ML20069H689
| ML20069H689 | |
| Person / Time | |
|---|---|
| Site: | 05200003 |
| Issue date: | 05/18/1994 |
| From: | Lythe G, Surry D WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | |
| Shared Package | |
| ML19304C237 | List: |
| References | |
| BLWT-SS4-1994, WCAP-14084, NUDOCS 9406130276 | |
| Download: ML20069H689 (34) | |
Text
__
m.w ' c.. x n -
n w a%.3 m.,&.;.4.p. @m u n n, m'.a,WNd/
nwm t y?.M&.
h C, Wrig.mi.
- m. n ~. g s,,, 4 l A r.
%',9,4 Rs*p'[' h I ".. s w
.. a - f Mt y n %
' r
- n.
.kr..*
9,,,
= +,
-, '..ik (,k. ;,. -- - -"
, es
. g h'd, a J.,.'g M 0
.e ?. y y a,.e,f
',, %. W f #' * .a'
-C,.
m L
S-.
.h
'."#.4'.'(
s y
' h $.,hf#
,,R.w+ s'. [n?$m$Abfr
. L.
e ; w., ' x.
y h*
L x
.c u n v-y w.
m
/
.n.
, WCAP-14084 [
Westinghouse Non-Proprietary Class 3 PHASE IVa WIND TUNNEL TESTING i
FOR THE WESTINGHOUSE AP600 REACTOR 1
r
/
,o..e w,
- w. ye v
4
,/
I Al'.
=4 g
!M -
5 '
E,.,
.t^3
,, i
- 9 v.'
j
%, q.
r yr n
6#
d
.,g m
~. <
'.rA w g
1l.
5' n.[
y
.gl1
.g.,
f i
?.gi g
,n.,...
z,
[m.'M\\
- '.em, I(
'I'*
[4
,(f.,4,,.,
r.
4
+
g pf g.',.,
s,
. fr av
.g; f,'Q 4, :.
4 4.- - *
,- w y-3.*,
.,s c,
3 r..w /..c ;
- n. h. >
- n. -;:.9
.-~;
s'
,y g h %. w; s
r#,..
M.J, - ', v-r.-
s.p' % s a
..M...A,u, '. >
~
-u s
7
',, f ;
.4,'
6:
hs
_ s,
s
- h y,'o. y YO y.A'.f_.
C sg
%.., t.*y b
- a
' y m.*A '1'
-4'.
.1
_ '.. s"
--[r*O
(
--k 4,
a
,s s
" N, r a
s b
., A e -
. J.l' /,.~ ^
.s.
e*.'
V'"
"s,s*
, 5,, _. f'
+r s %' F.
s*1'.~.4...g a
? 'l * '.
MF ge,y
,.,d..',%,
i. "-
1 e-v
.,74
,A
=r.e n
r.
-s
,y',=*
~
' I,
...3%
%,7. g <, 3l.' -**
I' f.
6,,.e.,..,m, l'., w
[ ;.,a
+
",*".48
. = ' 4; 1*
%.Aly itt.T;] % G & ?
\\..i 0, p -
o '** 3
],[ ,; ;*,
1 dn e@
+
I
- r y f w;<
L, n., -
- @ U,2.g g. 1,s.s,. J t,,.'
, "., y.' "., y,4 %. s '-
";y e,
a a.
'S r,j T
' ' ^ n a
-
- ew c
4 I& _, '
-.$s, 4
I' w,.b... _4,... '"* ',.., ; (a1.. N.,.
-,, ' ' ' g.f
., v.%
s.,,
ie
.m e
+
r 1
rqm
- 4., ". %
-1. u,.3
, e, -t.,
s e/
,.,/.,
A'"
t, + j *,,
- ,*,,,) '
e x
,o k
s r
M -
.d.*,.
'OT.3 k bg M
8
(
^ - ' ' '
N N @4.'e-
~
J Mfill' i i! Mi r ;*[-
",,. W $.h ~ ; \\.,n*, ~ -G 'n.:.
g u-m % W:.2 w r.-
f ;-
v i' ~.,r +,,s n.
r 2
J'
,. e :^ e
...9 O
a
- ,P*-"(',,
b, * ', p,,'.
9,
f-
-+/.
e, ~ _gy s.
"y..F c
4 y/-16 g *,
- as,i-.
g
,\\
,p
';n; *L, j
_y
- =
4
,' r y x(#'5f. w..h ;; #g sib 4,,.~.
j.
EE
~
'[
~ '
s g.d t
g w-. ~.m,., + n u,
a 4m w".c s.,
a.
' ++. g 4*
,,t,
., g,3
,v, j ffa
.g.ym g/
4.
3..
d i <.
P
, 'i, p
p,.
.. ;i E
' c
..s;-
3, *;Q g '
~ p, f*
V3'.,,
Mr*.W y 7,4 d Q y )r "h. o.,*: Q ~q ij k: y i. g g~ &e/
N.M
"- +
Y..pQ.Q c' )m.y[m?q, :QY. JR ','
.; T
_M. [: j 6
~
~
+k(
1 v,
m w~.&g,n@n'.Qg -
v--
5-(, (. 0 [.h? '64 % 3
.... m,~
)4
'.' b b I JJI hA d
.I,,
f."."';f;m<.ij,M,dhW.,9%r #.:
e W, _ \\'hi..,W.W, pM,+: * [@. c% 4",[r
,j.'~N *' 3.o ['0,
- r P '
.s sh s
W;q w~
. n..,, h. p r e.*g,. c r A. Q4 F g.
e-r 4,
.g, x.M.a. >p,g e. s wa..* g..
.c d h>.,YM 7 Q+ y.4'd $ k~}M.M,+ j, A. g%v.. y n
N o "h.s,h m'*'h %,>d.;[,
w m
. yu ;
4 -
.> <,~.
s
=,
- . 9,. n,.g.f a.,
n.2] [.e b O N b bh) n+-
Ud qo
-.-. MhMM2W7k~l k#.N k [,.
- Di' Q
~bb, kh,.,.
dh
% W N E W M
- k N" W D M*;g.c g n, g M n.
N s
x @4 9406130276 940606 W
W 4-s.g.
4WNAIMNp...., g,.,.
