Anda di halaman 1dari 10

Air Concentration and Velocity Ikleasurements on Self-aerated Flow

down Stepped Chutes

Jorge Matos (1) and Kathleen H. Frizell (2)

(1) Technical University of Lisbon, IST, DECivil, 1049-00 1 Lisbon, Portugal; PH (3 5 1 2 1)


8418 145; FAX (351 21) 8497650; e-mail: jm@civil.ist.utl.pt
Downloaded from ascelibrary.org by University of California, San Diego on 04/12/15. Copyright ASCE. For personal use only; all rights reserved.

(2) U.S. Bureau of Reclamation, Denver, CO 80225, USA; PH (303) 445-2144; FAX (303)
445-6324; e-mail: kfrizell@do.usbr.go\

Abstrnct

The U.S. Bureau of Reclamation (Reclamation) developed instrumentation to measure air


concentration and water velocity in large flumes as well as in smaller indoor models, namely
the large outdoor flume facility located at Colorado State University (CSU) in Fort Collins,
Colorado, and the stepped spillway model chute assembled at the National Laboratory of Civil
Engineering (LNEC), Lisbon.

The basic components of both the air concentration probe and the back-flushing Pitot tube are
briefly reviewed in this paper. Typical model and near prototype air concentration and velocity
profiles are shown, illustrating the highly-aerated, highly-turbulent feature of skimming flow
over stepped spillways.

The experimental results suggest that self-aerated skimming flow over stepped spillways
behave similarly to self-aerated flow on smooth chutes. The shape of the velocity distribution
seem however affected by the large flow resistance.

Introduction

The growth of RCC use in dam construction, as well as the development of new construction
concepts, such as the lightweight wedge-shaped concrete blocks, has renewed the interest in
stepped spillways.

In skimming flow over stepped spillways, large quantities of air entrain upstream of the
spillway toe, even for moderate to high unit discharges. Therefore, the study of the highly
aerated flow is particularly important for the design of the chute sidewalls and the energy
dissipator at the spillway toe.

In order to obtain prototype scale data, the U.S. Bureau of Reclamation (Reclamation)
developed instrumentation to measure air concentration and water velocity in a large, outdoor
flume (Frizell, Ehler and Mefford, 1994). These instruments were subsequently redesigned by
Reclamation for obtaining air concentration and velocity data in a stepped spillway indoor
model, assembled at the National Laboratory of Civil Engineering (LNEC), Lisbon (Matos and
Frizell, 1997).

Copyright ASCE 2004 Water Resources 2000


Building Partnerships
In the present paper, the authors review the operation and calibration of the air concentration
and velocity probes. Some interestin, ~7results on the use of the above instrumentation are
described and compared with those obtained in other experimental studies.

Air concentration instrumentation

Air concentration probe

The construction of the air concentration probes by Reclamation was based upon the resistivity
Downloaded from ascelibrary.org by University of California, San Diego on 04/12/15. Copyright ASCE. For personal use only; all rights reserved.

probe developed by Cain (1978) and tirther investigated by Bachmeier (1988). The probe used
at the CSU large outdoor flume include a 0.2 mm platinum wire set in a 1.59 mm outside
diameter stainless steel tube 50.8 mm long with waterproof expoxy. Further details on
Reclamation’s initial electronic and probe modifications and output can be found in Frizell,
Ehler and Mefford (1994).

In the new design of the air probe developed by Reclamation for use in smaller models, the
conductors are two 0.2 mm platinum wires set in a 6.35 mm outside diameter protective
support, with waterproof epoxy. Only one wire has to be in contact with the bubble to
interrupt the current flowing between the two conductors. Reclamation improved the probe
design by usin,0 the same metals for the circuit, thus eliminating flow induced voltage and
providing better durability of the tip. The electronics were also improved by providing
increased stability, increased accuracy and operation within a range of water conductivity
(Jacobs, 1997).

Air concentration probe calibration

The calibration of the air probes was conducted using the same test apparatus described in
Frizell, Ehler and Mefford (1994). The calibration apparatus consists basically of an inverted
“U” shaped water pipe system with the water and air volumes entering the system measured
with rotameters. Water is supplied to the pipe system on the up leg and an air line with
pressure regulator and rotameter is connected to the water pipe on the down leg of the system.
The air concentration probe tips were located flush with the end of the pipe system nozzle. The
air concentration was then determined as a ratio of the air volume over the air volume plus the
water volume measured in the pipe system.

Although the experimental data gathered for the initial probe was consistent and repeatable
(Frizell, Ehler and Mefford, 1994), improvments have been made in the new design. namely to
linearize the results (Frizell, 1996). Figure 1 sho\vs the experimental data of the new air
concentration probe obtained by Frizell (1996), as well as the readjusted calibration curve
(after Matos and Frizell, 1997).

