<<~ZI 0 Ja. 0O I CD CD I
<<L: <<L'.
z K I CV C74 H
UJ I z I H I I-LU I
0 CL-0M 0 ~
0 CD /)
I~
P4 I- Ol I I 8 a. LLJ I I
I I
I I I I
/ ~ I Y7
\
4 0
z UJ J LLI UJ v) 9 0 ~
I CV U LIJ
Q REACTOR 90 MIL SCREEN EL 22'-2-1/4" PLATE EL 23'-0" LADDER BOTTOM OF TRENCH EL 12'-0" Q REACTOR DR TANK =
EL 12'-0" PLATFORM CHKD Q EL 11'-0" BOTTOM OF TRENCH EL 12'4"
~/g / DIVIDER SCREEN 90 MIL BOTTOM OF SUMP EL 7'-7" /
TRAIN B TRAIN A FIGURE 2 ESF SUMP SECTION
The sump water temperature can vary between 55 and 225oF, with a surface elevation between 2l and 26 ft. The outlet line flow can vary between 300 and 7300 gpm per line and the containment pressure varies between 0 and 44 psig.
The geometry, features, and piping in the sump area were described by Ebasco Drawings (l). In addition, a site visit was made and measurements as well as photographs made of sump details.
SIMILITUDE The study of dynamically similar fluid motions forms the basis for the design of models and the interpretation of experimental data. The basic concept of dynamic similarity may be stated as the requirement that two systems with geometrically similar boundaries have geometrically similar flow patterns at corresponding instants of time (2). Thus, all individual forces acting on corresponding fluid elements of mass must have the same ratios in the two systems.
For a situation in which a free surface is present, the Froude number u
F =
MgL where u and L are a characteristic velocity and length, respectively, should be the same in the model as in the prototype. In the present cases, the model was constructed the same size as the prototype which gives
'm = u
,u p
Thus, velocities, flowrates, and time scales for the model will be the same as for the prototype.
Scaling considerations then, depend only on secondary non-dimensional groups.
In particular, a loss coefficient or Euler number
~b 2
pu where hp is pressure drop and P the fluid density, will depend only weakly on the Reynolds and Weber numbers R =uL V
W = u
/o/pz,
where 9 is the kinematic viscosity, and 0 the surface tension. Thus, E = E(R,W) with the only variation due to temperature dependence of R and W. Further-more, since only results of very weak vortex activity, were acceptable, the Weber number effect was insignificant. Then, E = E(R) and the variation of R over the temperature range should be evaluated. For model operation at 68'F (typical) at the lowest flow tested, 4285 gpm 1 line, the pipe Reynolds number becomes R = 5.7 x 10 5 , at 68'F while at 225'F 6
R = 2.0 x 10 at 225 F Since it is known that parameter variation becomes asymptotic at such large Reynolds numbers, the, effect of temperature will be insignificant.
The model, being constructed at full scale, should thus exhibit similar flow patterns, loss coefficients, vortex formation, and swirl as the prototype operating at the same flowrate and water level. In addition, screening of identical size and construction was used so that the head losses would also be identical under similar operating conditions.
For the prototype operating at maximum flow, 7300 gpm, at maximum temperature, 225'F, the Reynolds number for an outlet line would be R = 3.4 x 10 6
10 Model tests were conducted at R values of 5.6 x 10 5 and 1.5 x 10 6 . Loss coefficients are usually constant in and above this range and thus should be the same at the highest prototype Reynolds number. The effect of vary-ing containment pressure with temperature and time does not affect the loss coefficients or inlet pressure loss, but only the NPSH at the pumps.
MODEL DESCRIPTION The model was constructed at a 1:1 geometric scale within the Experimental Facility for,Containment Sump Reliability Studies, Durgin, et al (3), at the Alden Research Laboratory. This fixed facility includes a large tank (L 70 ft, W 35 ft, D 12 ft) with a flow distribution system capable of sup-plying 20,000 gpm along three sides. A depressed sump (L 20 ft, W 10 ft, D 10 ft), located in the center, had provision for outlet lines at 15 loca-tions up to 2 ft in diameter. Installed equipment provided for water fil-tration, level control, and flowrate control. Both inlet and outlet line flow meters were provided as were pressure measurement systems. All data acquisition and reduction were under control of a mini-computer at the site.