PDR ADOCK 05200003 uf gy Me N r.,
ppp--,_..).,
c -
me.w--...m.._
WESTINGHOUSE MON-PROPRIETARY CLASS 3 WCAP-14084
.l I
PHASE IVa WIND TUNNEL TESTING FOR THE WESTINGHOUSE AP600 REACTOR G. R. Lythe D. Surry BLWT-SS4-1994 t
3 i
4 BOUNDARY LAYER WIND TUNNEL LABORATORY THE UNIVERSITY OF WESTERN ONTARIO FACULTY OF ENGINEERING SCIENCE LONDON, ONTARIO, CANADA N6A 5B9
- 1994 Westinghouse Electric Corporation
TABLE OF CONTENTS PAGE
SUMMARY
111 ACKNOWLEDGEMENTS v
1 INTRODUCTION 1
2 EXPERIMENTAL PROCEDURE UWO TESTS 2
2.1 Modelling Of The Surrounding Site And The Wind 2
2.2 Modelling Of The Containment Building, Including The Flow Path 3
2.3 Pressure Measurements 3
3 EXPERIMENTAL PROCEDURE - NRC 1:30 SCALE TESTS 4
3.1 Modelling Of The Surrounding Site And The Wind 4
3.2 Modelling Of The Containment Building 5
3.3 Pressure hieasurernents 5
4 EXPERIMENTAL PROCEDURE - NRC 1:30 SCALE TESTS 5
4.1 Modelling Of The Surrounding Site And The Wind 5
4.2 Modelling Of The Containment Building 6
4.3 Pressure hieasurennents 6
5 EXPERIMENTAL RESULTS AND DISCUSSION 7
5.1 General 7
5.2 Main Results 7
5.3 Effects Of Tornado Profile 9
5.4 Effects Of The Cooling Tower 9
5.5 Residual Uncertainties 10 REFERENCES 12 TABLES 13 FIGURES 20
APPENDIX A - CALIBRATION OF FLOW LOSSES A1 APPENDIX B COMPUTER LISTING OF PRESSURE COEFFICIENTS B1 APPENDIX C - COMPUTER LLSTING OF ADJUSTED PRESSURE COEFFICIENTS C1 APPENDIX D -INVESTIGATION OF SPEED DEPENDENCY IN THE NRC 1:96 RMS D-1 DATA r
P E
h i
t 9
5 o
)
. _ ~.
=.
L
SUMMARY
This report details the first part of the fourth phase of the wind tunnel testing of the Westinghouse AP600 nuclear reactor (there was no phase III). The design for this reactor employs passive means for emergency cooling, including natural draft cooling and water film evaporative cooling. This cooling is dependent on natural convection through the building, which could be affected by wind conditions. Phase I testing examined the potential for wind-induced flows through the building and the sensitivity of the wind effects to various changes ia the geometry of the containment building and its surroundings. Phase I testing is detailed in reference 8. Phase II testing included the modelling of the complete flow path within the building and was used primarily to provide l
l information for the design of the baffle wall. His testing is detailed in reference 9.
I
(
l The airns of this phase of the testing are to:
1.
Check for any sensitivity of the measurements to Reynold's number, 2.
Provide final design data for the baffle wall:
3.
More accurately examine the effects of the hyperbolic cooling tower, 4.
Examine the effects of a uniform velocity profile for determining potential tornado loads, and; 5.
To provide some information for the modelling of the cooling tower in Phase IVb (these data will be presented in the report for Phase IVb).
To accomplish these aims, two models were used, and testing was done in two wind tunnels. He first model was the 1:96.67 scale model which was used in previous phases (in the "ct'rrent design" configuration, with internal flow panages modelled). It was modified to include more pressure tapping locations and to allow the l
internal flow path to be sealed at the base of the chimney. The second model was a new model constructed at a l
scale of 1:30. This model did not include the internal flow paths, but did include an internal volume to correspond l
to the internal volume of the sealed 1:96.67 model. The testing was done in two wind tunnels: the Boundary Layer l
l Wind Tunnel Laboratory at the University of Western Ontario (UWO), as in previous phases, and the 30' x 30' wind tunnel at the National Reevcb Council of Canada (NRC) in Ottawa, Canada. In both wind tunnels and at both scales, the models were tested in turbulent boundary layer flow representative of the flow in an open country terrain (i.e. ANSI exposure C), excwt for the tornado loading case, where near uniform flow was used. The configurations tested are listed in tabd 1.
The bigblights and main findings of the study are as follows:
L The effects of Reynold's number (Re) can be see, in the plots of figure 18. At the higher Re's tested, the results become insensitive to increasmg Re. To account for the changes in results between the Re used for the main UWO tests and the Re above which the results do not change, the main UWO results for the baffle loads were adjusted based on the data in figure 18.This amounted to an increase in the mean coefficients by a factor of 1.23 and no change in the rms coefficients. Peak loads were then formed from the mean and rms loads and an appropriate peak factor.
1 2.
A summary of the baffle loads are presented in figure 19. These loads are in the form of coefficients which, for design, must be muitiplied by a design dynamic pressure at roof height. Appendix D of reference 9 contains a discussion regarding the calculation of such design dynamic pressures.
3.
The cooling tower substantially reduces mean baffle loads, but increases the rms loads, resulting in someu bat 'arger peak loads. Although the peak loads are somewbat larger, they remain within the l
upper bound of figure 19, tbus the data do not warrant any " adjustment" due to the cooling tower, j
l l
iii a
P I
4.
The loading under a tomado flow simulation decreases from that under a " normal" flow simulation, due to the significant reduction in the unsteady loads. Since the design dynamic pressure for tornado winds is almost the same as for hurricane winds (see appendix D of reference 9), the baffle loads for the latter, based on figure 19, will be conservative for tornado loading.
r a
i
[
f r
i
?
I t
e I
t
?
i I
l 1
~{
't 7
i j
t s
.1
/
iv
l ACKNOWLEDGEMENTS This study was initiated by Westinghouse Electric Corporadon. The co-operation and interest of the Sponsors, m particular Ms. S. V. Fanto, are gratefully acknowledged.
Acknowledgement is also made of the contributions by various members of the technical staff of the Laboratory: Mr. S. Norman carried out the experimental phase of the study at UWO and instrumented the models, Mr. D. Morrish was responsible for the data acquisition system and assisted with the testing at NRC, and Mr. A.
Burggraaf assisted with the data acquisition system. The 1:96.67 test model was originally built and subsequently modified by members of the University of Western Ontario Coordinated Machine Shops. The 1:96.67 proximity model was built by Mr. G. Stevens. The 1:96.67 cooling tower model was built by Mr. J. Collishaw. The 1:30 test model was built by Mr. S. Horvath. Mr. J. Golab took the photographs of the UWO tests.
The considerable assistance and advice provided by members of the NRC staff are also gratefully f
acknowledged. In particular Messrs. S. Zan and K. Cooper coordinated the NRC experiments and designed the devices that were necessary to produce appropnate boundary layers in the 30' x30' wind tunnel. The operation of the 30' x 30' wind tunnel was performed by Mike and Cameron of the NRC technical staff. Mr. J. McKellar took the photographs of the NRC experiments.
Finally, the authors would like to acknowledge the general direction, advice and encouragement received i
from Dr. A.G. Davenport. Director of the Boundary Layer Wind Tuncel Laboratory.
i f
i V
j
F INTRODUCTION The Westingbouse AP600 nuclear reactor is designed to use passive means for emergency cooling. These include natural draft and water film evaporative cooling which are made possible by an air flow path through the containment building. The air flows in inlets at the top of the building, downwards past a baffle wall, then around the bottom of the baffle, upwards between the baffle and the containment vessel and out the chimney at the top of the building.
A goal of the design is that the wind not resist the air flow through the building. Phase I testing, detailed in reference 7, examined the effects of various design changes on the potential for wind. induced flows through the building. In that testing, the flow through the building was not modelled, but the pressure difference between inlets and chimney (i.e. the pressure driving any flow) was measured. In the Phase II tests, reported in reference 9 the air flow path was modelled for two different building designs: the most wind neutral design found in Phase I testing, and the current design of the building. The purpose of the Phase II testing was primarily to provide information for the design of the baffle wall. The information sought was the loads on the wall and how uniform the flow was at various points along the flow path. Buoyancy was not considered since the driving pressure due to buoyancy amounts to only about I to 5% of the wind-induced driving pressure for the design wind cases.
