ML20059M789
| ML20059M789 | |
| Person / Time | |
|---|---|
| Site: | 05200003 |
| Issue date: | 11/12/1993 |
| From: | Pernsteiner A AFFILIATION NOT ASSIGNED |
| To: | |
| Shared Package | |
| ML20059M769 | List: |
| References | |
| NUDOCS 9311190321 | |
| Download: ML20059M789 (90) | |
Text
_
CONDENSATION IN THE PRESENCE OF NONCONDENSABLE GAS:
EFFECT OF HELIUM CONCENTRATION by Arthur P. Pernsteiner A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science (Nuclear Engineering and Engineering Physics) 1 at the University of Wisconsin - Madison -
1993 9311190321 931112
?P PDR ADOCK 05200003 d
'A~
PDR _ ;)--
i Abstraet CONDENSATION IN THE PRESENCE OF NONCONDENSABLE GAS:
EFFECT OF HELIUM CONCENTRATION Arthur P. Pernsteinct Under the supervision of Professor Michael L. Corradini An experimental investigation to determine the effects of helium and air on condensa-4 tion heat transfer rates was conducted. The test facility consisted of a rectangular channel, with one wall being a cooled aluminum plate. A series of forced flow tests were conducted with the plate in a horizontal, downward facing orientation. The mass fraction of noncon -
densables for these tests was varied from 0.65 to 0.78, and the mixture velocity was varied from 1.0 to 2.1 m/s. The helium concentration ranged between 0 and 39 molar percent of the total noncondensable content. In order to characterize natural convection effects, a second series of tests was conducted with the condensing surface in the vertical position and
' stagnant' flow conditions.
In this series of expedments the mass fraction of noncondensables was varied from 0.33 to 0.92, and the helium concentration ranged between 0 and 30 percent of the total noncondensable content. In general, heat transfer rates remained constant within the uncertainty margins when helium was substituted for air on a molar basis. However, results for the very high helium concentration run (~40 molar percent), m the forced flow experiments, indicated an enhancement of the heat transfer V
a F ' -
+W,i dM-,4N
-4 3
i d
h J-g ad--_i s-na R
--,i ii rates. The previously reported inhibiting effect of helium as a simulant for hydrogen by.
other investigators was not observed.
1 Approved:
Date Professor Michael L. Corradini l
Nuclear Engineering and Engineering Physics
?
t I
4
=
t t
7
?
ii I
L w
=
wg r
--~v
iii i
i Acknowledgement I would like to express my deepest gratitude to the time and effort spent by my advisor, Professor Michael L. Corradini on helping me achieve this goal. His suppon and guidance over the four years I have spent in this department, as a graduate student and as an under-graduate hourly, has been invaluable. He always finds time, and room in his office, for the hoards of students that hover around his door.
Thanks is also extended to the rest of the department. I believe we are blessed with the best group of professors and supporting staff on campus. I can't say it was easy to get up in February and go to a 8:00 lecture, but at least I could count on having people who knew who I was, and cared how I performed running the classroom, and the office.
Here's to the many hours spent in Rm.131. While the work I did there may be the most important of the office activities, it will never stand out like those intense games of' office hockey' or the heated political debates. I know that many of the friendships I have made here, will prove to last a lifetime.
Finally, I would like to give a special thanks to my parents. Their special kind of support over the years will never be forgotten. I share any goals I achieve, or success I may have, with them.
l
iv Nomenclature C,
Heat Capacity D,
Mass Diffusion Coefficient 3
f Molar Concentration h
Heat Transfer Coefficient k
Thermal Conductivity M
Molecular Weight p
Pressure q"
Heat Flux T,,,,
Coolant Temperature T,
Mixture Temperature g
T Surface Temperature m
V Flow Rate x
Distance Greek Symbols 9
Inclination Angle p
Density co Molar Ratio
Table of Contents Abstract i
ill Acknowledgements Nomenclature lv Thesis Outline v
vill List of Figures List of Tables x
1 Introduction and Background 1
1.1 Air / Steam Experiments 2
1.2 Air / Helium / Steam Experiments......
4 2 Literature Review 6
2.1 Previously Investigated Studies 6
2.2 More Recent Work 8
2.2.1 Separate Effects Studies....
8 2.2.2 Integral Experiments.........
11 2.3 Justification for Current Study...............
12 3 Force Flow Air / Steam and Air / Helium / Steam Experiments 14 3.1 Forced Flow Test Sectiun 14
vi 3.2 Experimental Procedure 19 3.3 Results and Discussion
... 21 4 Natural Convection Air / Steam and Air / Helium / Steam Experiments 30 4.1 Natural Convection Test Section 30 4.1.1 Air /SteamTests.
............ 30 4.1.2 Air / Helium / Steam Tests 33 4.2 Experimental Procedure 33 4.2.1 Air / Steam Tests 33 4.2.2 Air / Helium / Steam Tests 35 4.3 Results and Discussion 37 4.3.1 Air / Steam Tests
. 37 4.3.1.1 Heat Transfer Coefficient Measurements....... 37 4.3.1.2 Visuahzation Tests
... 41 4.3.2 Air / Helium / Steam Heat Transfer Coefficient Measurements. 46 5 Conclusions and Recommendations 53 5.1 Conclusicns..........................................
53 5.2 Recommendations................................... 5 4 Bibliography 56 A Error Analysis Calculations 60 Al Coolant Energy Balance Heat Flux Measurements...............
60 A2 Heat Flux Meter Heat Flux Measurements...................... 62 1
A3 Heat Transfer Coefficient Measurements..............
...... 63 1
i 1
i
vii-A4 Uncertainty Calculation forHelium Concentration Measurements
... 64 B Summary of Alr/ Steam Natural Convection Results 67 C Air /IIelium/ Steam Natural Convection Results 69 Cl Test Conditions and Average Heat Transfer Coefficients......... 69 C2 Heat Transfer Coefficient as a F:metion of Plate Position...
.. 71 D Attempts at Measuring Velocity P:ofiles 75 D1 MKS Pressure Transducer
... 75 D2 Hot Film Probe 77 D3 Visualization with Strobelight
. 78 L
~.
e- -. - _.,, _ _
viii List of Figures 1.1 Passive Safety Features of Westinghouse AP-600 Design..........
3 3.1 Schematic of Forced Flow Experimental Facility 16 3.2 Sideview oftest Section 16 3.3 Endview of Test Section.
17 3.4 Steam Diffuser and End Section..
17 3.5 Repeatability of Forced Flow Air / Helium / Steam Experiment...
. 22 3.6 Degradation of Heat Transfer Coefficients Due to Helium Injection..
29 4.1 Schematic of Natural Convection Experimental Facility 31 4.2 Steam Diffuser and End Section...
32 4.3 Schematic of Natural Convection Facility Modified for Helium Tests.
34 4.4 Repeatability Using Heat Flux Meter Method.........
39 4.5 Repeatability Using Coolant Energy Balance Method.........
39 4.6 Location of Heat Flux Meters (HFM) Along Test Plate.............
43 4.7 Flow Pattern Observed During Visualization Tests................
44 4.8 Repeatability Using Heat Flux Meter Method....................
48 4.9 Repeatability Using Coolant Energy Balance Method............... 48
1 ix 4.10 Air / Helium / Steam Heat Transfer Coefficients 50 4.11 Air / Helium / Steam Feat Transfer Coefficients..................
50 4.12 Air / Helium / Steam Hest Transfer Coefficients...................
51 4.13 Air / Helium / Steam Heat t ransfer Coefiicients...................
51 4.14 Comparison of Air / Steam Only and 30% Test Results
.... - 52 i
i j
l 1
l l
i, l
l l
4 s
X Y
- f* rri. 1 ust oi 1aoles 7
2.1 Summary of Previous Investigations 2.2 Summary of Recent Work in Condensation Heat Transfer
.9 12 2.3 Predi ed Conditions in Containment During Accident 3.1 Air / Helium / Steam Tests Performed at P
=1 bar 23 om 3.2 Pure Air Tests..
24 24 3.3 Transport Propenies of Noncondensable Gases (30 C) [18]
4.1 Steam / Air Natural Convection Results..
38 4.2 Air / Helium / Steam Natural Convection Results..
47
1 Chanter 1 r
Introduction and Background A series of experiments investigating the condensation of steam in the presence of noncondensable gases was conducted. The experimental facility simulates the likely conditions inside the containment building after a loss of coolant accident. Noncondensable gases are of primary concern due to the heat and mass barrier they present during the accident. This effect is due to the formation of a layer of noncondensable gas adjacent to the containment wall through which steam must diffuse before condensing. As a result, the heat transfer rates in this series of expenments, as in the containment, are controlled by a diffusion phenomena as well as heat conduction through the cold wall and any condensate film present on the wall.
In the event of a loss of coolant accident, large amounts of pnmary system water may flash into steam and be released into the containment atmosphere. Unless mechanisms are provided to reduce containment pressure, the containment building's structural integrity may be jeopardized. Past reactor containment systems relied primarily on active coolant
i 2
l sprays for containment cooling, requiring backup diesel generators, or the availability of i
off-site power. New advanced containment designs utilize passive cooling methods in an effort to improve system reliability and reduce the costs associated with active safety systems. One such system is employed by the Westinghouse Electric Corporation in its AP-600 nuclear power plant design.
A layout of the AP-600 and its passive safety features is provided in Fig.1.1. In a loss of primary system coolant accident, water is allowed to flow from the large reservoir at the top of the containment. The water forms a film which flows over the inner containment shell, providing evaporative cooling of the containment during the initial stages of the accident. Due to the limited inventory of water available, air inlets are provided. The natural circulation of air between the steel interior and the outer concrete shell will provide sufficient cooling capacity later in the accident scenario, when core decay heat levels are lower. Due to the high heat transfer coefficient associated with water in direct contact with t
the steel shell [1], the heat conduction through the wall and the heat transfer coefficient on the inside of the containment shell are important in determining the overall efficiency of the cooling process.
1.1 Aire team Experiments Condensation of steam in the presence of air is of pnmary importance when considering containment performance.
The goal of this series of experiments was to obtain experimental heat transfer data associated with steam in the presence of air. Tests were performed at ambient pressures with naturd convection flow near a vertical, cooled plate.
~...
3-CCNTAINM&MT C00*JeG Nf.ATED AIA OGCHAAGE 4
4,/
- AS$*.'E CONTAwutNT C000NG sv5TTM HATER STORAGE TANE g 3 gy 3 WATEM WITWIN THE AV
'l 00NTAWMENT COOUNG
@. VATER CLCOE THE 9V
\\
A.a WLET e
s
~'I
\\
f
\\ MIT T
\\
7,. :. E.Gr._., /
'y y',
11 Il
.ua eAFatt f
t l 64 4 ;
i f
N NS*!!L 00NTLNuf%T.Ess!L e
i.dl ;
!!l '
til f 6
f till!
i.