The model was installed in the test basin as shown in Figure 3, primarily using plywood panels to generate the fixed boundaries. Three of the shield wall portals were included through which flow could enter the sump area.
In addition, flow could enter over the floor on the east and west ends and through the pipe trench on the west end. The reactor drain tank was instal-led within the sump as shown in Figure 4 in which the sump divider screen and folded sere'ens can be seen. This screen was identical in size and spac-ing as that of the prototype.
All pipelines of finished diameter greater than 3 inches were installed in the model as were other objects which might, affect the flow patterns.
h Figure 5 shows a view looking towards the sump from the west while Figure 6 shows a view through the east drain opening. The floor grating and at-tached screen at EL 23.00 can be seen in Figure 7 where a portion of grat-ing has been turned to reveal the sump interior.
180 z
EAST WEST CI C) C) 0)
LU FACILITY BOUNDARY O CONTAINMENT I
I, SI I li C)
I - I I
I EL 11.00 I I I II il Ol VERTICAL (I li SCREENING i I oo EL 7.58 >I
)
I r i.oo I
<C Ii 7~ 00 Il
)I EL 1800 I
I
)I (TOP EL 22.00)
EL 15.5 I
EL 14.00 (TOP EL 23.001 (TOP EL 22.00)
I 35'0'LOW DISTRIBUTOR SHIELD WALL FIGURE 3 MODEL SUMP INSTALLATION
FIGURE 4 SUMP INTERIOR
,/
Hi
()j)
)g 0
l~~
I
'-h FIGURE 5 WEST PIPE TRENCH, VIEW FROIVI WEST
FIGURE 6 EAST PIPE TRENCH THROUGH SHIELD WALL OPENING
/r p
/
FIGURE 7 HORIZONTAL BAR RACK AND SCREEN, EL 23.00
INSTRUMENTATION AND OBSERVATION TECHNIQUES Each 24 inch outlet line was equipped with a vortimeter, Figure 8, piezo-meter taps for gradeline measurements, a flow meter, and a regulating valve.
The 12 gradeline pressure taps were connected to a scanning valve which sequentially connected each to a pressure cell. The computer system con-trolled the scanning valve and measured the voltage output from the pres-sure cell. The differential pressure output from the flow meters was simi-larly measured. Figure 9 shows the outlet piping giving pressure tap loca-tions The various devices used were
~
flow meters: 24 inch annubars (2) pressure cell: Sensotec +1 psiD (2) pressure cell: Sensotec +7.5 psiD (1) vortimeter: 24 inch ARL Each test was 30 minutes in duration. The observed vortex type, according to the ARL scale, Figure 10, was entered to the computer system every 30 seconds on a hand held computer terminal. The pressure gradelines were fully sampled every 60 seconds as were the accumulated vortimeter revolutions.
16 FIGURE 8 VORTIMETER
VORTIMETER PRESSURE TAPS GRADELINE PRESSURE TAPS TRAIN A VORTIMETER FLOWMETER CONTROL VALVE TRAIN B 12'6't 22YP 8'IGURE-9 OUTLET PIPELINES
18 VORTEX TYPE INCOHERENT SURFACE SWIRL SURFACE DIMPLE; COHERENT. SWIRL AT SURFACE DYE CORE TO INTAKE; COHERENT SWIRL THROUGHOUT WATER COLUMN VORTEX PULLING FLOATING TRASH, BUT NOT AIR TnaSH
~l VORTEX PULLING AIR BUBBLES TO INTAKE D aiA euseaES FULL AIR CORE TO INTAKE FIGURE 10 NUMERICAL SCALE FOR VORTEX TYPE CLASSIFICATION
19 The pressure gradeline was extrapolated to the entrance by a linear least squares curve fit of the pressure measurements. The area average velocity was used to calculate the pipe velocity head, which was added to the extrapolated pressure gradeline. The total head at the pipe inlet was subtracted from the sump water level outside the screens to determine the inlet loss. An entrance loss coefficient was calculated by:
K =
2 i
2g where K = loss coefficient i
6H. = inlet head loss, ft Average swirl in the suction pipes was measured by cross-vane swirl meter.