At the end of Phase II, there remained several outstanding questions. First, the effect of Reynolds number on the results. This could only be addressed definitively by testing a larger modelin a faster wind tunnel such that the Reynold's numbers were high enough that no further significant aerodynamic changes would be expected at full scale values. Secondly, the effect of a tornado wind profile (near uniform) on the results. This could be accomplished using the same test model as in previous phases, but with a different flow model. Thirdly, the effects of the hyperbolic cooling tower on the results. Some limited measurements were made in Phase II; however, the blockage of the cooling tower in the University of Western Ontario (UWO) wind tunnel was excessive. This question could be addressed by testing the 1:96.67 model in a larger wind tunnel where the blockage would be small. The current phase of testing, phase IVa (there was no phase III). is aimed at addressing these questions. A further question, the effect of severe terrain, will be the subject of Phase IVb.
Table I shows a summary of the tests that were performed as part of phase IVa. The tests can be subdivided into three series:
1.
The UWO tests. These tests (cases I to 4a in table 1), were performed using the 1:96.67 scale model used for previous phases, with some added pressure measurement locations. Tests were performed at the UWO wind tunnel, as in previous phases. Since testing in this wind tunnelis relatively economical. data were taken for a full range of wind azimutbs to form the basis of the design loads.
Data were also taken that would be used for comparison with data taken at the National Research Council of Canada (NRC) (see below). Finally, data were taken at a few wind angles in uniform flow to represent tornado loading conditions.
2.
The 1:30 NRC tests. For these tests, a new 1:30 scale model was built, without the internal flow passages and hence with fewer measurement locauons. The tests were performed in the 30' x 30' wind tunnel at the NRC for a range of speeds. The speeds ranged from a speed corresponding to the Reynolds number (Re) of the UWO tests, up to the maximum speed of the wind tunnel. The speeds and Re's used for all tests are listed in table 2. These tests would be used to examine Re effects. Note that the data for the lowest speed (case 5) are not expected to be as reliable as those for bigber speeds, but may be useful for observing trends.
3.
The 1:96.67 NRC tests. These tests were performed using the same model and instrumentation as was used for the UWO tests, but were performed in the 30' x 30' NRC wind tunnel for speeds ranging from the UWO test spead up to the maximum speed of the wind tunnel. Since the blockage of the hyperbolic cooling tower is neghgible at this scale in this tunnel, data were taken with the cooling I
tower in place in order to quantify its effect Two wind angles were chosen: one with the cooling tower directly upstream of the site and the second such that the sit? would be in the shear layer at the edge of the cooling tower wake.
All of the above tests were performed in turbulent boundary layer flow representative of the flow in an open country terrain (ANSI exposure C). In addition to the above measurements, measurements were made of the pressure distribution around the throat of the cooling tower and of the velocity distribution across the wake behind the cooling tower for use in Phase IVb. Those data will be reported with phase IVb.
2-EXPERIMENTAL PROCEDURE - UWO TESTS 2-1 Modelling of the Surrounding Site and the Wind The basic tool for these tests was the Laboratory's Boundary Layer Wind Tunnel. The 1:96.67 scale model of the ccntainment building was placed at the centre of a turntable and, for most of the tests, was surrounded by a " proximity" model consisting of the other plant buildings as shown in Figure 1 (some tests were done without the surroundings). This entire assemblage could be rotated to simulate different wind directions. The turbine building beight was 125.33 feet The model is shown in the wind tunnel m Figure 2.
The wind tunnel also allows extended fetches of coarsely modelled upstream terrain to be placed in front of the building under test. He wind tunnel flow then develops boundary layer characteristics representative of those found in full scalc. His methodology has been highly developed and is detailed elsewhere (1,2,3). In this case, a 1:96.67 scale boundary layer representadve of open country conditions (ANSI exposure C) was required.
To achieve this, spires and a trip were used along with floor roughnesses with heights of 0.75 and 2.0 inches. The upstream terram model is shown in Figure 3.
Vertical proRies of mean speed and the longitudinal component of the turbulence intensity, measured immediately upstream of the proximity model, are shown in Figure 4 compared with reference profiles for open country terrain. The latter have been developed by ESDU (4,5.6) through fitung theoretical models to available full scale data. The roughness length, z,,, (a charactensuc parameter of the mean speed proGle) calculated from the wind tunnel data, is approximately 0.02 metres. This is well within the acceptable range of a factor of 2 from the accepted median open country value of 0.03. The ratios of mean speeds at particular beights to those at roof height are shown in Table 3, along with similar reference values. Also shown are values of the local turbulence intensity, which is simply the root-mean-square (rms) speed divided by the mean speed at each height. The table includes heights up to 1.5 umes the building height. The table shows that except for the few points closest to the ground, the mean ratios are within 0.05 of the reference values and the local intensities are within 2 percentage points of the ESDU values. Hence this is a very good representation of the wind structure for an open country terram.
The simulation was funber checked by measurmg a spectrum of the wind speed at rooibeight. This spectrum is shown in Figure 5, along with the reference spectrum for open country terrain. The Ggure shows that the spectrum is well within the acceptable range of a factor of 2 over the entire range of wave numbers.
In addition to the main tests described above, a limited number of tests were done m a flow simulauon used to determine tornado loads. Figure 6 shows verucal profiles of velocity and turbulence intensity for this simulation, which is within the goal of a boundary layer less than 1/3 the beight of the building, with the flow over the upper 2/3 being uniform to within 5%.
2
22 Modelling of the Containment Building, including the Flow Path The model scale of 1:96.67 was chosen for Phase I to allow the model to be large enough to include small details and to be practical to work with. Also, the scale needed to be within the range of atmospheric model flows attamable in the wind tunnel and it was desirable to choose a scale that would accommodate standard material sizes to aid in the construction. The model built for Phase I and modified for use in Phase II, was further modified for the current tests by adding new circumferential nngs of pressure taps at the following locations:
1.
on the extenor of the main building at 2/3 the height of the inlets:
2.
on the extenor of the main building just below the inlets:
3.
just inside the inlet manifold; 4.
on the extenor 1/2 way up the chimney, and; 5.
at the top of the containment annulus.
These locations, along with all other pressure measurement locations, are shown as arrows in Figure 7.
Photographs of the model in the wind tunnel are shown m Figure 2.
In previces phases and for the UWO tests in the current phase, the extenor of the chimney was roughened. The reason for roughening was that since the full scale Reynold's number (Re) could not be matched in the tests, it was desireable to alter the flow around the chimney to more closely approximate the Dow expected at higher Re and to provide results that were insensitive to Re. Expenments in phase 1, based on the limited informadon available from the pressure taps inside the chimney (none had been installed on the exterior), and on tests done over a limited range of Re, indicated that roughening on the chimney produced the most appropriate results; without roughening on the cbimney, results changed with speed. It was recognized that roughening the chimney would increase its drag a little and bence make the chimney base pressure somewhat more negative; however, it was expected that rougbening the chunney would likely cause only second order effects on the results of interest, namely the pressures inside the chimney. These pressures would likely be dominated by effects of the separadon of the flow from the top of the chimney.