I I
+
. i t'f
~
64 'l ~
6 m
4. e.
,, - i, h1
'I
'1 D U. Il_t/
- . STEAM GENERATCM eu
..sssunizs3 i
f*
4:
bL 4 '
~' '
F. '. 2 5
= a P*fSSURIZEM CONNECTON O'I J
- 5. (AW5T CONNECT 0N i
~
- " h. -
3 lId
,7,5
- 6. CORE MAEEVP tANE X
gi t_g.p, i
a;-
2c0euuut0a 44.:i e>
= s se 6
j sp y
- s. c<.p.;
s
=
'ds%
EQ'
,g.I 1
Eh?'a.
25.3T Fig.l.1: Passive Safety Features of Westinghouse AP-600 Design i
I
=
m ei
- --w*a
.iw--
w wr+w.*
.c e
r+-W9-4
+
m-M #W
4 Results were compared with those presented by [1]. In addition, data collected in this series of experiments was used as a reference for later experiments involving helium as a second, noncondensable gas.
1.2 Air / Helium / Steam Experiments If the primary system water inventory is lost and emergency core cooling systems fail to provide sufficient cooling, the temperature of the fuel may reach levels high enough for water to chemically react with the cladding material. These oxidation reactions release hydrogen to the containment, which raises the concentration of noncondensable gas.
FurtherTnore, hydrogen, having a molecular weight only 7 percent that of air, may produce secondary condensation effects not directly related to the simple increase in noncondensable concentration a to its relative buoyancy. Although the overall hydrogen concentration is not likely to be above 13-20 molar percent (due to hydrogen detonation) it may play an important role in the condensation heat transfer rates of the cold surface.
The goal of this investigation was to obtain experimental data on the heat transfer rates with a noncondensable gas mixture of air and helium. Due to the potential danger of handling hydrogen, helium gas was used to simulate the hydrogen gas. Tests were performed at ambient pressure with a forced, parallel flow over a horizontal, dowTiward facing surface as well as natural convection flow next to a vertical, cooled plate.
Past tests [3,4] have focused primarily on air-helium tests under ' stagnant' conditions conducive to natural convection. In particular, the tests by Dehbi [4] utilized helium as a third gas in large molar concentrations (9 to 35 molar percent).
Although such
5 concentrations are useful in determining the effect of a light gas, they are not in a prototypic-'
range for severe accident conditions. The current test series focused primarily on lower heliurn concentrations (3 to 18 molar percent) than in past tests. The choice of this range is dictated by the likely conditions in an LWR contamment; i.e. below detonation limits.
Experimental results were compared qualitatively to the data by Dehbi. A new approach of quantifying the overall effect of helium is proposed and the results are discussed from this perspective.
c
?
e r
4
-. ~..
N 6
Chapter 2 Literature Review 2.1 Previous Studies A review of earlier works has already been discussed in detail by Huhtiniemi [2].
Huhtiniemi classified previous condensation heat transfer studies into two major groups:
- l. Separate Effects Experiments A simple test geometry is used to isolate a small number of selected factors for investigation. Typically, time dependence is elimmated, and steady-state conditions are achieved.
- 2. IntegraVLarge Scale Experiments An effort is made to include factors such as a realistic flow geometry or transient behavior. Generally, the complexity of the experiment complicates the analysis of the results so that contributions from separate factors cannot necessarily be isolated.
A summary of pertinent separate effects studies is provided in Table 2.1.
i 4
i
7 Parameter Bart,j [5]
Dallmeyer[6]
Debhi[4]
Gerstmann [7]
Henderson [8]
Gas air air air, He air air Vapor steam ccL,,c/I, steam freon-ll3 steam T.,, [*C]
51.3-88.2 95 sat NA NA m,/m,,,
0.47-0.92 0.02-0.16 0.25-0.9 trace 0.1-0.83 m
v., [m/s]
2.1-6.9 1-13
-0
~0 NA p[]
180 90 90 0-90 180 P [MPa) 0.1 sat 0.15-0.45 0.1 NA A T["C]
26.3-63.2 55-85 10-65 4.3-39.4 NA Geom plate plate tube plate tube L/D 610 410 3500/38 457.2 1220/29 Cho [3]
Kroger[9]
Kutsuna [10]
Robinson [11]
Slegers [12] l Spencer [13]
air Ar,He air air air N,,Co,,He steam potassium steam steam steam freon-113 sat 598-768 85-90 sat 26.7-65.6 sat 0-1.4E-5 NA 0.42-0.55 0.16-0.87 0-0.01 0-0.03
~0
~0 4-5.3
<2 m/s, N/A
-0
-0 0
0 180 90 90 90 0.31-1.24 sat 0.1 0.27-6.2 0.004-0.03 NA 35-100 2.3-733 5.0-15 4.0-10 1.4 20.8 N'A disk disk plate disk plate tube M.
137 101.6 800 46 127 NA/NA Table 2.1: Summary of Previous Investigations a
n--
a
8 A discussion on the focus of the separate effects studies, along with an overview of intergral/large scale experiments can be found in Ref. [2].
2.2 More Recent Work 2.2.1 Separate Effects Studies Several groups have contributed to the condensation heat transfer field since the work by Huhtiniemi in 1991. A summary of the recent separate effects studies is provided in Table 2.2. In this chapter, the previously published experiments are summanzed.
Lu and Suryanarayana [14] investigated vapor flow inside a horizontal rectangular duct.
The vapors used in the study were R-113 and its proposed replacement refrigerant FC-72" (3M fluid). The heat transfer coefficient was found to increase with increasing inlet vapor velocity, and an enhancement of the heat transfer coefficient was observed upon the appearance of interfacial waves. These effects have been observed before, and were verified here for the CFC R-113 and its replacement fluid (FC-72
). The results were correlated by two separate equations for the wave free regime, and another for the wavy regime.
Fox et. al [15] studied steam-helium and steam-air mixtures inside a reflux condenser. It was found that transport phenomena were greatly affected by the stability of the component
9 4
b Parameter Lu [14]
Fox [15 Siddique[16]
Kang /Kim [17]
i Gas none air He air air l
Vapor R-113/FC-72 steam steam steam l
T.,[ C]
50/60 sat.
100-140 82-100 0.75-0.95 0.10-0.35 0-1.0 m,/m,,,
v., [m/s]
0.3-4.4
-0.2 N/A 3-j 9[a]
180 90 90 184.1 l
i P [MPa]
sat 0.1 sat 0.1 Geom plate tube tube plate
'f l
UD 40 8.1 55 15.2 Table 2.2: Summary of Recent Work in Condensation Heat Transfer i
5 5
i
10 combinations used in the condenser. In general, stable flow pattems were observed when helium gas, with a molecular weight less than that of the vapor, was used in the condenser, and unstable flow patterns developed for the cases when a noncondensable gas heavier than l
the vapor (air) was implemented. Stable conditions were observed for both noncondensable gas loadings at high vapor mass fractions (co=0.95). It was found that certain unstable conditions exist which result in oscillatory recirculation regions which exhibit small temporal fluctuations. The results indicate that simple models which assume there is a stable gas / temperature front are not valid when using noncondensables with a molecular weight greater than that of the condensing vapor.
Siddique et. al [16] measured the local condensation heat transfer coefficient of steam, in the presence of air, in a vertical tube, with a downward flow. The experiment was developed to model condensation heat transfer inside the isolation condenser, a component of the General Electric SBWR passive cooling system. The inlet air mass fraction ranged from 0.10 to 0.35, with mixture inlet temperatures of 100,120, and.140*C. The ' local l
Nusselt number increased with the mixture Reynolds number and decreased with increased noncondensable mass fraction. A model was developed to predict local heat transfer coefficients on the inside of tubes, for the range of conditionc studied in the test series.
Kang and Kim 117] investigated the effect of noncondensable gas and a wavy water film on condensation heat transfer. A water film was injected into a 1.52 m long rectangular channel, at steady state thermal conditions, to produce a wavy film condition. Data was
M 11 collected for various air mass fractions (0-0.78), mixture velocities (17 m/s), and film flow rates. Even small amounts of noncondensable gas was seen to greatly affect condensation heat transfer rates. Also, the waviness of the condensate film increases the heat transfer up to several tens of percents. Although both of the previous observations were knowTt previously, this study extends the range of quantitative data to the conditions stated above.
2.2.2 Integral Experiments In addition to the heat transfer data, integral and large experiments provide valuable background information about physical conditions such as gas concentrations, temperatures, prevailing flow fields, and system pressures in the contamment during an accident.
The Westinghouse Electric Corporation has carried out several large scale tests involving an extemally cooled containment structure similar to that proposed for the AP-600. The 1/8th scale facility allows for the measurement oflocal heat fluxes and total heat transfer rates. The effect of steam injector location and the effectiveness of external surface cooling can be tested. Unfortunately large scale test results are often proprietary, as in the case of the Westinghouse study, and are not available to the scientific community.
A second large scale integral test is being performed at the Paul Scherrer Institute, which is very similar to the work being done by Westinghouse. The 1/10 scale SI3WR, PANDA, experiments will provide General Electric with performance characteristics for their new simplified boiling water reactor isolation condenser in a scaled drywell.
k r
12 i
2.3 Justification for Current Study The major motivation behind the current series of experiments is the need for representative-heat transfer coefficients under conditions which are likely to occur in a range of accident scenarios for the plant design base (Table 2.3). An estimation of the possible prevailing conditions is provided though computer analysis and large scale integral testing. Very few separate effects experiments have been carried out which provide quantitative data under the conditions listed in Table 2.3.
m,jm, Pres. [MPa]
Vel. [m/s]
Length Scale [m] Noncondensables
~0-0.95 0.1-0.4 0-3 1-10 Air, Hydrogen Table 2.3: Predicted Conditions in Containment During Accident Furthermore, the concentration oflight gas (hydrogen) is expected to be in the 0-18 molar percent (of the total mixture) range, i.e., below detonation concentrations due to hydrogen burns. A final qualifying feature, for comparison to this study, is the utilization of a test geometry which simulates external flow. The implementation of a flat plate of sufficient width to eliminate wall effects, or a large tube geometry with negligible curvature effects is 4
required.