Air concentration measurements

The experimental data was collected in two stepped chutes: (i) the CSU outdoor chute, a
26.6 degrees sloping flume (lV:2H), 15.20 m high and 1.52 m wide. with block step height of
6.4 cm and capable of passing unit discharges of 2.9 m”/s/m (i.e., critical depth dc = 0.95 m);
(ii) LNEC chute, a 53.1 degrees sloping stepped spillway model (lV:0,75H), 2.90 m high and
1.OOm wide, with step height of 8 cm. Unit discharges up to 0.2 m”/s/m (d(- = 0.16 m) can be
tested in the LNEC flume.
Copyright ASCE 2004 Water Resources 2000
Building Partnerships
C (volumetric air concentration)
-8

0.6
Downloaded from ascelibrary.org by University of California, San Diego on 04/12/15. Copyright ASCE. For personal use only; all rights reserved.

0.4

0.2
Cp (probe reading)
0
0 0.2 0.4 0.6 0.8 1

Figure 1. Calibration curve for the air concentration probe.


(after Mates and Frizell, 1997)

Figure 2 shows normalized air concentration profiles from tests in the CSU chute for a unit
discharge of 2.9 m3/s/m (i.e, de/h = 13.S, where h is the step height). Figure 3 shows the air
concentration profiles obtained at the LNEC stepped chute for q., = 0.08 m’/s/m (de/h = 1.1).
In both figures, y is the distance from the pseudo-bottom (formed by the external edges of the
steps) measured perpendicular to the flow direction. Y90 is the characteristic depth where the
air concentration is 90%, Cm,,, is the mean (depth averaged) air concentration and s is the
normalized distance (s = (L-Li)/di), w here L is the distance measured along the chute, Li is the
distance from the inception point of air entrainment and d; is the flow depth at the inception
point. Figures 2 and 3 also include the air concentration distribution in uniform flow on a
smooth chute of identical slope, following Wood’s (1985) turbulent diffusion model of the air
bubbles within the air-water mixture.

The similarity of the air concentration profiles on smooth and stepped surfaces is apparent on
Figures 2 and 3, which is in accordance with previous findings (e.g., Ruff and Frizell, 1994,
Matos and Frizell, 1997, Chamani and Rajaratnam, 1999). Assuming that the mean air
concentration in uniform flow is similar over smooth or stepped chutes, uniform flow was
likely not reached in the CSU flume (Figure 2). whereas equilibrium appears to have been
attained in the LNEC model (Figure 3). This different result is not unexpected if it is noted
that, in the former situation, the normalized distance at the donwstream end is much smaller.
The results of the growth of the mean air concentration down the stepped chute presented by
Frizell, Matos and Pinheiro (2000) as well as by Matos (2000a), also concurs with the above
reasoning.

The existence of an air concentration boundary layer is apparent on Figure 3, in particular near
the model toe (s = 82.7), whereas in Figure 2 such result is not noticeable. It should however
be noted that in the CSU flume the mean air concentrations are considerably lower due to the
flatter slope. In addition, the air concentrations close to the pseudo-bottom may be slighlty
underestimated, due to the lower sensitivity of the instrumentation in detecting small air
bubbles,
Copyright ASCE 2004 Water Resources 2000
Building Partnerships
1 y/y90 . . J2-, r--
,x=; X Cmean = 0.25 (s = 5.1)
,/’/ c
0.8 , ,
,’ ,’, I“../’ A Cmean=O31 (~~31.1)
/’
I’ /
I
/ I Cmean = 0.33 (s = 39.8)
0.6
,/icyA++ I
Downloaded from ascelibrary.org by University of California, San Diego on 04/12/15. Copyright ASCE. For personal use only; all rights reserved.

/ - - - - Wood (19Sj): Cmean = 0.2:

- - - - Wood (1985): Cmean = 0.3

chute): Cmean = 0.39 ___ Wood (1985): Cmean = 0.3


(Matos. 1999: lV:2H)
-Wood (1985): Cmean = 0.3

I , -

0.2 0.4 0.6 0.8


d

Figure 2. Air concentration distribution on the CSU large outdoor stepped chute.
(qw = 2.6 m3/s/m, dclh = 13.8)

I ~ X Ctnean = 0 39 (s= 6.1)


,’ ‘1 1
0.8 i _’ ,
I xx ~ A Ctnean = 0.53 (s ~33.7)

- Ctnean = 0.6 I (s = 82.7)

Wood (1985). Cmean = 0.39i

~- - -Wood (1985): Cmean = 0.53

~~ Wood (1985): Ctnean = 0.61

chute): Cmean = 0.63 -Wood (1985). Cmean = 0.63


(Matos, 1999: IV 0.75H) --I
1
0 0.2 0.4 0.6 0.8 1

Figure 3. Air concentration distribution on the LNEC stepped chute.