In modelling the flow path, the most important aspect to model is not stnctly the geometry, but the flow losses m the vanous parts of the Dow path. The loss values to model were derived from expenments conducted by Wesdnghouse Electnc Corporation (WEC), which are detailed m reference 7. In phase II, the model was designed to have lower losses than these values and calibration expenments and model modiGcations were perfonned to adjust the losses to the WEC values. Details of this modelling of the flow losses and the Phase 11 calibradon expenments are given in appendix A of reference 9. For the current phase IVa tests, the calibration expenments were repeated to ensure that the model modifications, which required partially dismantling the model, bad not appreciably changed the losses throughout the flow path. Results of this recalibradon are presented in appendix A of this repon.
23 Pressure measurements All pressure data were measured usmg a solid state pressure scanner system which sampled pressures at a rate approximating 5 samples per second in full scale for a penod approximating 22 minutes in full scale. All of these samples are kept for later analysis (e.g. to determme the maximum, mimmum, mean and rms values in each case). During the tests, the speed, denoted Vref, was monitored by a pitot-static tube mounted just upstream of the proximity model, at a beight near the roof of the wind tunnel. All of the pressure data are presented in this report in the form of non-dimensional pressure coefficients as defined in reference 1. They are referenced to the mean 3
dynamic pressure at roof heigbt,1/2 pV roof, where Vroof is calculated from the Vref measured during the test using the Vroof/Vref ratio from the measured velocity profile.
Pressures were measured at the locations sbown in Figure 7. As well, several combinations of the pressure measurements were made at each sampling mstant to form data records for new " combination" taps. The combination taps are all numbered in the 600's and are defined in Table 4. Many of these combinations are the same as those used in Phase II, while others are new combinations made possible by the new pressure tap locations. Combination taps 601 to 624 and 643 to 650 represent loads on the baffle at particular locations, since they are the difference between the pressure in tne shield building annulus (SB.A) and the pressure in the containment annulus (CA). Combmation taps 625 to 632 represent the difference between the pressure at the inlet taps and the pressure at the upper level SB A taps. Combination taps 633 to 640 represent somewhat approximate baffle loads at the top, since they are the difference between SB A and C A taps that are not directly opposite one another (taps 643 to 650, which were unavailable for Phase II, are better taps to use here). Three of the most important combination taps are the average of the inlet taps (tap 651), the average of the chimney taps (tap 652)
~
and the difference between the average of the inlet taps and the average of the chimney taps (for this test, tap 642; in phase II tap 641 was used).
3-EXPERIMENTAL PROCEDURE - NRC 1:30 SCALE TESTS 34 Modelling of the Surrounding Site and the Wind The basic tool for tese tests was the 30' x 30' wind tunnel at the NRC. The 1:30 scale model of the containment building was bolted to the floor of the wind tunnelin a position such that the surrounding buildings could be added if necessary. For this reason, it was not centrally located. Provision was made for including surrounding buildings if significant Re number effects were found without the surroundings, based on the rationale that Re sensitivity is likely to reduce as interference effects from the surroundings increase. The model is shown in the wind tunnel in figure 8.
In the 30' x 30' wind tunnel the boundary layer is generated by the use of spires and trips. In this case, a 1:30 scale boundary layer representative of open country conditions (ANSI exposure C) was required. The spires and trip used to achieve this are shown in figure 9.
Vertical profiles of mean speed and the longaudinal component of the turbulence intensity, measured on the centreline of the wind tunnel without the model present, are shown in Figure 10 compared with reference profiles for open country terrain. The htter have been developed by ESDU (4,5,6) through fitting theoretical models to available full scale data. The roughness length, z,, (a characteristic parameter of the mean speed profile) calculated from the wind tunnel data, is approximately 0.03 metres. This is well within the acceptable range of a factor of 2 from the accepted median open country value of 0.03. The ratios of mean speeds at particular heights to those at roof beight are shown in Table 5. along with similar reference values. Also shown are values of the local turbulence intensity, which is simply the root-mean-square (rms) speed divided by the mean speed at each height. The table includes bcights up to 1.5 times the building height. The table shows that except for a few points close to the ground and the two bigbest points, the mean ranos are within 0.05 of the reference values (where they differ, the actual values tend to be bigher than the reference values and hence are conservative). The local intensiues are all within 2 percentage points of the ESDU values. Hence this is a very good representation of the wind structure for an open country terrain.
The simulation was further checked by measuring a spectrum of the wind speed at roof height. This spectrum is shown in Figure 11, along with the reference specuum for open country terrain. The figure sbows that the spectrum is weM within the acceptable range of a factor of 2 over the upper range of wave numbers; however, 4
1
it is missing some of the low frequency fluctuations. This was considered acceptable since the overall energy in the flow as measured by the turbulence intensity is comparable to the reference values.
3 2 Afodelling of the Containment Building The model scale of 1:30 was chosen with the aun of achieving the highest Re possible. This scale was the largest scale for which a reliable boundary layer could be generated in the NRC wind tunnel. In order to examine the effects of the chimney roughening, tests were performed with and without chimney roughening. He size of the roughness used was the same relative size as for the 1:96.67 scale test. Photographs of the modelin the wind tunnel are shown in Figure 8.
33 Pressure measurements All pressure data were measured using a solid state pressure scanner system which sampled pressures for a period approximating 20 minutes in full scale. The pressures were sampled at a rate approximating 9 samples per second in full scale at tiae highest wind speed tested and about 70 samples per second at the lowest wind speed tested (in the analysis, only frequencies up to 2 bz in full scale are used). All of these samples are kept for later analysis (e.g. to determine the maximum, minimum, mean and rms values in each case). During the tests, the speed, denoted Vref, was monitored by a pitot-static tube mountedjust downstream of the test model, at a height l
well above the test model. All of the pressure data are presented in this report in the form of non-dimensional pressure coefficients as defined in reference 1. Rey are referenced to the mean dynamic pressure at roof height, 1/2 pV roof, where Vroofis calculated from the Vref measured during the test using a Vroof/Vref ratio measured l
as part of the velocity profile.
I Pressures were measured at the same, but fewer locauons than on the 1:96.67 model. Five circumferential rings of tapping locations were used (see Figure 7):
1.
Ring 1 is at level 2a. on the exterior of the main building at 2/3 height (taps 85 to 100);
2.
Ring 2 is at level 4a. on the extenor of the mam building,just below the inlets (taps 109 to 124);
3.
Ring 3 is at level.ia just inside the inlets (taps 125 to 140);
4.
Ring 4 is at level 6, on the exterior of the chimney (taps 141 to 148);
5.
Ring 5 is at level 6, in the interior of the chimney (taps 149 to 156).
As well, several of the combination taps. as described in section 2.3, were formed; namely, the average of the inlet taps (tap 651), the average of the chimney taps (tap 652) and the difference between the average of the inlet
{
taps and the average of the chimney taps (tap 642).