The study which most nearly fulfills the previous criteria was that carried out by Dehbi et.al [4]. In these experiments, condensation occurred on a cylindrical, cooled tube, within a vertical,3.5 m outer tube. Steam in the presence of known concentrations of air and helium was admitted into the test section, and heat transfer data was collected at an elevated
i 13 pressure (~3 atm). The helium mass fraction ranged between 1.7 and 8.3%. Decreasing heat transfer coefficient values were measured for increasing helium mass fractions.
Stratification pattems were observed, with high helium concentrations migrating to the top of the vessel. These experiments, although useful in comparison to data in the current test series, was performed at relatively high helium concentrations (corresponding to 9-36 molar I
percent). In addition the tube type geometry is intended to model conditions relative to an i
isolation condenser, and is not intended to simulate extemal flow conditions within a large
-P building. Also, the experiments by Dehbi were performed at higher ambient pressures (1-5 bar) and thus higher temperatures.
The current study models airihelium/ steam condensation under the conditions presented above. Helium concentrations ranged between 4.0 and 35.9 molar % of the total mixture for P
the forced convection experiments and between 1.7 and 20.1 molar % for natural convection experiments. Forced convection flow rates ranged between 1.0 and 2.1 m/s. The flow 1
geometry was in the developing region, which simulated the large flow areas associated with the reactor containment.
l 1
i 1
1 7
w
14 Chapter 3 Forced Flow Air / Helium / Steam Experiments The experimental facility used in the forced convection tests was a modification of the facility described in Chapter 4 of [2]. The forced convection test modifications consisted only of a new end section capable of mixing the incoming mixture of steam and noncon-densable gases.
3.1 Forced Flow Test Section i
The focus of this series of experiments was to measure the condensation heat transfer coefficient with varying steam / air! helium mixture compositions and flow velocities. The experimental apparatus was designed to model a section of the containment upper dome that j
i is exposed to the condensing mixture. A channel configuration instead of a spherical dome i
structure was adopted because it allowed better control of the parallel flow over the condensing surface. The parallel flow geometry was considered to be more appropriate for containment analysis purposes than the stagnant conditions used in the previous
=-
t 15 investigations [3,4] because the natural convection in the containment dome'is likely to
- generate a flow field that is primarily parallel to the walls in the upper dome region. In our opinion, the scaling of ' natural-convection-driven' flows is quite difficult, and thus, these flows are difficult to interpret relative to the size of the containment. The turbulent flow characteristics of this test series may be more representative oflarge scale mixtures because they reduce the stratification of the gases reported by past investigators [4].
The experimental facility is illustrated in Figure 3.1. The test section consists of 190.5 cm long plates that are sandwiched to form a 15.2 cm square duct. The transparent, polycarbonate-plates were used to allow visual access into the test section. The test section 4
was designed to withstand test temperatures of 140*C and an absolute pressure of 2 bars.
The first 83.8 cm portion of the test section acts as an entrance length. This length was calculated to be long enough such that the majority of free stream turbulence is climinated.
The damping of the flow was necessary to simulate the developing flow pattern that would be present in the actual containment dome. Measurements of the turbulence level at the exit of the entrance length were at or below ~ 10% [5] for noncondensable gas flow. Because the facility represents a small portion of a much larger containment dome, the total length of the test section was scaled so the boundary layers of the opposite walls did not interact.
Condensation occurs in the section immediately following the entrance length. The portion of the section which simulates the containment wall is a 3.8 cm thick aluminum plate mounted as the duct ceiling.
16 Or QT r
r kb w housy
.a i
mutal warcu i q
]
b 3179(T }
t :t i i
- t
- i P d.
r
!I I
I II I
A tts, scer o.
acworrs M
sin.cn l
u:v m s w FE2VafER Pur m
w M
.co. msa FCEDvatta pusp ON py g Fig.3.1: Schematic of Forced Flow Experimental Facility
)
HEAT FLUX METER STATICNS (HFM-1 etcJ 4
- ,- 0.8500 I21.593 i
i
-L ij 6
i e
3 I
I i
1 4
e 9
e o
3.0000 (76 2 03 -
64000 t152.403 -
Fig.3.2: Sideview of Test Section
17 reocmc
.AsCT133 j
a.
I l
l I
- no,i f _ _ f
_ l _
]-.-
--11m (?7.79l 4 = na.n l
l
__ l_ _ l _ _ l _ :.
l i
I t
ll l
I
- eme. ruan t
am. ruma -
l Fig.3.3: Endview of Test Section AIR 1_
V ooouooh (
STEAM oooooc nonceof_
s Fig.3.4: Steam Diffuser and End Section
?
i 18 The aluminum plate is 106.6 cm long, with an effective heat transfer width (width of plate exposed to steam /airihelium mixture) of 15.2 cm. The condensing surface of the aluminum.
plate was sandblasted and coated with Carbon Zinc II* paint. This paint simulates the coating to be used on the containment shell of advanced reactors. See Figures 3.2 and 3.3.
for test section diagrams.
Cooling of the aluminum test wall is provided by seven identical cooling plates mounted on the backside of the condensing area. Each of the seven cooling plates contains an independent coolant conduit, machined in a spiral pattern to provide uniform cooling.
The plates are mounted symmetrically above the heat flux measurement stations as illustrated in Figure 3.2. The use of separate cooling plates enabled us to measure local heat i
fluxes more accurately.
The steam / air / helium mixture enters through a mixing tee and passes through the diffuser located at the inlet of the test section (Fig.3.4). Steam which is not condensed in the test section passes directly into a secondary condenser which consists of a cold water 3
spray and heat exchanger. Condensed steam is collected in a r:servoir and pumped back to the boiler to be recycled.
The heat energy removed though the cooling plate is measured using two separate techniques: 1) Heat Flux Meters (HFM) and 2) Coolant Energy Balance (CEB). The heat flux meters consist of an array of four thermocouples embedded in a Teflon strip. The I
Teflon strips are located in grooves which were machined into the test wall.
The thermocouples are placed at approximately equal spacing, with positions measured to within 0.05 mm. The distance between the outermost thermocouplejunctions was approximately -
-m' r
m r
e
- -...~', - + -,..
19 20.5 mm. A linear least squares fit was used to obtain the temperature gradient inside the wall. The coolant energy balance method requires a measurement of the coolant flow rate (V) and temperature rise (ATcoa) between the inlet and outlet for each coolant plate.
The standard error of the heat flux measurement was typically less than 5% for the heat flux meter method and 12% for the coolant energy balance method [ Appendix A]. The temperature rise in a cooling loop is inversely dependent on the flow rate. Therefore, a criteria between muumum flow rates and temperature rise across the plate can be calculated so the relative error for each measurement is minimized (Appendix A).
The local condensation heat transfer coefficients are defined as:
h ' = r q"
- r,,,
3.2 Experimental Procedure The experimental procedure for the forced convection experiments is similar to that described by Huhtiniemi [2]. However, the introduction of the noncondensable gas helium required modifications to the procedure which is outlined in this section.
The test begins by setting the air mass flow rate using a static pressure gauge and a rotameter. A second rotameter was used to deliver the desired helium flow rate. As Figure 3.4 indicates, the air and helium flows were delivered into the test section together after passing through a mixing tee and associated piping to the test section. Our experience indicated that this premixing arrangement resulted in a uniform mixture of the
i T
20 noncondensable gases before the mixture is combined with steam.
However,. no measurements were made of this during the tests. After the air and helium mass flow rates e
were set, the steam mass flow rate was adjusted so a saturated mixture condition existed in the test section. In past experiments [2], stratification of steam from the noncondensable gas was observed unless special precautions were taken. These involved the design of a steam diffuser to mix gas and vapor uniformly. The mixture conditions were checked with a temperature measurement, to verify uniform condensible/ noncondensable mixture.
i 1
The temperature and flow rate of the coolant through each plate was adjusted so a uniform surface temperature of the plate was achieved. This ensured the heat flow in the condensing plate was one-dimensional. After the coolant flow settings had been set to give the desired surface temperature, the experiment was run continuously for 0.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> to ensure steady state conditions had been reached.
After the steady state conditions had been achieved, the data acquisition was started for an air test. Due to limitations of the data acquisition system, only 16 chaimels of data could be acquired simultaneously. The data acquisition procedure ' as composed of a 2 minute w
acquisition period followed by a five minute interval over which the data was transferred '
from the computer memory to a hard disk. This procedure was repeated for each of the
{
three thermocouple cards. Therefore, the total data acquisition period was 16 minutes.
After the test, the measured gas inlet temperatures with observed gas line pressures were j
used to calculate an updated estimate of velocity in the test section. Following the air test, the flow rates of air and helium were adjusted appropriately for a helium test and the facility
21 was run for five minutes to allow the new test conditions to reach steady state. The data acquisition procedure for the helium test was identical to the air test.
A few tests were repeated in order to make sure the test results were con.sistent and the test apparatus was operating in steady state. R.esults of such comparison are shown in Figure 3.5. The results are very consistent and within the uncertainty margins (<!2%).
t 3.3 Results and Discussion The air-helium experiments performed in this test series are summarized in Table 3.1. The corresponding air tests are listed in Table 3.2. In the first twelve air-helium and air tests, the molar fraction of noncondensable gas to steam was maintained at 0.69 (70*C at 1 atmosphere). In the air-helium tests, this means a fraction of the air molecules originally present at a given temperature were replaced by an equal number of helium molecules.
Comparing the air / helium / steam and air / steam results, it appears the effect of helium on the heat transfer coefficient is negligible if the molar fraction of noncondensable gas in the i
mixture is held constant. This was initially somewhat surprising because the diffusion coefficient for a steam / helium mixture is significantly higher than that for the steam / air mixture (Table 3.3). The higher diffusion coefficient would enhance the mass transport and, consequently, should lead to higher condensation heat transfer rates.