Lee and Durgin (4) have shown that a swivel meter with vane diameter about 75% that of the pipe diameter best approximates the solid body rotation of the flow. The rate of rotation of the vortimeter was determined by count-ing the number of blades passing a fixed point in two minutes.
An average swirl angle was defined as the arctangent of the maximum tan-gential velocity divided by the axial velocity. 'The maximum tangential velocity is the rotational speed times the circumference of the pipe, Tf d N, and the average swirl angle is defined- by:
6 = arctan (
TdN ) (l4)
U where N = revolutions per second d = pipe diameter, ft U = mean axial velocity, ft/sec
20 TEST PROCEDURE Tests were conducted at the available water supply temperature. The model was filled to the proper level, and all piezometer and manometer lines were purged of air and the differential pressure cells zeroed. The required flow-rates were then set and allowed to stabilize for 15 minutes. The water level was checked and re-adjusted, if necessary. The data acquisition system pro-ram was then started so that all measurements were automatic.
gram The system signaled the model operator every 30 seconds to request observed vortex type data. Subsequent to test completion, all data were transferred to'hard disk files. An analysis program was then run to determine average vortex type and standard deviation, swirl angles, and loss coefficients.
The test plan involved systematically varying the water level, flowrates, flow distribution, blockage of vertical screens, and blockage of the horizontal screens. To this end, symbolic representation of various parameters was adopt-ed as given in Table 1.
Three water elevations were chosen: 21.00, 23.00, and 25.00 ft.
Flowrates of 0, 4285, and 8785 gpm were specified by Ebasco Services as repre-sentative. These could occur in various combinations in the two outlet lines.
Flow distributions would occur primarily because of blockage of the drain openings through the shield wall. These were systematically blocked, Figure ll, entirely except for the west end which was fed by numerous portals. A maximum blockage of 1/2 was adopted for this entrance.
Vertically; screen blockage was always 50% of the available screen area but arranged in various patterns. Longitudinally, the screens were divided into segments, Figure 12, so that any two segments could be blocked. Vertically, the screens were divided in half so that only the top or bottom was blocked.
21 TABLE, 1 TEST PLAN DEFINITIONS Water Surface Level (ft)
Sl 21. 0 S2 23.0 S3 25. 0 Flowrates (gpm)
Ql 4285 4285 Q2 4285 0 Q3 0 4285 Q4 4285 8785 Q5 8785 4285 Q6 8785 8785 Flow Distribution (see Figure 10)
DO none Dl A + B D2 A+B+C D3 D+1/2E D4 C+D+1/2E Screen Blockage =(vertical; see Figure 11)
VO none Vl A + D top to bottom V2 B + C top, to bottom V3 B + D top to bottom V4 A + C top to bottom V5 top 1/2 all U6 bottom 1/2 all Screen Blockage (horizontal)
HO none Hl outer 1/2 H2 inner 1/2 H3 all
22
-+- O.
C FIGURE 11 FLOW DISTRIBUTION DEFINITIONS B /
Qo iQo Qo FIGURE 12 SCREEN BLOCKAGE DEFINITIONS
23 Horizontal screen blockage consisted of the outer 1/2, inner 1/2, or all being blocked, and was only used for S3 (EL 25.00) tests.
The initial 23 tests concentrated on determining the flow distributions and screen blockage most conducive to adverse performance. To this end, the flo'ws were fixed at 8785 gpm in each line and only the highest and lowest water levels used. All vertical screen blockages, *all horizontal screen blockages, and flow distributions were tested independently.
Four combinations of blockage, Cl through C4, were then selected for further evaluation. Abridged testing as above, enabled two worst tests, designated Wl and W2, to be selected.
Complete variation of water level and outlet flowrates was then conducted, using Wl and W2, to evaluate the sump performance.
24 RESULTS AND DISCUSSION The results for all tests are shown in Table 2. It can be seen that the worst vortex type observed was type 2; which was predominant.
For the initial 23 survey tests, the loss coefficients varied from 0.71 to 1.09 for Train A and from 0.75 to 1.11 for Train B. The swirl angle ex-tremes were 2.48'o 5.15'or Train A and -1.76 to 0.38'or Train B.