4-EXPERIMENTAL PROCEDURE - NRC 1:96 SCALE TESTS 41 Afodelling of the Surrounding Site and the Wind The basic tool for these tests was the 30' X 30' wind tunnel at the NRC. The 1:96.67 scale model of the containment building that was used in the UWO tests was used here with the same instrumentation and the same
' surroundings model. The test model and the surroundings model (when used) were mounted on a 1" thick base that extended 6" beyond the outer edge of the surroundings model. De protruding 6" were sloped down to floor 5
level at the outer edge. The base was used to allow more efficient model changeovers (e.g. turning the model to simulate another wind angle) since the models could be screwed to the base instead of having to be bolted to the wind tunnel floor. The presence of the base is not expected to significantly affect the results, since it does not raise the model significantly and since it is relatively close dowTistream from the flow trip (see below). The model is shown in the wind tunnel in Figure 12.
In the 30' x 30' wind tunnel the boundary layer is generated by the use of spires and trips. In this case, a 1:96.67 scale boundary layer representative of open country conditions (ANSI exposure C) was required. The spires and trip used to achieve this are shown in figure 13.
Vertical profiles of mean speed and the longitudmal component of the turbulence intensity, measured on the centreline of the wind tunnel without the model or base present, are shown in Figure 14 compared with reference profiles for open country terrain. The laner have been developed by ESDU (4.5,6) through fitting theoretical models to available full scale data. The roughness length, z, (a characteristic parameter of the mean o
speed profile) calculated from the wind tunnel data. is approximately 0.03 metres. This is well within the acceptable range of a factor of 2 from the accepted median open country value of 0.03. The ratios of mean speeds at parucular beights to those at roof beight are shown tn Table 6, along with similar reference values. Also shown are values of the local turbulence intensity, which is stmply the root-mean-square (rms) speed divided by the mean speed at each height.The table includes heights up to 1.5 umes the building height.The table shows that for all heights the mean ratios are within 0.05 of the reference values and the local intensities are all within 2 percentage points of the ESDU values. Hence this is a very good representation of the wind stn:cture for an open country terrain.
The simulation was furtber checked by measunng a spectrum of the wind speed at roof beight. This spectrum is shown in Figure 15, along with the reference spectrum for open country terrain. The figure shows that the spectrum is within the acceptable range of a factor of 2 over the upper range of wave numbers; bowever,it is missing some of the low frequency fluctuations. Thts was considered acceptable since the overall energy in the flow as measured by the turbulence intensity is cornparable to the reference values.
42 Modelling of the Containment Iluilding The same 1:96.67 scale model that was used for the UWO tests was used bere (see section 2.2). Chimney rougbenmg was used for the initial tests and then removed to provide companson data. Since the model had not been altered smce the UWO tests, the flow losses through the flow path did not need to be recalibrated; they will remain the same as those detailed in appendix A. Pbotographs of the model in the NRC wind tunnel are shown in Figure 12.
43 Pressure measurements All pressure data were measured using a. solid state pressure scanner system which sampled pressures for a period approximating 20 minutes in full scale at the lowest wind speed tested and about 35 minutes at the bigbest wmd speed tested. The pressures were sampled at a rate approximating 3 samples per second in full scale at tbe highest wind speed tested and about 6 sranples per second at the lowest wind speed tested (in the analysis, only frequencies up to 2 bz in full scale are used). Al.1 of these samples are kept for later analysis (e.g. to determme the maxunum, mimmum, mean and rms values in each case). During the tests, the speed, denoted Vref, was monitored by a pitot-static tube mountedjust downstream of the test model, at a height well above the test model. All of the pressure data are presented in this report in the form of non-dimensional pressure coefficients as 2
defined in reference 1. They are referenced to the mean dynamic pressure at roof beight,1/2 pV roof, where Vroof is calculated from the Vref measured during the test using a Vroof/Vref ratio measured in a separate experiment after the testing (with the model base in place, but without the test model or surroundings).
6
Pressures were measured at the same locations as for the UWO tests (see Figure 7) and the same combination taps were fortned (see table 4).
5-EXPERDIENTAL RESULTS AND DISCUSSION 51 General
. For all tests, statistics of the pressure coefficient records (maximum, mmimum, mean and rms) have been determined for all taps and are tabulated in Appendix B. Taps are numbered in accordance with the numbering system shown in Figure 7 and Table 4. These data are examined more closely in the following sections.
Some of the current data frotn the UWO tests (case 4) are comparable to data taken in Phase II. Figure 16 shows some comparisons; the repeatability is quite good, lending confidence to both studies.
He main purpose of these tests was to determine the best estimates of the loads on the baffle wall inside the containment building under normal conditions, and then to determine the effect of a tornado wind profile and the cooling tower on these loads. The most complete set of data is the UWO test data. The NRC data was obtained to assess concerns with the UWO data regarding Re scaling and blockage and can be used to determine if any adjustments. ire necessary.
Since the baffle loads are of primary interest and because the 1:30 scale model did not include the internal flow path, the data that are most important for observing trends are the data from the following three
" combination" taps: the spatial average of the inlet taps (tap 651), the spatial average of the inside chimney taps (tap 652) and the spatial average inlet minus spatial average chimney tap (tap 642). The last of these is the most important of all. If further information on trends seen in the data from these taps is needed, then the data from the j
rings of taps on the outside of the building and chimney can be used.
1 52 Main Results Figure 17 shows the data from the three primary taps for all cases, referenced to the case numbering in tables 1 and 2. His figure gives a good indication of the overall vanability of the results over all Re's, with and without chimney roughness, with and without surroundings, with the chtmney open and closed and in three e
different simulations.De last two sets of data show the effect of adding the cooling tower.He main observation from this plot is that even with all of the differences between configurations, there is not a great deal of vanability i
in the data. Trends with Re can be seen, as can open/ closed chimney trends and smooth / rough chimney trends. but differences tend to be small.
I In analysing the data for the NRC 1:96 scale tests, it was observed that the rms data increased with wmd speed, while the mean data remamed constant. It was determmed that this was pnmarily due to acoustic noise in the wind tunnel itself and not a real trend in the data. Details of this investigation are given in appendix D.
In order to determine if the UWO data needs to be corrected for Re effects and/or other effects, mean and rms data for matebed closed chimney, no surroundings cases have been plotted in figure 18 vers. 4 the main I
building Re for the three pnmary taps.The following observations can be made from figure 18:
1.
For the inlet tap, the chimney rougbening has very little effect.
7
2.
At the inlets, there is some variation of the mean pressure with Re up to about 1.2 x 10', but very little above this. The rms pressure shows little variation with Re.
3.
At the inlets, the difference between the mean UWO data and the high Re mean data is small; a Cp difference of-0.03 or a ratio of 1.17. For rms data, the difference is negligible.
At the chimney, the effect of roughening is significant for the lower Re, but less so at the higher Re.
4.
Roughening tends to make the variations at low Re's less severe and more predictable.
At the chimney, there is significant variation of the mean pressure with Re up to about 2 x 10' and 5.
some resideal variation above this. Note that the Re used is based on the containment building diameter. Relative to the local chimney diameter, the Re would be reduced by a factor of about 1.8.