With the downward facing cold surface, the noncondensable gas would tend to stratify at this location. For a two component mixture (helium and steam compared to air and steam) this would seem to favor more vapor diffusion. However, the partial molar density of helium is much smaller than steam by the ratio of their molecular weights. This would have the
22 r
Repeated Tests @ 70 C,12% He 200 6
2 180 Tested -
Tes:W7
+
p Testn8 o y
3 160 E
.2 3
'1 uy 140 s
i e
l,'
t o
120 n
o I
100 0
0.2 0.4 0.6 0.8 1
Distance Downstream [m]
Fig.3.5: Repeatability of Forced Flow Air / Helium / Steam Tests
t 23 Test #
T
[*C]
He conc. [%)
vm [m/s]
HFM [W/m K]
CEB [W/m; K]
2 1
68.6 26.0 1.1 123 133 s
2 69.9 4.0 1.0 118 129 t
l 3
70.5 8.0 1.0 125 138 4
70.2 12.0 1.0 130 136 5
70.4 16.0 1.0 125 133 6
70.9 8.0 1.0 122 134 7
70.0 12.0 1.0 115 127 8
69.5 12.0 1.0 120 133 9
70.6 38.9 1.5 174 187 10 70.2 12.0 1.5 148 159 11 70.7 8.0 1.5 145 159 12 70.4 4.0 1.5 140 151 13 79.5 3.9 2.0 246 259 14 79.4 3.9 2.0 249 262 15 79.0 7.5 2.0 233 251 16 78.8 10.9 2.1 247 261 l
Table 3.1: Helium / Air / Steam Tests Performed at Pm=1 bar 7
l
24 2
Test #
T_[*C]
HFM[W/m K]
CEB[W/m K]
1 70.1 120 124 2
70.2 120 134 3
69.8 113 123 4
70.0 117 124 5
70.3 114 118 4
6 69.9 120 128 7
69.3 112 120 8
69.5 120 133 9
NA NA NA 10 70.2 144 155 11 70.2 143 152 j
12 69.9 133 145 13 79.8 249 263 14 NA NA NA 15 80.0 234 245 16 79.9 242-246 Table 3.2: Pure Air Tests 2
Gas Component Thermal Conductivity [10-2W/mK] Diffusion Coef. in steam [10"m f3)
Air 2.6 2.7 l
He 15.1 22.2 l
H,,
18.4 13.4 i
Table 3.3: Transport Properties of Noncondensable Gases ( 30 *C ) (Ref.5 He)
l 1
'l H
25 counterbalancing effect of causing a larger helium boundary layer to develop at the cold j
downward facing surface as steam condenses, thereby distilling a " deeper" light gas boundary layer. Thus the two effects may quantitatively cancel each other out.
Unfortunately,it is difficult to assess the importance of these effects separately, because the magnitude of the difYusion coefficient for the air / helium / steam mixture is not knovm and local species sampling was not possible without disturbing the flow. However, an estimate of the mixture diffusivity can be obtained using a method recommended by Bird et. al [18].
The mass diffusivity Da for a binary mixture is approximated by the formula:
r
= a( JTa a )b Eqn. 3.1 (papa)m(Ta a)m1(Q+y)^
r r
For H O with nonpolar gas:
2 a = 2.745x10d b = 2.334 Equation 3.1 can be solved directly for the air / steam mixture, but for the case of air and helium in the presence o? steam, an estimate was necessaiy for the critical properties of the air / helium combination. The approximations used were simple molar averages for the gas pair [18].
p', = E x,p,,
T, = Z x4T,,
Eqn. 3.2, 3.3 b
' - - +
e
26 The heat transfer to the cooled plate in the presence of noncondensable gases is governed by the diffusion of steam to the plate, and the thermal resistance of the mixture to conductive heat transfer. In this case with noncondensable present, the heat transfer due to conduction can be ignored, and the heat transfer rate is linearly dependent on the mass diffusivity D l
a.
q" = Constant x Das Eqn. 3.4 Calculations using the diffusion coefficient estimate were performed for various helium concentrations. Results from the calculations indicated that the magnitude of the increase in heat transfer rates is about the same as the uncertainty margin with small helium i
concentrations (4 to 26 percent). Therefore, a test was performed with a relatively large helium concentration (test #9), so the magnitude of the increase due to helium was larger than the uncertainty margin. The measured increase in the heat transfer coefficient of ~50 2
2
[W/m K] compared well with the predicted increase of ~46 W/m K. This would suggest that the effect of helium is obscured by the measurement uncertainties at low. helium concentrations and the counterbalancing mechanism could be more important with small flow velocities (<1 m/s) or with very large light gas molar fractions.
It should be noted, that the preceding calculations are not presented as quantitative i
evidence supporting the effect of helium. Rather, they are a crude first step in an effort to l
characterize the measured increase in the heat transfer coefficient. Bird et. al (18] caution, l
l that the critical pressure and temperature approximation is not very accurate, particularly if the critical properties of the components differ greatly, as in the case of air and helium.
Time constraints prohibited further analytical study on this matter.
H
i t
27 Several approaches to quantify the overall effect of helium were considered. Dehbi [4]
kept the helium mass fraction constant and varied the air fraction. The degradation of heat transfer rates due to hemua was then deduced by comparing these results with pure air and equal noncondensable gas mass fraction. However, this approach accentuated the effect of helium because replacing equal weight of air with helium results in the higher molar l
concentration of the noncondensable gas; i.e., so high as to be unphysical for the reactor contamment situation. The inhibiting nature of helium on heat transfer rates that was reported by Dehbi was not evident even with the higher helium concentrations in the current test series. This would suggest that helium suppressed natural convection in Dehbi's i
apparatus, which likely had accentuated the stagnant flow conditions and the light gas effect. No bulk velocities were measured in Dehbi's tests.
l Although we have compared our results to those of Dehbi, it is partly due to the scarcity
- of data on this subject, and therefore it is important to note that there are differences between the two experimental conditions. The experiments by Dehbi were performed at -
higher ambient pressures (1-5 bars) and thus higher temperatures under natural convection conditions. Perhaps the most important differences between the two studies are the cold -
wall orientation and the range of helium concentrations: vertical and larger amounts of helium [4] versus horizontal and downward facing with lower helium amounts for our study The natural convection experiments presented in Chapter 4 are rnore closely related to those conducted by Dehbi, the discussion of which will be provided later.
Considering the fact that the air mass in the containment is constant, one can imagine helium (hydrogen) molecules are substituted for steam molecules in the flow entering the
-h i
i
28 containment due to metal oxidation. This is a much more physically realistic approximation of what may occur in a degraded core accident. In the last four tests of the series (test #
13-16), the air mass flow was held constant and the heat transfer rates were measured with different helium injection rates. As expected, the saturation temperature decreased with -
increased helium injection rate at a constant pressure, because of the increase in the noncondensable gas fraction. However, the heat transfer rates remained constant within the uncertainty margin, confirming the modest effect oflight gas addition to this mixture.
A similar comparison for the low velocity tests can be obtained by using data by Huhtiniemi (2) who measured the heat transfer coefficients as a function of mass ratio for air and steam mixtures. A mass ratio of air to steam was calculated for the case in which the helium molecules are interchanged with steam molecules. This lower mass ratio of gas to steam was then used to find the corresp anding increase in the heat transfer coefficients. The results are plotted in Figure 3.6, which shows the heat transfer coefficients with the helium -
and in the case where helium was replaced by steam. The degradation by helium is clearly evident only with the higher helium concentration. This indicates that the degradation of L the heat transfer rates with small amounts of helium is counte.rbalanced with higher steam diffusion fluxes. When the helium concentration is higher than 10 percent, the added noncondensable gas should begin to noticeably decrease the heat transfer rates.
29 l
l 200 190 *
[
180 -
h
< Measureo MTC e
g 170a e HTC H2O sucstattiang He i 3
1604
'Q t
==*
150*
b n
j 140 -
i '
i l
130 -
.6 9
I 120 -
110
- 100 0
10 20 30 He Concentracon (% of total nonconcana4 Die gas)
Fig.3.6: Degradation of Heat Transfer Coefficients Due to Helium Injection 1
l
30 Chapter 4 Naturai Convection Air / Steam and Air / Helium / Steam Experiments 4.1 Natural Convection Test Section 4.1.1 Air / Steam Tests The experimental facility used in the forced convection experiments was modified for use in the natural convection series. The major modification was the positioning of the test wall in a vertical position at the top of the test section (Fig. 4.1). Positioning the test plate vertically allows for the development of gravity induced flow patterns. The ends of the test -
section were sealed, except for the steam inlet at the bottom of the test section and the condensate dram. Steam enters into the bottom end plug and through a diffuser (Fig. 4.2).
The diffuser was designed to mitigate forced convection effects and to aid in the mixing of:
l the noncondensables in the steam / air / helium natural convection tests. Condensed steam is collected inside the bottom of the test section. A plastic float and stopper valve release excess condensate before it reaches the level of the steam diffuser. The instrumentation and we u
w ra se
a a.
,. ~
31
+
a i
)
.L CHILLf 3 watts
..E P MICIM 3WF(T r agV
"!ELAfsf kidssFfLB TC37 SECTDI P
VATOR 8%MCP TM T
g l
._ ii= _
e-w a
R w
w LOV Persm wres partsuet FEDVATER W rggvarga pgep d6 RCC3tCLLAfDs LDEC Fig.4.1: Schematic of Natural Convection Expenmental Facility i
J i
I 1
ya.-,
g
32 l'8 0oC 0oC 0oC l
l Phf h
,\\du 6
l i,,
ecarmenssat.ts }
b situe Fig.4.2: Steam Diffuser and End Section
~
w 33 support systems associated with the facility remained identical to those described for the forced flow experiments.
4.1.2 Air / Helium / Steam Tests The experimental facility used in the natural convection air / steam tests was modified only sli;;htly for the air / helium / steam tests. Helium was supplied from a high pressure tank, and the building air compessor provided the second noncondensable gas (Fig.4.3). The air and helium supply lines passed through separate flow meters and into a mixing tee. The mixed gas line was attached to the steam inlet line, allowing for the further mixing of the noncondensables as they passed through the steam diffuser.
4.2 Experimental Procedure 4.2.1 Steam / Air Experiments In the steam / air natural convection tests the experiment began by introducing steam into the test section. The temperature was brought to the desired level by adjusting the steam mass i
flow rate. The temperature and flow rate of the coolant were adjusted so a uniform surface -
temperature of the test plate was achieved. The experiment was. allowed to run for approximately 15 minutes in order to ensure steady state conditions. The time spent waiting for steady state conditions was reduced in this case, because a true steady state system was difficult to achieve, requiring an exact balance between steam mass flow rate and condensation on the test plate. The situation actually encountered was a quasi-stable state,
34 r foamtafwE T F P Put33Urt TC37 stCf!38 C
r rtev mura
=
o f-b e,. --
- f --
A U._
l P
~
l vafta marry tw T
,-r-u
~
m ortem
@- W tev missc
,,,3. m s u e rtggvatta Pupe..
rttavafER fue at UslltA.Af!38 LIE Fig.4.3: Schematic of Natural Convection Facility Modified for Helium Tests J
1 4
9
i 35 where the mixture temperature would not drift significantly over the 15 minute period (no more than 0.5'C). The time of 15 minutes was chosen because it is approximately the time required to collect all necessary data at a specific setting.