Combination blockage Cl, Figure 13a, was selected by observing that test ll had both high swirl and high loss for Train A and that test 17 had the highest combined swirl for both pipes which was due to horizontal blockage.
Thus, it might be expected that vertical blockage V6 combined with horizon-tal blockage Hl, would produce even higher swirl. This was the case because the Train A swirl increased to 5.74', test 24, but the observed vortex type decreased and the losses for each line changed only slightly. The swirl in Train B also remained about the same. Since horizontal blockage could only be tested at, elevation S3, a test with H2 'blockage was added but did not produce interesting results.
Combination blockages C2 and C3, Figures 13b and c, were selected by observ-ing that the D4 flow distribution, tests 16 and 23, respectively, produce high swirl and above average loss coefficients. This was combined with ver-tical blockage V3 and the antisymmetrical blockage V4 with the thought that one of them ought to combine with D4 to produce even higher swirl and loss coefficients. This proved to be the. case in test 26, C2, where the Train A swirl reached 5.28'. The observed surface vortex was type 1 only. A signi-ficantly high loss coefficient was observed for C3 blockage in Train B, test 29
'ombination blockage C4, Figure 13d, was selected by observing that the high-est loss coefficient with significant swirl occurred in Train B, test 21, dis-tribution D2.. This was combined with the highest swirl in Train B, test 3, screen blockage V2. The loss coefficients for both Trains A and B reached their highest values in these tests with significantly high swirl angles in Train A.
TABLE 2'est Results Train Train Vertical Vortex Loss Swirl Loss Swirl Flow Screen Horizontal Surface Test ~Ty e Coeff. e Coef f. e Blockage e Distribution Blockage Blockacle Piowrate Elevation 1 2.0 0.83 0.90 0.0 Screen DO Vl HO Q6 S3
. 2 2.0 1.09 1.11 0.38 Screen DO Vl Q6 Sl 3 2.0 0.71 4.68 0.86 -1.76 Screen DO V2 HO QG S3 4 2.0 0.86 2.48 0.92 -1.31 Screen DO V2 Q6 Sl
-5 1.8 0.86 4.89 0.82 0.0 Screen DO V3 HO Q6 S3 6 2.0 0.99 3.11 0.86 0.0 Screen DO V3 Q6 Sj 7 2.0 0.81 4.41 0.82 -l. 54 Screen DO V4 HO Q6 S3 8 2.0. 0.94 3.16 0.95 -1.08 Screen DO V4 .Q6 Sl 2.0 0.84 4.60 0.81 -0. 94 Screen DO V5 HO Q6 S3 10 2.0 0.94 3.61 0.91 ,0.89 Screen DO V5 Q6 Sl 11 2.0 0.81 4.94 0.90 -0. 93 Screen DO V6 HO Q6 S3 12 2.0 0.93 4.67 0.99 -0.56 Screen DO V6 Q6 Sl 13 2.0 0.82 4.57 0. 87 -1.54 Flow distribution Dl VO HO Q6 S3 14 2.0 0. 82 4.85 0. 