The equivalent Re to the onset of stable results at the inlets as observed above in point 2 would be about 2.2 x 10', consistent with the results of figure 18b for the chimney. For the roughened chimney, the pressures appear to have stopped changing at the highest Re tested. It can also be seen (from data for tap 147) that the magnitude of the wake pressure is smaller for the rough chimney r
than for the smooth chimney. At very high Re, where the wake pressure is independent of further changes in Re, roughness is known to increase the wake pressure. Hence,it is expected that the smooth chimney results would eventually approach the roughened chimney value, or something slightly less negative, at higher Re's. The rms pressure shows smaller variation with Re.
6.
At the chimney, the difference between the mean UWO data and the rough chimney mean data at the highest Re is a Cp difference of 0.16 or a ratio of 1.23. For rms data, the difference is negligible.
7.
For the inlet minus chimney tap, the effect of roughening is significant, since it retlects the effects seen at the chimney. Roughening tends to make the variations at low Re's less severe and more predictable.
8.
For the inlets minus chimney tap, there is significant variation of the snean pressure with Re up to about 1.2 x 10' and some residual variation above this. For the roughened chimney, the pressures appear to have stopped changing at the highest Re tested and it is expected that the smooth chimney results would eventually approach this value at higher Re's. The rms pressure shows smaller variation with Re.
9.
For the inlet minus chimney tap, the difference between the mean UWO data and the rough chimney mean data at the highest Re is a Cp difference of 0.12 or a ratio of 1.23. For rms data, the difference is negligible.
The conclusion from the above observations is that the UWO data requires a correction to the mean inlet minus chimney difference of a factor of 1.23. The rms requires no correction. This factor has been derived from closed chimney cases, but it is assumed, reasonably, that it can be applied to the relevant open chimney case.
For conservatism, the total inlet minus chimney correction can be applied to all pressure differences across the baille wall. Since the mean data are to be adjusted, the peak data (maximums and minimums) will have to be formed as follows:
Cp(maximum) = adjusted Cp(mean) + g
- Cp(rms)
Cp(minimum) = adjusted Cp(mean)- g
- Cp(rms) where g is an appropriate peak factor. This peak factor was calculated as the average peak factor from all positive and negative peaks from all relevant combination taps on the bafilw (i.e. taps 601 to 624 and 643 to 650) for all wind angles. It was calculated to be 4.58. This process also tends to smooth out some of the variability associated with the single. point peak results.
8 I
l i
The final pressure coefficients relevant to the baffle design, adjusted as) it described, are tabulated in appendix C. A summary of the worst values is given in table 7. Figure 19 shows i summary of this data for each wind angle, the largest positive (inward acting) and negative (outward-acting) feak coefficient was found from all taps at each level and is plotted at that level (the positive and negative peaks have been slightly offset vertically for clanty). The figure includes an upper bound line to the data showing linear behaviour for the lower portion of the baffle, extending to the centreline of the mlets where pressure differences would be expected to be maximum.
Above this, the upper bound pressure difference is shown as constant, nis upper bound line has shifted to the right somewhat from the corresponding line from Phase II, largely as a result of the adjustment to the current data as desenbed above.
53 Effects of the Tornado Profile Figure 20 shows the data from the tornado profile tests superimposed on comparable data from a
" normal" profile test (neither set of data includes the adjustments discussed above), for the combination taps on the baffle. De figure shows that in all cases the peak pressure coefficients are lower for the tornado case, pnmanly resulting from the reduction in unsteady components. This means that the design values of figure 18 would be conservative under tornado conditions.
54 Effects of the Cooling Tower In the NRC 1:96.67 scale tests, tests were done at 310* both with and without the cooling tower. This angle represents the condidon where the containment building is on the edge of the wake from the cooling tower.
I A test was also done with the cooling tower in place for an angle (280*) representing the condition where the containment building is directly downstream of the cooling tower.
i Figure 21 shows a comparison of the with and without cooling tower data at 310'. Figure 21a includes I
data from all 3 speeds, while Figure 21b includes data from only the lowest speed. Each point on these plots
. represents data from one tap under two configurauons; the x coordinate of the point is the Cp for the tap without the cooling tower in place and the y coordinate of the point is the Cp for the tap with the cooling tower in place.
Tbus potnts that fall on the solid 45* line indicate taps where the data is the same for the two configuradons; points above the 45* !ine indicate taps where the Cp with the cooling tower is greate. than the Cp without it. Data for baffle difference taps (taps 601 to 624 and 643 to 650) are shown. A least squa.es fit to the data, forced through zero, is shown as a dashed line and its parameters are shown in the lower right corner of each plot. The slope indicates the ratio of with to without cooling tower data (i.e. a slope greater than I would indicate that Cp's are greater with the cooling tower than without) and the correlation indicates how well the line fits the data (a correlation of 1.0 would indicate all data pomts fall on the line).The actual numerical values for the data from the bigber two speeds are not as reliable as the data for the lowest speed, since the bigber speeds were subject to the acoustic noise discussed above and in appendix D; however, the comparison between two cases for the two bigber speeds should still be valid, since both cases were subject to the same noise, which only constituted 5 to 10% of the rms signallevel(hence less of the peak values).
Figure 21a indicates that the cooling tower actually reduces the peak loads somewhat, with a large reducuon in the mean load and little change in the ras. Figure 21b, which is based on the mest reliable data i
indicates that with the addition of the cooling tower, positive peak loads are reduced, negauve peak loads are increased, mean loads are substandally reduced and rms loads are increased. Although some peak loads are increased, they remain within the upper bound line of 6gure 19.
Similar comparisons for 280* are somewhat more difficult to make, since data were not taken at NRC l
without the cooling tower at this wmd angle. Therefore, UWO data without the cooling tower for this angle must be used for the companson; bowever, the configurauon is slightly different(UWO mcludes chimney roughening).
9
)
i I
To take into account differences in configurauon, data at azimuth 310" from both wind tunnels, without the cooling tower, were compared in a similar manner to that shown in figure 21.The slopes from the least squares fits were used to "conven" the UWO data for 280* to equivalent NRC data for 280*.These adjusted data are compared with the NRC data that include the cooling tower m figure 22.This figure agam shows a substantial reduction in the mean load and a slight increase in the rms.The posiuse peak loads are decreased and the negative peak loads are increased slightly. Although some peak loads are increased, they remam within the upper bound line of figure 19.
The conclusion from the above is that the cooling tower reduces mean loads but increases the rms loads.
resulting in somewhat larger negative peak loads and somewhat smaller posiuve peak loads. It is important to note that it is the positive peak loads that detine the upper bound in figure 19 and that in all cases examined the cooling tower reduces these loads, altbough sometimes at the expense of a smallincrease in the negative peaks.
Tbus the largest peak loads observed with the cooling tower remain within the upper bound of figure 19. and no adjustments to the data for the presence of the cooling tower are warranted. Note that this trend of reduced mean loads and increased rms loads is similar to that seen in Phase II: bowever, the Phase Il cooling tower peak loads were exagerated due to blockage effects.