After the stable conditions had been achieved the data acquisition was started. The collection of data was identical to that described in the forced convection experiments.
When all data was collected at a given temperature, the steam mass flow rate was increased, and the mixture temperature was stabilized at the next point of interest. Data was collected between 50"C and 90*C with specific points at 50,60,70,75,80,85, and 90*C, 4.2.2 Air / Helium / Steam Experiments Running an Air / Helium / Steam natural convection test began by charging the test section with an air and helium mixture at the desired concentrations. The two gases passed through separate rotameters and into a mixing tee. The mixed gases then traveled through the steam diffuser and into the test section. The flow of air and helium was maintained for a length of time required to displace the test section volume a minimum of ten times.
There was some concem that after the air / helium flow was tumed off and steam was introduced to the system, a " distilling" process would take place, in which helium rich gas would leak out of the test section and be replaced by air. However,it was noted that the test facility was capable of maintaining modest positive pressures (1-3 psig) over a short period of time. Because there is no driving force behind the hypothetical " distilling" process, the judgement was made that the mixed gases would not unmix, and the helium / air mass ratio i
would remain constant throughout the test.
i I
u 36 After a sufficient volume of the helium / air mixture had been fed into the system, the noncondensable gas flows were turned off and the steam line was opened. Care was taken to increase the test section temperature slowly. The reason for this is the fundamental relationship between mixture temperature and the ratio of condensable to noncondensable gases. The test volume stayed at nearly atmospheric pressure due to small leaks. Increasing :
the temperature above the point of interest would force a ponion of the helium rich gas i
inside the test section through the small leaks. Dropping the temperature back to the original point ofinterest would pull air back in, resulting in a decreased concentration of helium.
The steam flow rate was increased until the desired temperature was reached. As in the air / steam experiments the system was allowed to achieve a semi-stable state. An initial set of data was taken to measure the cooled plate temperature and to make any fine adjustments on the steam flow. If the cooled plate temperature was not 30 C, the bath temperature -
and/or coolant flow rate was adjusted. The system was allowed to restabilize, and the process was repeated until system conditions were within acceptable limits (approximately 0.5'C for both the plate temperature and the mixture temperature). After a successful test,-
the steam flow rate was slowly increased until the mixture temperature reached the next predefined level and the process was repeated. Typically, several sets of data were collected
~
at each temperature step.
r I
i
37 4.3 Results and Discussion 4.3.1 Air / Steam Tests 4.3.1.1 Heat Transfer Coefficient Measurements Average heat transfer coefficient results for the air / steam natural convection test series are given in Table 4.1. Mixture temperatures ranged between 50*C and 90"C with data being taken at 50,60,70,75,80,85, and 90"C. Although more experiments were run, only those which correlate well with the desired experimental conditions are given here. A complete list of results is given in Appendix C.
The experimental error in the heat transfer measurement was determined to be approximately 10 percent (Appendix A). Tests were repeated in order to check the consistency of the heat transfer results. Figures 4.4 and 4.5 are graphic illustrations of the heat transfer coefficient along the length of the condensing plate at 60*C (Test #3 and Test
- 4) and 85*C (Test #10 and Test #11). Figure 4.4 contains data collected using the heat flux meter method, while Figure 4.5 contains data collected using the coolant energy balance method. The Coolant Energy Balance (CEB) data is displayed as a horizontal line, spanning the distance over which the particular coolan! channel was positioned. The Heat Flux Meter (HFM) data is displayed as a single point, located at the center of the CEB line. Presenting the data in this format represents the data in its true form: the CEB data an average over a length of the cooling plate, and the HFM data a single point measurement of the heat transfer coefficient at the center of the CEB area.
1
38 2
Test #
Tmix [C]
HFM [W/m K]
CEB [W/m K) l 1
51.1 69 42 2
50.4 62 30 3
60.7 62 56 4
59.9 68 62 5
70.9 103 104 6
69.8 109 91 7
75.2 121 118 8
79.7 156 142 9
79.9 157 149 10 85.6 194 183 11 85.3 193 182 12 89.6 264 261 13 89.8 263 265 14 90.2 267 251 Table 4.1: Air / Steam Natural Convection Results
39 AtriSteam fests at 60 C and 55 C-300 HFM 60 C -
HFM 60 C ---
HFM 85 C J, O HFM 85 C - ~
7
./D y
400
\\<. :f
~
}
v
..u.
g..
150
- %/
0 2
100 p
So 0
1 2
3 4
5 6
7 Bottom Position Along Test surface Top Fig.4.4: Repeatability Using Heat Flux Meter Method Atr/ Steam Tests at 60 C and 83 C 300 CEB 60 C -
CEB 60 C ---
CEB 85 C ---
30 CEB 85 C -
g
?y 3:.
200 It 3
E 150 8u E
~
100 u3 50 0
1 2
3 4
5 6
7
- Bottom Position Along Test surface Top 1
l Fig.4.5: Repeatability using Coolant Energy Balance Method -
1
- r-e:
m.
J. '
m m
.m
40 The difference between the average heat transfer coefficients measured using the coolant-energy balance versus those using the heat flux meters was typically 5-20 percent, with the larger discrepancies being associated with the lower temperature runs.
According to Huhtieniemi [2], part of the discrepancy is due to the nature of the measurement techniques:
the heat flux meter measured a local heat flux at the midplane, whereas the coolant balance method averaged the heat flux over the cooled area. A non-linear heat flux profile would a
lead to slightly different results for the two techniques. The higher uncertainty associated with the coolant energy balance method (Appendix A) indicates the discrepancy at low temperatures may be due to limited accuracy in coolant flow temperatue readings.
The discrepancy near the bottom of the test plate for separate runs under identical test conditions was a cause for concern. Test results consistently showed a relatively large discrepancy between identical runs near the bottom of the plate, with more precise values higher on the test surface, as evident in Figures 4.4 and 4.5. The measured average heat transfer coeflicients varied by up to 30 percent for identical test conditions near the bottom of the plate, while values corresponding to positions near the top of the plate typically remained within 2-3 percent. Note, however, the 30 percent discrepancy is an upper bound,
')
1 l
with most repeat heat transfer coefficient measurements remaining within 10 percent of the initial test.
The fact that the upper portion of the test section is not subject to the variation in test results, indicates an anomaly in the test conditions rather than an experimental or systematic error. While no evidence can be provided,it is possible that the turbulent nature of the flow in this region, as discussed in the next section, is transitional. That is, the assumption that i
I
41 steady state conditions are established does not hold. Extended periods of turbulent flow patterns may be followed by periods of relatively calm, parallel flow. The time periods involved would have to be on the same order, or longer, as the data acquisition period (approximately 15 minutes), because no transitions were observed while data was being collected. Any evidence of oscillatory behavior would partially support the findings of Fox et. al [15]. Tests were not run in the region where Fox et. al reported a stable mixture (noncond=nble mass fraction below 0.10).
4.3.1.2 Visualization Tests The characteristic shape of the heat transfer coefficient curves (See Fig. 4.4 and 4.5) with higher values on either end of the plate, and relatively lower values near the center was evident throughout the series of experiments. Traditional heat transfer results for forced flow would lead us to expect a high heat transfer coefficient near the leading edge of the plate, with an asymptotic decrease towards the end of the cooled surface. Huhtienimi [2]
attributed the slight increase evident in earlier forced convection experiments to the onset of a natural circulation pattern near the test section exit. The orientation of the platein the natural convection experiments was such that natural convection flows would be produced parallel to the cooled plate rather than in the orthogonal direction. Two explanations for the increasing heat transfer coefficient at the bottom of the vertical cooled plate are: Increasing parallel flow velocities with distance down the plate, and turbulent velocity profiles other than those addressed by Huhtienimi. Although it was not possible to measure velocity profiles (Appendix D), both of these phenomena are supported by the visualization tests.
42 The pnmary reason for constructing a Lexan" walled test section was to allow visual observations of the flow patterns and condensate film inside the control volume. In these particular runs a piece of the insulating material was removed from the test section wall. A i
heat gun was used to raise the temperature of the Lexan sheet ahcue the mixture saturation temperature (approximately 80 C). With the test wall clear from condensate, video was obtained.
Observations were made at discrete locations along the test section. Special emphasis was placed on isolating flow patterns adjacent to the individual heat flux meter locations (Fig.4.6). Following is a summary of the observations recorded:
- Adjacent to HFM 1,2 (Fig.4.7)
Very turbulent velocity profiles. Rising steam / air mixture along adiabatic wall. The rising plume tended to roll towards the cooled plate. A heavier, misty layer near plate fell rapidly and tended to curl away from cooled plate. The net result was a system of vortices approximately 6-10 inches tall ranging in width from ~2 inches to l
the entire width of the test section (6 inches). No regular or repeating pattern was evident.
- Adjacent to HFM 3 Some of the same characteristics of the flow near HFM i and 2 were evident at this location but turbulent flow patterns were less pronounced. This location appeared to serve as a transition area between turbulent flow patterns below and parallel flow associated with positions above.
l j
i i
43 Light Source v
- s
- W '
HFM 7 f@f;Q}u.ljgif HFM 6 Observation Window ig8tR$
EQ79%i HFM 5 ONkk
,Mf'3$e&
HFM 4 7s 3
4 7 ':('
s
'i.
W HFM 3 r1+ #
~'~
'&M i
's%w$% 9 HFM 2
'W g g@l.
HFM1 m.
m, u m, a
s Entrance Length n
i Steam Inlet i
Fig. 4.6: Location of Heat Flux Meters (HFM) along test plate I
44 O/ %0 HFAI 7 p?
y t
4 t_
E HFM 6 4
t 4
+
4 HFM5 a.
Decreas.ing Water Droplet Density
?
u as HFM 4 T-@ z" aqw :yerrggg df shdEu HFM 3 a-[Ms@jgp%4$* 7 s
w,ww n
I @N,2.,.kkkh[d2h5
[G d?%ihrhkhN HFM 2 n~ '
' gj'
' Vi HFM1 Fig. 4.7: Flow Pattem Observed During Visualization Tests
45
- Adjacent to HFM 4,5 Rising plume of air / steam near adiabatic wall with very limited turbulence. Similar downward flow of heavier, misty layer near plate.
.3
- Adjacent to HFM 6 Same parallel flow pattern as at position 4 and 5, however, velocities were reduced.
In addition, number density of airborne water mist was limited. The flow was beginmng to stagnate with velocity near both sides of the test section reduced to a fraction of those at lower positions.
- Adjacent to HFM 7 The flow at the top of the test section appeared to be largely stagnant. There was an area of flow reversal at the very top of the test section, but flow velocities were low.