86 -1.54 Flow distribution D2 VO HO Q6 S3 15 1.0 0.80 4.46 0.83 -1.54 Flow distribution D3 VO HO Q6 S3 16 1.0 0.79 5.10 0.86 -1.41 Flow distribution D4 VO HO Q6 S3 17 1.0 0.79 5.15 0.75 -1.33 Horizontal DO VO Hl Q6 S3 18 1.1 0.81 4.54 0.76 -1.23 Horizontal DO VO H2 Q6 S3 19 1.9 0.83 4.51 0.85 -1.15 Horizontal DO VO H3 Q6 S3 20 2.0 0.94 3.95 1.11 -0.75 Flow distribution Dl VO Q6 Sl 21 2.0 1.06 4.42 1.09 -1.15 Flow distribution D2 VO o6 Sl 22 2.0 0.90 3.16 0.91 -0.94 Flow distribution D3 VO Q6 Sl 23 2.0 0.91 4.80 0.90 -1.14 Flow distribution D4 VO QG Sl 24 1.0 0.88 5.74 0.79 -1 ~ 13 C-1 DO V6 Hl Q6 S3 25 1.0 0.87 5.27 0.79 1 ~ 23 C-1 DO V6 H2 Q6 S3 26 1.0 0.81 5.28 0. 76 -0 96
~ C-2 D4 V3 HO Q6 S3 27 2.0 0.97 4.24 ,0.90 -0.39 C-2 D4 V3 HO Q6 Sl 28 1.0 0.79 3.93 0.86 -1.35 C-3 D4 V4 HO Q6 S3 29 2.0 0.96 2.46 1.13 -0.50 C-3 D4 V4. HO Q6 Sl 30 .2 ~ 0 0.92 5.32 0.97 -0,. 97 C-4 D2 . V2 HO Q6 S3 31 2.0 1.16 2.55 1.29 -0. 39 C-4 D2 V2 HO Q6 Sl
TABLE 2 Test Results (continued)
Train A Train B Vertical Vortex Loss Swirl Loss Swirl Flow Screen Horizontal Surface Test Type Coeff. e Coeff. 6 Blockage Ty e Distribution Blockacle Blockage rlowrate Elevation 32 1.0 0.97 5.18 0.96 -0. 39 W-1 D2 V2 HO Ql S3 33 1.0 0.93 5.10 W-1 D2 V2 HO Q2 S3 34 1.0 0.97 -0. 19 W-1 D2 V2 HO Q3 S3 35 1.0 0.86 5.38 0.84 -0. 57 W-1 D2 V2 HO Q4 S3 36 1.0 0.86 3.95 0.92 -0. 57 W-1 D2 V2 HO Q5 S3 37 1.0 1.11 3.95 1.14 -0. 20 W-1 D2 V2 HO Ql S2 38 1.0 1.01 3.95 W-1 D2 V2 HO . Q2 S2 39 1.0 0.99 -0.39 W-1 D2 V2 HO ~
Q3 S2 40 1.0 l. 39 4.53 1.09 -0.38 W-1 D2 V2 HO Q4 S2 41 1.0 1. 01 3.44 1.47 -0.39 W-1 D2 V2 HO Q5 S2 42 1.9 l. 16 3.35 1.26 -0.39 W-1 D2 V2 HO Ql Sl 43 1.0 0.97 3.14 W-1 D2 V2 HO Q2 Sl 44 1.0 1 ~ 04 0.0 W-1 D2 V2 HO Q3 Sl 45 2.0 1.67 4.38 1.10 -0. 19 W-1 D2 V2 HO Q4 Sl 46 2.0 1.07 2.70 1.70 -0. 58 W-1 D2 V2 HO Q5 Sl 47 1.0 0.81 1.58 0.86 0.0 W-2 D4 V4 HO Ql S3 48 1.0 0 '0 0.99 W-2 D4 V4 HO Q2 S3 49 1.0 0. 84 0.0 W-2 D4 V4 HO Q3 S3 50 2.0 0 81 F 2.10 0.97 -0.38 W-2 D4 V4 HO Q4 S3 51 1.0 0.83 2.30 0.87 0.0 W-2 D4 V4 HO Q5 S3 52 1.0 0.89 3 '4 0.98 0.0 W-2 D4 V4 HO Ql S2 53 1.0 0.88 2.56 W-2 D4 V4 HO Q2 S2 54 1.0 0.95 0.0 W-2 D4 V4 HO Q3 S2 55
'.0 1.06 3.14 0.97 -0. 29 W-2 D4 V4 HO Q4 S2 56 1.0 0.91 2.10'.60 1.15 0.0 W-2 D4 V4 HO Q5 S2 57 1.9 0.96 1.10 0.0 W-2 D4 V4 HO Ql Sl 58 1.0 0.91 0.95 W-2 D4 V4 HO Q2 Sl 59 1.0 0.96 0.0 W-2 D4 V4 HO Q3 Sl 60 2.0 1. 20 0.79 1.02 -0. 38 W-2 D4 V4 HO Q4 Sl 61 2.0 0.96 1.15 1.30 0.0 W-2 D4 V4 HO Q5 Sl
27 A) COMBINATION BLOCKAGE C1 c
I B) COMBINATION BLOCKAGE C2 C) COMBINATION BLOCKAGE C3=W2 D) COMBINATION BLOCKAGE C4=W1 FIGURE 13 COMBINATION BLOCKAGE
28 Since surface vortex activity did not appear significant in any tests through 31, the criterion for selection of combinations Wl and W2 was based on loss co-efficient. Clearly, combination C4 showed the highest values for both lines and was selected as combination Wl. Of the remaining tests, test 29 showed the highest. coefficient so that combination C3 was selected as W2.