5 5 Residual Uncertainties The Phase 11 and Phase IVa testing bave investigated the wind-induced pressures on the AP600 reactor in a very comprebensive manner, consistent with the current state-of-the-art in wmd engineenng. The extrapolation of these test results to full scale appears to be very reasonable in the light of the arguments presented in the preceding sections and the significant degree of repeatability that has been obtamed oser a range of test condiuons. In this context, it is worth bearing in mmd a number of factors that contribute to the differences observed, and also that contribute to the confidence that the primarv variadons in results have been explored.
Residual differences in the test results can be associated with a number of factors.
1.
The three wind simulations used all differ slightly, both in their pnmarv characteristics of mean speed distnbution away from the reference speed at the top of the buildings and their local turbulence intensity distnbutions as well as m their secondary charactensucs associated with the spectral distnbution of the turbulence. There are also other charace.nir* r.ssociated with cross-tlow components, spatial non-uniformiues etc. that were not ma% red, but undoubtedly differ a little from case to case and contribute to small differences m the resulung aerodynamic responses.
2.
Model locations within the wind tunnels differed, and blockage effects idiscounting the unacceptable blockage associated with the cooling tower m the UWO tunnel) were always present to differing degrees. No corrections were attempted for these, but it is known that they tend to lead to conservative results due to the increased acceleration of the fiow past the model and its wake.
3.
Over the vanety oflength scales and speeds that were used in the tests, record lengths were mamtamed at approximately the same full scale value tabout 20 mmutes). The sampling rate was also vaned according to the time scaling wbere feasible to approxtmate a constant full scale value.
Nevertheless, differences did occur. Moreover, many of the expenmental frequency response charactertsues mevitably remamed constant. effecuvely changmg thett associated full scale characterisucs between tests. These differences will have added vanability to the natural vanability associated with the random processes exammed.
10
It is difficult to quandfy precisely these various factors. Perhaps their best measure is the degree of consistency in the overall results as presented in the text. In extrapoladng these results to full scale, which is an additional factor of about 200 in Reynolds number. a number of positive factors should be borne in mind.
1.
The tests carried out here extend to Reynolds numbers in the so-called transcritict.! range for smooth flow around two-dimensional circular cylinders. It is also well-established that the presence of shear, turbulence and three-dimensionality all act to reduce the effects of Reynolds number. All these factors are strongly in evidence for the AP600 system.
2.
Full scale wind characteristics vary considerably, depending on terrain and other elements of surface roughness. The tests carried out here simulate an open country exposure, which is generally conservative for loads. Mean and peak wind speeds are reduced as terrain roughness increases.
Special terrain cases are considered further in Phase IVb.
A SUhBIARY OFTHIS REPORT APPEARS ATTHE BEGINNING l
l 1
11 o
REFERENCES Davenport. A.G. and 1syumov. N.,"The Applicauon of the Boundary Layer Wind Tunnel to the 1.
Prediction of Wind Loading", Intemauonal Research Seminar on Wind Effects on Buildings and Structures Ottawa, Canada. September 1967 University of Toronto Press,1968.
2.
Whitbread. R.E., "Model Simulauon of Wind Effects on Structures". NPL Intemational Conference on Wind Effects on Buildings and Structures, Teddington, England,1%3.
Surry, D. and Isyumov, N., "Model Studies of Wind Effects - A Perspective on the Problems of 3.
Experimental Technique and Instmmentation",ICIASF-75 Ottawa, Sept.1975.
ESDU, " Characteristics of atmosphenc turbulence near the ground. Part I: definitions and general 4.
information". Item number 74030. ESDU international Ltd, London England,1974.
ESDU. " Characteristics of atmosphenc turbulence near the ground. Part 11: single point data for 5.
strong winds (neutral atmosphere)". Item no. 85020. ESDU International Ltd, London,1985.
ESDU. " Strong winds in the atmospberic boundary layer. Pan 1: mean hourly wind speeds", item no.
6.
82026. ESDU International Ltd, London England,1982.
7.
Stewart, W.A. and Pieczynski, A.T., " Tests of Air Flow Path for Cooling the AP-600 Reactor Containment", Westingbouse Electnc Corporation, Report 88-8E9-ADLWR-R2, March 28,1988.
Lythe, G.R. and Surry, D., " Phase 1 Wind Tunnel Testing for the Westingbouse AP600 Reactor",
8.
The University of Western Ontano. BLWT-SS36-1991 Lytbe, G.R. and Surry, D., " Phase 11 Wind Tunnel Testing for the Westinghouse AP600 Reactor".
9.
The University of Western Ontano, BLWT-SS15-1992 12
The information contained in the following Tables, Figures, and Appendices is classified Westinghouse Proprietary Class 2, and thus, has been excluded from this non-proprietary version of this report:
TABLE:
2 Reynold's Numbers for Phase 4A 3
Comparison of Velocity and Turbulence Intensity Profiles for the UWO Tests 4
Defmition of Tap Combinations 5
Comparison of Velocity and Turbulence Intensity Profiles for the 1:30 Scale NRC Tests 6
Comparison of Velocity and Turbulence Intensity Profiles for the 1:100 Scale NRC Tests 7
Phase 4A (UWO 1:96) - With Surroundings, Chimney Open & Rough, Extreme Values of Pressure Coefficients Over All Azimuths i
FIGURE:
4 Vertical Profiles of Mean Wind Speed and Turbulence Intensity Compared with Theoretical Profiles for Open Country Terrain - UWO 1:96.67 Tests 5
Spectrum of Velocity at Roof Height Compared with Theoretical Spectrum for Open Country Terrain - UWO 1:96.67 Tests 6
Vertical Profiles of Mean Wind Speed and Turbulence Intensity for the Tornado Wind Simulation - UWO 1:96.67 Tests 10 Vertical Profiles of Mean Wind Speed and Turbulence Intensity Compared with Theoretical Profiles for Open Country Terrain - NRC 1:30 Tests 11 Spectrum of Velocity at Roof Height Compared with Theoretical Spectrum for Open Country Terrain - NRC 1:30 Tests 14 Vertical Profiles of Mean Wind Speed and Turbulence Intensity Compared with Theoretical Profiles for Open Country Terrain - NRC 1:96.67 Tests 15 Spectrum of Velocity at Roof Height Compared with Theoretical Spectrum for Open Country Terrain - NRC 1:96.67 Tests 16 Comparison of Pressure Coefficients From Between Phase II and Phase IVa Testing 17 Pressure Coefficients for Average Inlet Tap (651), Average Chimney Tap (652), and Inlet Minus Chimney Tap (642) Versus Case Number 18 Pressure Coefficients From Closed Chimney Cases 19 Distributian of Worst Pressure Differences Across the Baffle 20 Compariscu of Pressure Coefficients From the Tornado Case With Those From the Comparable Non-Tornado Case 21 Comparison of Pressure Coefficients With and Without the Cooling Tower (Azimuth 310) 22 Comparison of Pressure Coefficients With and Without the Cooling Tower (Asimuth 280)
APPENDIX:
l A
Calibration of Flow Loises B
Computer Listings of Pressure Coefficients i
C Computer Listings of Adjusted Pressure Coefficients D
Investigation of Speed Dependency in the NRC 1:96 RMS Data 1
i
I TABLE 1 - PHASE 4A CONFIGURATIONS CASE
- TUNNEL SCALE SURROUND-COOLING CHIMNEY ANGLES NO.