Also, very little mist was present, making visual observations difficult. Small areas of turbulence were observed at top corners of test section. However, irregular flow patterns near top end plug may have been due to heat from light source positioned immediately above.
The observations outlined above were used to interpret the shape of the heat transfer coefficient' when plotted versus position along the test plate. The high heat transfer coefficient at the top of the test section was a result of high steam concentrations and little or no boundary layer near the cooled plate. Increasing boundary layer thickness and increasing noncondensable gas fraction would result in the initial drop in the heat transfer coefficient moving from position 7 to position 6. As the flow accelerates, two competing effects, decreasing steam concentration, and increasing mixture velocity result in the modest
46 increase in values adjacent to positions 5 and 4.
Position 3 was difficult to characterize because turbulent flow patterns were evident in this region, although not as prevalent as those near positions 1 and 2, which would indicate an increase in the heat transfer coefficient.
However, the small turbulent eddies, along with the decreasing steam concentration, and possibly a reduction in the parallel flow velocity near the plate, reduced the heat transfer near position 3. Finally, turbulent flow patterns near the bottom of the plate resulted in increased heat transfer coefficients at positions 2 and 1. The small decrease in heat transfer values going from position 2 to position 1, was attributed to a further reduction in steam concentration.
4.3.2 Air / Helium / Steam Heat Transfer Coefficient Measurements Average heat transfer coefficient results for the air / helium / steam natural convection test series are given in Table 4.2. Mixture temperatures ranged between 60*C and 90 C with data being taken at 60,70,80,85. nd 90 C. Tests covering this range of temperatures were repeated for helium molar concentraticas between 4 and 30% of the noncondensable gas present (between 1.7 and 20% of the total mixture). Although more experiments were run, only those which correlate well with the desired experimental conditions are given here. A complete list of results is given in Appendix C.
The experimental error in the heat transfer measurement was determined to be approximately 10 percent (Appendix A). Tests were, again, repeated in order to check the
47 2
Test #
T.,,, (*Cl He [%)l.HFM [W/m Kl CEB [W/m K1 1
70.3 3.9 107 103 2
80.4 3.9 159 147 3
84.8 3.9 l 189 180 4
85.4 3.9 191 189 5
84.7 3.9 191 191 6
89.4 3.9 l 245 250-7 90.5 3.9 266 272 8
59.4 7.4 73 61 9
61.1 7.4 l 79 65 10 70.2 7.4 l 104 91 11 80.1 7.4 l 158 146 12 85.6 7.4 l 194 189 13 89.7 7.4 237 253 14 69.7 15.2 95-93 15 70.2 15.2 103 94 16 79.7 15.2 156 146-17 85.0 15.2 198 188 18 89.9 15.2 258 261 19 68.0 30.0 89 85-20 69.8 30.0-98 90 21 80.3 30.0 154 141 22 84.3 30.0 186 177 23 84.5
.~ 0. ^
181 162 24 90.1 30.0 l 263 264 Fig.4.2: Air / Helium / Steam Natural Convection Results
48 Atr/HebunvSteam f ests at 00 C and 55 C 300 HFM 85 C 4% Helium -
HFM 85 C 4% Helium --
250 HFM 60 C 7.4% Helium -
HFM 60 C 7.4% Helium -
'd
}
..... - ~
3
- m 3
j 5
150 8o U
100
- /)N
,8
)
\\
f E
s..
50 0
1 2
3 4
5 6
7 Bottom Position Along Test Surface Top Fig.4.8: Repeatability Using Heat Flux Meter Method Atr/HebunVSteam fests at 60 C and 55 C 300 CEB 85 C 4% Helium -
CEB 85 C 4% Helium -
250 CEB 60 C 7.4% Helium --
2 CEB 60 C 7.4% Helium -
?
t E'c 5
150 8v E
l 100 A
3 9
0 1
2 3
4 5
6 7
Bottom Position Along Test Surface Top Fig.4.9: Repeatability Using Coolant Energy Balance Method i
)
49 i
consistency of the heat transfer results. Figures 4.8 and 4.9 are plots of the heat transfer coefficient along the length of the condensing plate at 60"C (Test #8 and Test #9) and 85"C (Test #3 and Test #5). Figure 4.8 contains data collected using the HFM method, while Figure 4.9 contains data collected using the CEB method. The uncertainty associated with the helium concentration was 7-11% (See Appendix A for uncertainty calculation), with the i
higher uncenainties associated with the low helium concentration runs. Figures 4.10 i
through 4.13 are graphic illustrations of the heat transfer coefficient along the length of the condensing plate for the entire range of temperatures 4 and 30% helium concentrations.
Again, the coolant energy balance (CEB) data is displayed as a horizontal line, and the heat f
flux meter (HFM) data is represented by a single point at the center of the horizontal line.
A comparison of the air / steam and air / helium / steam restdts offers no discernable difference between the respective heat transfer coefficients. Figure 4.14 is a typical illustra-tion of the consistent heat transfer coefficient measurements collected in the two series of-tests. Even at the relatively high helium concentration of 30%, no large differences from t
the air only runs is seen. It was concluded that helium concentrations in the range studied have no measureable effect on heat transfer rates in this test series within the experimental error.
7
50 Four Percent Hebum 350 HFM 70 C -
CEB 70 C -
300 HFM 85 C -
CEB 85 C -
- e Yj 25 0 1
{
+
E I
J 2T e
I O
150 I
f s
p 100 3=
i 50 0
1 2
3 4
5 6
7 Bonom Position Along Test Surface Top Fig.4.10 Air / Helium / Steam Heat Transfer Coefficients Four Percent Hebum 1
350 HFM 80 C -
CEB 80 C -
300 HFM 90 C -
CEB 90 C -
{
u
?
t 4
250 is
+
6 i
a g
U 150 y
t ii4 100 3=
$0 0
1 2
3 4
5 6
7 Bot:om Position Along Test Surface Top Fig.4.11. Air / Helium / Steam Heat Transfer Coefficients
51 Thirty Percent Helium 350
,HFM 70 C -
CEB 70 C -
300 HFM 85 C -
CEB 85 C -
w 73 250 3
I 2
I
.g 2M I
e
+
u 0
150 t
+
E e
s 100 2
'- 7-- T y
.... s
$0 0
1 2
3 4
5 6
7 Bottom Position Along Test Surface Top Fig.4.12: Air / Helium / Steam Heat Transfer Coefficients Thirty Percent Helium 350 HFM 80 C -
CEB 80 C -
300 HFM 90 C -
7 CEB 90 C ~
r w
2
?
1
+
g no 5
-1 a
.2 200
'\\
E e
1 u
150 3
g a_..
100 3
'l
=
l 50
^1 0
1 2.
3 4
5 6
7 Bonom Position Along Test Surface Top l
l Fig.4.13:. Air / Helium / Steam Heat Transfer Coefficients
.i
52 Att only and Duny Percent Hebum 350 30% He HFM 90 C -
30% He HFM 70 C -
Air HFM 90 C 300 I '*
Air HFM 70 C -- -
x-f-
I
?
v E
250
.. },.
A..
S7
.y
.9 200 5'J 3
0 150 U
T I.
l 2
-\\
/
^
100 N
g
.s 1
50 0
1 2
3 4
5 6
7 Bottom Position AlongTest Surface Top Fig. 4.14: Comparison of Air / Steam Only and 30% Helium Test Results i
53 Chapter 5 Conclusions and Recommendations 5.1 Conclusions An experimental investigation to determine the effects of helium and air on condensation heat transfer rates was conducted. Results indicate the heat transfer rates are not sensitive to small amounts of helium (4-30 molar percent of the noncondensable gas) when helium replaces air on a molar basis. Higher helium concentrations may promote condensation heat transfer, because helium in steam has a higher thermal conductivity and larger diffusion coefficient than air in steam. However, due to its low density, helium can stratify and decrease the steam concentration gradient near horizontal, downward facing, condensing surfaces. The decrease in the steam concentration gradient would lead to diminished heat transfer rates. Under sorne circumstances the two effects may counterbalance each other, as our data suggests, making it necessary to quantify the effects separately. While the inhibiting effect of helium on heat transfer rates reported by previous investigators was not observed, it should be noted that care must be taken when comparing similar, but
54 fundamentally different, experimental studies.
For example, Dehbi [4], measured the decrease in the heat transfer coefficient associated with helium by comparing mixtures of constant mass fraction. This approach accentuated the effect of helium because by replacing equal weight of air with helium results in higher molar concentrations of the noncondensable gas. In contrast, this study effectively replaced previously existing air with an equal amount of helium, on a molar basis, maintaining a constant noncondensable to steam molar ratio.
5.2 Recommendations Experimental work in the field of condensation heat transfer in the presence of noncondens-able gases is a challenging and diverse field. Funher characterization of the effect oflight l
gas, particularly in the case of condensation on a vertical, cooled surface is needed.
Improved instrumentation would be a critical aspect of any proposed study, including accurate local heat flux measurements, strict control on the noncondensable mass fraction, and if possible, a method of accurately determining flow fields within the test facility. The measurement of flow velocities proved to be an especially challenging task, with failed attempts documented in Appendix D.
Problems associaad with the various technigt.es discussed could likely be circunwented. In particujar, a hot-film probe could be used under conditions with very limited airborne moisture. While velocity data could not be collected under all the conditions of-interest, any detailed insight into the flow behavior of th:
noncondensable/ steam mixture would be valuable. Another aspect of condensation in the containment which must be characterized is the trajectory at which the air / steam or i
i j
55-air / helium / steam impinges upon the test surface. Many geometries must be considered to a
model all the possible design base accident scenarios.
O l
(
l l
l
56 Bibliography
[1] Westinghouse Electric Corporation. Tests of Heat Transfer and Water Film Evaporation from a Simalated Contamment to Demonstrate the AP600 Passive Containment Cooling 1
System, Westinghouse Report NSE-90-0013, Jan 1990 i
[2] Huhtiniemi 1. K., Condensation in the Presence ofa Noncondensable Gas: The Effect of l
Surface Orientation, PhD Thesis, University of Wisconsin,1991.
[3] Cho D. C., Stein R. P., " Steam Condensation on the Underside of a Horizontal Surface",
1
(
Proceedings of Third International Topical Meeting on Nuclear Power Plant Thermal Hydr.
and Operations, Nov.1988.
[4] Dehbi A.
A., Analytical and Experimental Investigation of the Effects of Noncondensable Gases on Steam Condensation under Turbulent Natural Convection 4
l Condidtion. PhD Thesis, Dept. of Nuclear Engineering, MIT, Jan.1991.