In both cases, the flow patterns and screen blockages caused significant amounts of flow .to pass through the relatively narrow spaces between the vertical screens and'the corners at the junctions of the pipe trenches and sump. Typically, surface elevation differences of 2 or 3 inches were ob-served between upstream and downstream flows through these spaces.
For combinations Wl and W2, complete variation of water level and flowrate according to Table I was conducted.
The observed surface vortex type, T, for the Wl and W2 series tests (32-61) are plotted against Froude number in Figure 14. Also shown is an envelope developed during full scale generic studies of containment sumps (5). In those studies, vortex activity was not observed outside the envelope. It can be seen that the observed vortex types for the present study fell well within the envelope and that only types.l and 2 were observed. The consider-able amount of piping within the sump area of this plant apparently served to reduce surface vortex activity to less than might otherwise be expected.
Inlet loss coefficients, K, are plotted against Reynolds number, R, in Figures 15 and 16 for all tests with Wl type blockage (30-46) and W2 type blockage (28, 29, 47-61). The data are identified with respect to submergence by use of dif-ferent symbols. Furthermore, some tests were run with the flows in the outlet lines equal at two values, Ql and Q6, and at two surface elevations, Sl and S3.
The corresponding points are connected by solid lines in each figure. Since the outlet line flows were equal, these solid lines represent the behavior of the coefficient with Reynolds number at fixed flow patterns.
29 0 = W1 BLOCKAGE ge W2 BLOCKAGE
~
ENVELOPE FROM GENERIC TESTS (5)
NO OBSERVED SURFACE VORTEX ACTIVITY lm lm 332 Bmm 28 82 332 0
0 0.1 0.2 0.3 0.4 0.5 F = uses FIGURE 14 VORTEX TYPE VS. FROUDE NUMBER
1.8 1.8 1.6 1.6 S1 SUBMERGENCE H S2 SUBMERGENCE Q+ S3 SUBMERGENCE
~ 1.4 1.4 z
O 1.2 1.2 8
1.0 1.0 Qo Q
0.8 0.8 0.6 5 X 10 106 3 X 10 106 3 X106 REYNOLDS NUMBER, R REYNOLDS NUMBER, R A) BLOCKAGE W1 8) BLOCKAGE W2 FIGURE 15 INLET LOSS COEFFICIENT VS REYNOLDS NUMBER, TRAIN A
1.8 1.8 1.6 1.6 4 Sl SUBMERGENCE H S2 SUBMEBGENCE I- Q S3 SUBMERGENCE 1.4 1.4 O
O O
N 1.2 1.2 O-po 1.0 1.0 Qi 0.8 0.8 0.6 0.6 3x10 106 Q Q 106 3 x 106 REYNOLDS NUMBER, R REYNOLDS NUMBER, R A) BLOCKAGE Wl B) BLOCKAGE W2 FIGURE 16 INLET LOSS COEFFICIENT VS. REYNOLDS NUMBER, TRAIN B
32 The data show that the loss coefficient is generally inversely related to submergence. Since the measured head losses include screen loss, this de-crease of coefficient with increasing submergence is likely due to greater screen area exposure. For Train A, the data show no variations or slight decrease of. loss coefficient with Reynolds number. For Train B, the data show no variation or slight increase of loss coefficient with Reynolds num-ber. Within measurement uncertainty, it is probable that the coefficients are constant with Reynolds number as would be expected in this range.
The loss coefficients for blockage combination Wl were generally greater than for W2 and it is clear that blockage has significant effect. The largest val-ues recorded were K = 1.67 for Train A with blockage Wl (QA
= 4285 gpm, QB
= 8785 gpm, Sl = 21.0 ft)
K = 1 ~ 70 for Train B with blockage W2 (QA = 8785 gpm, QB = 4285 gpm, Sl = 21.0 ft) occurring at minimum submergence with the high coefficient associated with the outlet line carrying the lower flow. Since blockage Wl was selected on the basis of its corresponding high loss coefficients, these would be the highest values expected under the blockage selection scheme.