TORNADO?
1 INGS?
TOWER?
SPEEDS i
1 UWO 1:96 NO NO
- OPEN, 315 1
NO ROUGH 2
- CLOSED, i
ROUGH 3
YES
- CLOSED, FULL 360 ROUGH
+ 315 4
- OPEN, ROUGH 4a
- OPEN, 315,225,
YES ROUGH 135 5 - 10 NRC 1:30 NO NO
- CLOSED, 315 6
NO SMOOTH 11-15
- CLOSED, 5
ROUGH 16-18 NRC 1:96 NO NO
- CLOSED, 315 3
NO ROUGH 19-21
- CLOSED, SMOOTH 22-24
- OPEN, j
SMOOTH 25-27 YES
- OPEN, SMOOTH j
28-30 OPl.N, 310 SMOOTH 31-33 YES
- OPEN, SMOOTH 34-36
- OPEN, 280 1
l SMOOTH 37 NRC 1:96 COOLING TOWER PRESSURE DISTRIBUTION AND WAKE
- For cases with multiple speeds, cases are numbered from lowest to highest speed.
1 l
1
\\
J 4
i i
i
!!I l
llI p
1 T
N A
N LP t
0 g
n 0
g n
e n 6
di ni P
inl A
d l
i iou N
i u
b tb n
A e
o i
c R
n b
O r
F u
G t
N I
WAR H
W D
T U
OYA L
I E
RU G
IF g r n e i w lo o ot c
i N
k j
./..
'. g '
1
(
l J
i l
j g
que
.. ~. 4' - @ e 'k,
e-
- I j
M.
\\
\\'
/
l 1
1 1
I u
1 w
1 i
1 1
i T*
7,'
I, i i i yin -- -
1 s
..a w
-..w
~
'5 ?
~
w
~
.i i
l i
l l
1 l
FIGURE 2 PHOTOGRAPHS OF THE MODEL IN THE WIND TUbHEL l
UWO 1:96.67 TESTS i
3 L.-.. -.....
.I
y the e
- w..
_ ~
7
?
ag.
A
,--(7
~-
3 2
4 4
l
=
+
FIGURE 3 VIEW OF THE MODEL IN THE WIND TUNNEL SHOWING THE UPSTREAM TERRAIN MODEL USED UWO 1:96.67 TESTS
8 l V/
LEVEL 6
-~
9 y
t
/
LEVEL 5 I
),,,
b
/
s LEVEL 4A A
x
/
s l
l LEVEL 4
--~
l s
LEVEL 3
~f
- +--
t I
s s
/
)
{
]
LEVEL 2A
--+-
l l
l l
l s
s h
l s
i, l
LEVEL 2
-~
-+
s I
l l
l 1
s'
.I s
l s
LEVEL 1 A
!~
~h l
l s
s 1
l
- I:::
s s
l l
s s
s 8
s g
j l
l s
s STAIRVELL TYPICAL LONGITUDINAL SECTION 1
FIGURE 7a VERTICAL SECTION OF THE MODEL SHOWING PRESSURE TAP LEVELS
9 u
11 7
a 10 l
12 8
l l
I I.
13 ~:
W 14
- - - - -l- - - - -
6 -ai f=- 5 e
l l
16 4
2 Q
u 15 3
h 1
LEVEL 1 FIGURE 7b HORIZONTAL SECTION OF THE MODEL SHOWING PRESSURE TAP NUMBERING - LEVEL 1
25 9
27 23 a
26 28 24 l
22-__ [
- --- f ----
h~ 30 29 -+:
21 e
I i
32 20 1
18 l
v -
31 19 n
i 17 LEVEL 2 FIGURE 7c HORIZONTAL SECTION OF THE MODEL SHOWING PRESSURE TAP NUMBERING - LEVEL 2
93 92 94 91 95 90 96 89
.I
y.---
e l
97 i
88
)
98 87 l
)
m 99 86 100 85 LEVEL PA I
1 l
FIGURE 7d HORIZONTAL SECTION OF THE MODEL SHOWING PRESSURE TAP NUMBERING - LEVEL 2A
41 9
43 39 Ji 42 N
44 40 i
I h
[+- 4 6
h----
38 -+i
- +--37
-iz-f:=-
45--a j
i i
48 36 34 i
U 47 35 JL 33 LEVEL 3-FIGURE 7e HORIZONTAL SECTION OF THE MODEL SHOWING PRESSURE TAP NUMBERING - LEVEL 3
53 52 105 104 54 51 06 10 l
s
_. _ _ 4. ___ _ _
e l
107 l
10 l
55 50 108 101 56 49 LEVEL 4 FIGURE 7f HORIZONTAL SECTION OF THE MODEL SHOWING PRESSURE TAP NUMBERING _ LEVEL 4
117 U
133 132 y
134 119 13 11 5 x gay 135 65 66 64 130 12 Il4 67 63 68 62 l
129
f ----
121 h
$ 69 61 l %
11 3 e
f 137 l
70 i
60 128 71 59 12 12 138
, 72 58 57 12 7
/,
s 123 111 139 126 140 8
125 124 110 109 LEVEL 4A
)
i FIGURE 7g HORIZONTAL SECTION OF THE MODEL SHOWING PRESSURE TAP NUMBERING - LEVEL 4A I
77 144 145
/
78 NibV 14 3 o
152' N~/
7 15g 14 6 f5 l
1513 e
o 14 2 s
14 7
$4
/
141 g
148 73 l
l i
(
LEVELS 5 L 6 l
FIGJRE 7h HORIZONTAL SECTION OF THE MODEL SHOWING PRESSURE TAP NUMBERING - LEVELS 5 AND 6
- j:
i Ji vi EM g g,.[-
M h L-.23
__.2 M E Eca a
a bs..
.N
~.
..)
~-
A s
m i
FIGURE S PHOTOGRAPHS OF THE MODEL IN THE WIND TUNNEL NRC 1:30 TESTS
)
I k
/
Er I
i mama 1
~
4 1
8
/.
I l
/
l l
1 i
w
.-m i
N _,
s__
e x (Y
k' l
It l
~
J3g.
l
- -l l l
l l
l i
l FIGURE 9 VIEW OF THE MODEL IN THE WIND TUNNEL SHOWING THE UPSTREAM TERRAIN MODEL USED NRC 1:30 TESTS
a yn, y
~..
-.3-f
~,
- - c 'dii,r:
/
m 4
'
- W:~: ;:.yg6-fg.~
n..
~
w t
- j. '
_,N
' g:.
.i 1
i i
l l
i j
i i
1 l
i 1
i F~
'g V
p-1 i
i i
FIGURE 12 PHOTOGRAPHS OF THE MODEL IN THE WIND TUNNEL NRC 1:96.67 TESTS 1
6
I
~ _ _
f _ ~r* ' -
t w
3-l
~.
l
' q,.;
d
?:
4 el
,sg I,I f
f T
~,,
e l
~
../-
l
[,
pr.
0 i
1 l
1 i
l FIGURE 13 VIEW OF THE MODEL IN THE WIND TUNNEL SHOWING THE 4
UPSTREAM TERRAIN MODEL USED NRC 1:96.67 TESTS
.