57
[5] Barry J. J., Efects of Interfacial Structure on Film Condensation, PhD Thesis, University of Wisconsin,1987.
[6] Dallmeyer H., Stoff-und Wbrme&bertragung bei der Kondensation eines Dampfes aus einem Gemisch mit einem nicht kondensieren Gas in faminarer und turbulenter Strbmungs-gren:schicht, VDI-Forschungs-Heft 539, pp. 5-24,1970.
P
[7) Gerstmann J., Griftith P., Laminar Film Condensation on the Underside ofHori:ontal and Inclined Surfaces, Int. J. Heat Mass Transfer, vol 10, pp. 567-580,1967.
[8] Henderson C. L., Marchello J. M., Film Condensation in the Presence afa Non-condensable Gas Transactions of ASME, J. Heat Transfer, vol. 91(3), pp. 447-50, 1969.
[9) Kroger D. G., Rohsenow W. M., Condensation Heat Transfer in the Presence ofa Non-condensable Gas, Int. J. Heat Mass Transfer, vol.11, pp.15-26,1968.
[10] Kutsuna H., Inoue K., Nakanishi S., Filmwise Condensation ofBapo Containing Noncondensable Gas in a HorizontalDuct, int. Symposium on Heat Transfer, Beijing 193' U
[11] Robinson J. A., Windebank S. R., Measurement ofCondensation Heat Transfer '
Coeficients in a Steam Chamber Using a Variable Conductance Heat Pipe, Proc. 2nd UK National Conference on Heat Transfer, vol.1, pp. 617-637, Sep.1988.
i i
J
)
~
]
58 i
[l2] Slegers L., Seban R, A., Laminar Film Condensation ofSteam Containing Small Concentrations ofAir, Int. J. Heat Mass Transfer, vol.13, pp.1941-1947,1970.
[l3] Spencer D. L., Chang K. I., May H. C., ExperimentalInvestigation ofStability Effects in Laminar Film Condensation on a Vertical Cylinder,4th international Heat Transfer Conference, Paris, vol. 6, Paper Cs 2.3,1970.
1
[l4] Suryanarayna N. V., Malchow G. L., Film Condensation on inclined Plane Surfaces.
l Transactions of ASME, J. Heat Transfer, vol. 97(1), pp. 79-82,1975.
[15] Fox R. J., Nagasaki T., Hijikata K., Peterson P. F., Heat Transfer and Stability Phenomena in Gas Loaded Condensers, Dept. of Nuclear Engineering, University of I
California, Berkeley, Paper submitted for publication.
1
[16] Siddique M., Golay M. W., Kazimi M. S., Local Heat Transfer Coeffivientsfor Forced Convection Condensation ofSteam in a Vertical Tube in the Presence ofAir, Dept. of Nuclear Engineering, MIT, Paper submitted for publication.
l l17] Kang H. C., Kim M. H., Characteristics ofCondensation Heat Transfer with Wavy Interface in the Presence ofNoncondensable Gas, Full length paper submitted to NURETH 6.
l
59 i
[18) Bird R. B., Stewart W. E., Lightfoot E. D., Transport Phenomena, Jo
[19) Box G., Hunter W., Hunter J., Statistics for Experimentr s, John Wi 1
60 Appendix A Error Analysis Calculations A complete error analysis is presented by Huhtiniemi [2] for this facility and the experimentally measured quantities. The error calculation for the heat transfer coefficient is reproduced here, with the permission of the author, along with an uncertainty calculation for the helium concentrations which was not covered in Ref.[2].
A1:
Coolant Energy Balance Heat Flux Measurements The goveming equation for calculating the heat flux using the coolant energy balance is:
9m Cp(5'AT,,1):
N Of
- Ag The maximum change in water density, p, and heat capacity, C,, over the temperature range of interest is 0.2 percent. Therefore, the error in the computed heat flux is essentially a function of the measured quantities i' and ATcoa.
I 61
-a Error in the flowmeter reading is certified by Dwyer Instruments Inc. to be less than two percent of the full scale reading. The maximum flow rate through the meter is 3.8 liters per minute, resulting in an accuracy of i'm, = 1.3 ml /sec.
Error in the temperature rise ATcoa is caused by uncertainty. of the measured temperature difference between T;, and T.,
Using the error propagation formula,' the cumulative uncertainty can be obtained:
6Ter,o, = )(T'",,,,)2 + (Tlyo,)2 1
In this case the absolute accuracy of the thermocouple is not important, because' the difference between two thermocouple signals is measured. Therefore, the uncertainty is independent of the absolute irder and outlet temperatures. Instead, the uncertainty can be traced to the resolution of the measurement, which is 0.02 *C. This leads to a AT,,,,, of 0.03*C.
Combining the AT,,,,r equation with the heat flux equation presented earlier, the relative error can be written as:
//
- '~ ' = (' ^ ~~' )2, ( V,-- ) 2 q"
ar j,
The temperature rise in a cooling loop is inversely dependent on the flow rate.
Therefore, a criteria between minimum flow rates and temperature rise can be obtained so I
that the relative error stays below some preset value. In this case, the error due to the heat flux measurement were maintained below 10 percent.
-~
l 62 A2:
IIcat Flux Meter IIcat Flux Measurements If the thermocouple temperatures are measured at specified locations, the heat flux is given by the following equation:
9: = N~ar, u
The thermal conductivity of the aluminum plate is known within 0.6 W/m*C [Ref.28 IK].
Comparing the uncertainty with the thermal conductivity of 121 W/m*C, the uncertainty related to the thermal conductivity of the aluminum can be ignored.
The temperature gradient is dependent on the relative temperatures and spacing between thermocouple junctions. The error estimate for the temperature differences are related to the resolution of the thermocouple measurement.
Therefore, the error estimate for the temperature measurement is T,,, = 0.02 "C.
Consequently, the variance of the error (T,m,,)
can be assumed constant. Positions of the thermocouples in the heat flux meters were recorded with an accuracy of x,,,, = 0.05 mm. Now, assuming the error in the position measurement is negligible, the following relationships can be used to obtain an estimate for error in the heat flux measurement [19, page 459]. First, the experimental error variance is obtained from the following expression:
k(Tr-h*
2 n-1 4
where n is the number of thermocouples in a heat flux meter (n = 4).
m
.m m
I 63' Once the error variance has been computed, the estimated standard error of the heat flux can be expressed as:
2
//
g Gerror = k a
Ei=11!
This estimate for the standard error is dependent on test conditions and is different for each heat flux meter. Typically, the standard error of a heat flux estimate was less than 3 percent. However, the reader should note that the previous analysis ignores possible unknown systematic errors due to technical difficulties, e.g., imperfect thermal contact between a thermocouple assembly and the test plate.
A3:
Heat Transfer Coefficient Measurements The heat flux errors calculated above can be used in conjunction with the error propagation formula to calculate an estimate of the heat transfer coefficient uncertainty:
q','=,
2 (Twn)*=' )2 4 [ (T,5),-,
h,-,
2 u
T,-T_y T,u-T a h
q Note: Although this analysis is taken from Huhtiniemi [2], the preceding equation was corrected from the original.
The above equation was used to calculate the uncertainty in the heat flux measurements, which are included as error bars on the heat transfer coefficient figures.
i
64 A4:
Uncertainty Calculation for IIelium Concentration Values The uncertainty in the helium concentration in both the forced flow and natural convection test series can be estimated using the accuracy of the flowmeters which delivered the noncondensable gases and the accuracy of the T, measurement.
The volume flow rate of the Dwyer RMC series flowmeters are certified to within 2 percent of the full scale reading. The Dwyer RMB series flowmeter, which was used in the 30 percent helium natural convection run, is certified to within 3 percent of the full scale reading.
The mass flow rate of gas into the test section is given by the equation:
ga = pga V a m
g The density term is considered constant in this analysis because the pressure of the gases in the helium tests was constant at I bar to within the resolution of the pressure gauge.
Because of the constant density assumption, gas concentration and their uncertainties are calculated using volumetric flow rate measurements only.
l The highest relative uncertainty in the helium concentration will occur at the lowest helium concentrations. The lowest helium concentration tests were the natural convection f
\\
experiments with 4 mole percent helium (of the noncondensable gas present) at 90*C.
Which corresponds to a helium mole percent of 1.2 for the entire mixture.
1
65 The uncenainty in the helium volumetric flow rate is given by:
. He
. He
. He V,,,,, = V,,,,,(resolution of reading)+V,,,,,(accuracy of flowmeter)
The flowTneter reading resolution was determined to be 0.1 ft'/hr for the RMB flowmeters (10 ft'/hr full scale), and 0.1 ft'/ min for the RMC (20 ft'/ min full scale) flowmeters.
Along with the certified 3 percent accuracy of the flowmeter, the corresponding uncertainty for the RMB flowmeters is 0.4 ft'/hr. The corresponding uncertainty for the RMC flowmeter, with a 2 percent accuracy,is 0.5 ft'/ min.
The He concentration in the noncondensable mixture alone is given by:
. He. He V i V,,,,,,
gf, _
I
~. no. ns
.Aw ou V tVam,+V k Vam, Using the error propagation formula, the following equation is developed for the uncenainty of the helium concentration in the noncondensable mixture:
e Air
- H*
e He
/$or = (Verror)2(.u[".,,
)2 + (V,,,or)2C.u[.a, )2 (V +V 32 (V +V
>2 U
The calculation for the 4 mole percent helium tests, resulted in f ' = 4.0 0.4%, which corresponds to a 10% uncertainty in the helium concentration.
When the steam flow is added to the noncondensable mixture, the uncertainty in the saturation temperature must be taken into account. The uncertainty in the T reading is m
66 O.5"C which corresponds to a maximum uncr.rtainty in the steam mole fraction of 0.01. The uncertainty in the helium concentration in the mixture is then given by:
i GHe =)"' X b)noncondensables A worst case scenario is apparent, when the concentration of helium is low, and the molar _
ratio of noncondensable gases is also low. The case of 4 molar percent helium at a mixture temperature of 90*C will result in the highest relative uncertainty in this case. Working through the error propagation for this test results in a helium concentration uncertainty of f
10.5 percent in the final mixture, which represents the maximum uncertainty for the test series. A similar calculation for a 30 percent helium concentration test resulted in a max-imum uncertainty of 7.5 percent.-
b I
B ww n
g
--,-w---+
+wo, m
e
67 Appendix B Steam / Air Natural Convection Results 2
2
- Test #
Tm ["C]
HFM [W/m K]
CEB [W/m K]
Therm 3 81.7 163 165 Therm 4 50.4 62 30 Therm 5 61.0 52 60 Therm 6 70.9 103 104 Therm 7 75.2 121 118 Therm 8 84.9 207 202 Therm 9 85.3 193 182 Therml0 90.2 267 251 Therm 11 89.6 264 261 Therm 12 49.6 51 40 Therm 13
- 60. '.