All measured loss coefficients (all tests) fell within the range 0.71< K < 1 70.
~
This range corresponds to generic studies (5) of containment sumps where the loss coefficients varied from 0.7< K < 1.6 with an average of K = 1.2.
Swirl angles for Train A are plotted against Reynolds number in Figure 17 for the type Wl and W2 blockage. A similar plot. for Train B was not made since those angles were less than about 1.5'n absolute value. It can be seen that
33 the swirl angle was substantially dependent on the type of blockage but showed no systematic variation with either submergence or Reynolds num-ber. The largest swirl angle observed was approximately 5.4'or type Wl blockage with the water surface at EL 23.0 ft at 4285 gpm flow in Train A and 8785 gpm in Train B (Q4).
For both trains together (all tests), the observed swirl angles varied from 0'o 5.74'n magnitude which falls within the 0'o 9'ange found by Padmanabhan (5) for 24 inch outlets.
bK S1 SUBMERGENCE H S2 SUBMERGENCE Q S3 SUBMERGENCE V) tU CC ~
U 4
O 3
K g
Qo Qo H C) gl Q Q 106 3 X 10 106 3 x 106 I
REYNOLDS NUMBER, R X x REYNOLDS NUMBER, R Fl A) 'LOCKAGE W1 8) BLOCKAGE W2 FIGURE 17 SWIRL ANGLE VS. REYNOLDS NUMBER, TRAIN A
SUMMARY
AND CONCLUSIONS Loss coefficients were found to depend on submergence and blockage, but to be essentially independent of Reynolds'umber over the parameter ranges and blockage configurations- tested. Values between 0.71 and 1.-70 were found and can be used for system hydraulic calculations, including NPSH.
Swirl angles were found to depend primarily on screen blockage and flow dis-tribution with no systematic variation with either submergence or Reynolds number. The largest value observed was 5.74', which can be compared to pump manufacturers specifications, taking the length and geometry approach piping into consideration.
Free surface vortex activity was found to be type 2, at worst. This reflect-ed the fact that only coherent surface swirls or dimples were observed; debris or air-ingestion were not observed in any tests.
The proximity of the sump screens to the pipe trench walls at the junction with the sump resulted in large amounts of flow passing through the avail-able space under certain types of screen blockage and flow distributions.
This did not appear to cause any particular problems, but only resulted in higher (overall) inlet loss coefficients.
36 REFERENCES
- l. Ebasco Drawings I ESF SUMP COMPOSITE SKETCH SK-2998-M-710 R6 II GENERAL ARRANGEMENT DRAWINGS 2998-G-065 R 2998-G-067 R III MECHANICAL PIPING DRAWINGS 2998-G-198 R7 2998-G-199 Sh 2 R6 2998-G-200 Sh 1 R6 2998-G-200 Sh 2 R7 2998-G-205 Sh 1 R5 2998-G"212 R4 2998-G-215 Sh 1 R2 2998-G-215 Sh 2 R2 2998-G-215 Sh 7 R2 IV STRUCTURAL DRAWINGS 2998-G-495 R5 2998-G-496 Sh 1 R4 2998-G"496 Sh 2 R3 2998-G-520 Rl 2998-G-521 R2 2998-G-797 Sh 4 R3 2998-G-797 Sh 13 R2 2998-G-797 Sh 14 R2 Dividing Screen Structural Sketch Rl SK - 2998-AS-226 V LINE LIST 2998-B-052 Rlo
- 2. Rouse, H., Handbook of Hydraulics, John Wiley 6 Sons, 1950.
- 3. Durgin, W.W., M. Padmanabhan, and C.R. Janik, "The Experimental Facility for Containment Sump Reliability Studies," ARL Report 120-80/M398, August 1980.
37
- 4. Lee, H.L., and Durgin, W.W.,'The Performance of Crossed-Vane Swirl Meters," Symposium of Vortex Flows, ASME Winter Annual Meeting, Chicago, 1980.
- 5. Padmanabhan, M. "Containment Reliability Studies," ARL Report No. 49A-82/
M398F, August, 1982.