75 62 Therml4 69.8 109 91 Therm 15 79.7 156 142 Therm 16 79.9 157 148
i 68 Test #
T., ["C]
HFM [W/m K]
CEB [W/m K]
l 2
2 Therm 17 85.6
.194 183
.l Therml8 90.3 270 253 Therm 19
'89.8 263 265 Therm 20a 51.1 69 48 i
Therm 20b 60.7-62
.56 Therm 21 59.8 68 62.
i i
T S
L
')
l
. 1
69 Appendix C Air / Helium / Steam Natural Convection Results CI:
Test Conditions and Average Heat Transfer Coefficient Results 2
2 l Test #
T, [*C]
He co,,,,,,,
HFM [W/m K]
CEB [W/m K]
Therml01 58.8 4.7 0.19 77 78 Therml06 80.9 4.7 0.49 162 156 Therm 107 59.7 3.9 0.20 54 59 Therm 108 62.1 3.9 0.22 66 56 Therml10 70.3 3.9 0.32 107 103 I Thermi11 79.7 3.9 0.47 156 145 Therml12 80.4 3.9 0.48 159 147 lThermll3 85.8 3.9 0.60 201 191 l Therm 114 84.4 3.9 0.56 192 183 Thermil5 84.8 3.9 0.57 189 180 Therml16 85.4 3.9 0.59 191 189 Therm 118 84.7 3.9 0.57 191 191 Therm 119 89.5 3.9 0.69 253 253 Therm 121 89.4 3.9 0.69 248 253 Therm 122 89.4 3.9 0.69 245 250 Therm 123 90.3 3.9 0.71 266 272 Therm 124 59.6 7.4 0.20 68 61
70 2
2 Test #
.T
[C]
He[%) co,,,,,,,
HFM [W/m K]
CEB [W/m K]
Therm 125 59.4 7.4 0.19 73 61 Therm 127 61.1 7.4 0.21 79 65-Therm 128 70.2 7.4 0.31 104 91 Therm 129 80.1 7.4 0.48 158 146 Therm 130 84.7 7.4 0.57 199 194 Therm 131 85.2 7.4 0.58 198 193 Therm 132 85.6 7.4 0.59 194 189 Therm 133 89.7 7.4 0.69 237 211 Therm 134 90.4 7.4 0.71 255 262 Therm 135 69.7 15.2 0.31 95 9'
Therm 136 72.5 15.2 0.35 105
'/4 Therm 137 68.3 15.2 0.29 98 94 Therm 139 71.3 15.2 0.33
-107 95-Therm 140 68.7 15.2 0.30 97 87 Therm 142 70.2 15.2 0.31 103 94 Therml43 79.1 15.2 0.46 151 138 Therml44 79.7 15.2 0.47 156 146 Therml45 85.3 15.2 0.58 198
'187 Therml46 85.0 15.2 0.58 198 188 i
Therml47 90.2 15.2 0.71 270 265 Therm 148 89.1 15.2 0.68 250 251 Therm 150 89.9 15.2 0.70 258 261 Therm 153 68.0 30.0
-0.29 89 85 Therm 154 69.8 30.0 0.31 98 90 q
Therm 155 80.9 30.0 0.49 162 150 i
Therm 156 78.6 30.0 0.45 145 138 Therm 157 80.3 30.0 0.48 154 141 Therm 158 84.3-30.0 0.56 186 177 Therm 159 84.5 30.0 0.57 181 162 Therm 160 90.6
-30.0 0.72.
279 263 Therm 161 90.1 30.0 0.70 265 257 Therm 162 90.1 30.0 0.70 263 264
-1
71 C2:
IIeat Transfer CoefHelents as a Function of Plate Position Four Percent Helium 350 HFM 70 C -
CEB 70 C -
300 HFM 85 C -
CEB 85 C -
M e4 I
30 E
}
+
E I
.2 200 2
1
+
n 8
U 150 5
h 0
.. L y-p 100
+
I!
f
=
50 0
1 2
3 4
5 6
7 Bottom Position Along Test Surface Top 1
Four Percent Heliurn 350 HFM 80 C -
CEB 80 C -
300 HFM 90 C -
CEB 90 C -
I
+
I' t
w r
1 5
no
.4 200 e
i i
U 150 B
l 100 i
5.
50 0
1 2
3 4
5 6
7 Bottom Position Along Test Surface Top 1
72 Seven Percent Helmm 350
,HFM 60 C -
CEB 60 C -
HFM80C -
300 CEB 80 C -
HFM 90 C -
o i
J 250 l
CEB 90 C
--4 m
3 s
e e
E i
e
.2 200 1
s 3
U 150 1
+
r u
"E 1
A 100
+
b 50 F.--
0 1
2 3
4 5
6 7
Bottom Position Along Test Surface Top Seven Percent Helium 350 HFM 70 C -
CEB 70 C -
300 HFM 85 C -
~
x CEB 85 C -
?
E 30 I
3 I
T E
i
}
.g 2M e
8
+
0 150 U
q 1
y 100
,r-y 50 0
1 2
3 4
5 6
7 Bottom Position Along Test Surface Top t
I e
n
73 PLu P rcent Hehum 350
,HFM 70 0 -
CEB 70 C -
300 HFM 85 0 -
CEB 85 0 -
-x
?g 30 1
+
--I
.f 200 gg
+
8 0
1$0 E
I T
p 100 i
+
.3 I
50 0
1 2
3 4
5 6
7 Bonom' Position Along Test Surface Top Fifteen Percent Hehum 350 HFM 80 C -
CEB 80 C -
HFM 90 C -
300 t
CEB 90 C ~
k a
r 3
30
+
+
ic.
4
.o 200 G
E
=
]._..
8 u
150 1
1 y
I 3#
100 50 0
1 2
3 4
5 6
7 Bonom Position Along Test Surface Top 4
9 0
74 Thiny Percent Helmm 350 I
HFM 70 C -
CEB 70 C -
300 HFM 85 C -
CEB 85 C -
g Yg 250 7
E I
.g 200 3
c l
A y
+
1 0
150
+
3 a
B p
100 I
g
.4 e
50 0
1 2
3 4
5 6
7 Bonom Position Along Test Surface Top Tluny Percent Helium 350 HFM 80 C -
CEB 80 C -
HFM 90 C -
300 7
CEB 90 C '--*
7 y
+
250 3
.h 2%
a I
150 5
?
g 100 E
50 0
1 2
3 4
5 6-7 Bonom Position Along Test Surface Top
75 Appendix D Attempts at Measuring Velocity Profiles D1 MKS Pressure Transducer
==
Description:==
MKS Baratron Pressure Transducer Type 223B 0.1" H O Full Scale 2
Basic Operation:
The Type 223B Pressure Transducer measures differential pressure according to its full scale range and provides a 0 to I volt signal which is linear with pressure. The transducer is composed of an inconel sensor, printed circuit board, and cover. The sensor consists of three parts: a thin metal diaphragm, a ceramic electrode which senses the diarhragm's deflection, and a glass to metal feedthrough terminal which passes the -
electronic signal to the printed circuit board. A standard Pitot tube is connected to the transducer using stiff walled tubing. The static and dynamic pressures experienced by m
_--_a-,,
- - - - - - - - - ~ - - - - - - - - - - - - - -
76 the Pitot tube are transferred though the tubing to opposite sides of the metal diaphragm.
The measure of diaphragm displacement is converted to the O to i volt output signal.
L Results:
l No useful data was collected using the MKS Pressure Transducer. As soon as the steam valve was opened the signal began to oscillate widely (0.5 volts, ~10 ms). There is a second, overlying signal of higher frequency (0.5 volts, ~1ms)which also distorts the output. After approximately 10-20 seconds the Pitot tube plugs up with condensed steam, resulting in a near 0 volt signal.
Explanation:
i i
It was apparent that fluctuations in the steam flow rate cause small changes in the system pressure. The transducer does not sense changes in the system pressure on both sides of the diaphragm at the identical times, and therefore a dynamic pressure reading is produced. The second, overlying signal, may be a result of two-phase considerations.
It is possible that water droplets are impinging on the Pitot tube and are producing these fluctuations.
7 1
?
77 D2 Hot Film Probe
==
Description:==
TSI Hot Film Probe Model 1210-60w (designed for use in liquid flows). Designed for use with TSI Intelligent Flow Analyzer (IFA 100).
Basic Operation:
i The probe is a small resistance element which is heated to some preset temperature.
l The amount of electrical current needed to maintain the probe at its elevated 1
temperature is a measure of the level of cooling provided by the fluid flowing past the In general, when the IFA 100 is turned on, current flows though the sensor sensor.
which is part of an internal bridge circuit. The control amplifier senses any imbalance j
in the bridge, caused by varying probe resistance, and feeds back more or less current i
until the bridge comes into balance. The amount of cooling supplied by the fluid is a function of the mass flow, fluid properties, and the temperature difference between the sensor and the fluid.
Settings:
The bridge operating resistance was set to 8.06 O which corresponds to a sensor temperature of 250* C. The fluid temperature inside the test section is limited to 100*C, which means there is a minimum 150*C tempertture difference between the hot-film and the adjacent fluid. The probe was calibrated in air with a Pitot tube and a
78 manometer. The resulting calibration curve (volts vs velocity) was consistent with that described in the TSI manual.
Results:
No discernable data was collected from the hot-film measurements. Readings taken from the IFA 100 fluctuated about a value of ~1.75 volts (110%) regardless of position.
or orientation of the probe within the test section. The signal was not viewed with an oscilloscope so an accurate frequency measurement was not possible, but a timescale.of
~1 see was estimated. A value of 1.75 volts was read with the probe inside the test section before steam was added.
Explanation:
The sporadic behavior is likely due to the fact that the fluid whose velocity is being measured is two-phase. The level of cooling experienced by the sensor is influenced by fluid properties, and a water droplet impacting the sensor and then boiling off is one possible explanation for the signal.
D3 Visualization with Strobe Light A final attempt was made at measuring the velocity components of the condensation mixture during the visualization runs. A strobe light was directed into the test section during the run, and video was taken. The intention was to measure the characteristic i
s
?
79 times associated with the turbulent velocity components of the flow. Unfortimately, not enough light was reflected by the airborne moisture, and the measurement was unsuccessful.
L t
e 1
_... ~. _