(q, = 0.0s m2/s, dc./h = 1.1)

Copyright ASCE 2004 Water Resources 2000


Building Partnerships
Velocity instrumentntion

Back-flushing Pitot tube

A Pitot static tube designed for mounting on the fuselage of an airplane was used for
measuring velocity in the CSU outdoor flume. The external diameter of the Pitot tube was
approximately equal to 2 cm. The Pitot tube measured the pressure difference between the
stagnation pressure at the upstream end of the tube (through a 1.2 mm hole), and the static
Downloaded from ascelibrary.org by University of California, San Diego on 04/12/15. Copyright ASCE. For personal use only; all rights reserved.

pressure, measured by a port parallel to the velocity, located several diameters downstream of
the tip. A pressure transducer was used to convert these pressures to a voltage difference that
was read from a multimeter.

The velocity from the measured pressure of the Pitot tube is then determined by (Wood, 1983)

2AP
V= (1)
Pw (I- w

where V is the velocity, AP the difference between the total pressure head and the static
pressure head, pw the density of water, C the local air concentration and h a tapping coefficient
which accounts for the non-homogeneous behaviour of the air-water flow approaching the
stagnation point of the Pitot tube.

In order to maintain solid water within the Pitot tube and connecting tubing during operation,
continuous back-flushing of the Pitot tube was provided by means of a constant head source
which fed both the static pressure ports and the total head port of the Pitot tube. The back-
flushing flow rate to each port was controlled by low flow rate rotameters. A complete
description of the schematics of the velocity probe and details of its operation is included in
Frizell, Ehler and Mefford (1994).

The principle of operation of the back-flushing Pitot tube used at the LNEC model is identical
to the above described. The standard Pitot tube had an external diameter of 3.2 mm and the
total and static pressure heads were measured through 1.O mm and 0.5 mm holes, respectively.
The back-flushing flow rate to each port was controlled by needle valves.

Back-flushing Pitot tube calibration

The same apparatus used to calibrate the air concentration probes was used to calibrate the
velocity probes for varying air-water volumes. as detailed in Frizell, Ehler and Mefford (1994).
Air concentrations up to 70 % (volume of air per total volume) were used in the velocity probe
calibration. The water velocity impinging on the total pressure port of the Pitot tube was
assumed to be equal to the water velocity internal to the calibration apparatus tip. The latter
was determined as the water discharge divided by the area of the tip minus the area of the
entrained air.

Copyright ASCE 2004 Water Resources 2000


Building Partnerships
The back-flushing flow rate in air was chosen through laboratory testing as the minimum that
could be passed and still produce a continuous flow out of the Pitot tube ports in air (e.g,
3.79 Uhr, for the probe used in the CSU flume).

In Figure 4, the coefficient C, obtained from the calibration tests is plotted as a function of the
local air concentration, The results show- good agreement with the theoretically predicted
coefficients for air concentrations up to 0.4. For higher air concentrations, the coefficients
from the calibrations are higher. A probable reason for this discrepancy is the non-
homogeneous behaviour of the air-water flow approaching the stagnation point of the Pitot
Downloaded from ascelibrary.org by University of California, San Diego on 04/12/15. Copyright ASCE. For personal use only; all rights reserved.

tube. In actual fact, the ratio of the air bubble size versus the diameter of the pressure ports
should be significant for high air concentrations. On the other hand, even a small back-flushing
flow rate out of the Pitot tube might influence the measured velocity. It seems reasonable to
accept that, for a given flow rate, the measured velocity decreases with the increase of the local
air concentration of the air-water mixture (i.e., C, increases with C). The results expressed in
Figure 4 seem to sustain this physically based thought.

Researchers such as Viparelli (1953) and Lai (197 1) have also used a PitoUPrandtl tube (along
with a mechanical sampler or an electrical probe to measure the local air concentration) to
estimate the local water velocity in self-aerated flow. Figure 5 shows the tapping coefficient
obtained experimentally by Lai (in Cain, 1973) and that estimated by using the experimental
data gathered on Reclamation’s calibration (Frizell, Ehler and Mefford, 1994. Frizell, 1996).

A 1st. data series


16 (Frizell. 1996)
~ -Regression curve

/ ~ X Frizell et al. (1994)

1 - Theoretical curve

08

Figure 4. Velocity probe coefficient determination plot based on


Reclamation’s calibration test results.

Copyright ASCE 2004 Water Resources 2000


Building Partnerships
From Figure 5 it can be concluded that for the data of Frizell, Ehler and Mefford (1994), as
well as for that collected by Frizell (1996), h might be largely greater than 1, due to the back-
flushing flow out of the Pitot tube. With regard to Lai’s (197 1) data, the tapping coefficient h
is relatively close to unity. regardless of the air concentration. However, the application of
Eq. (1) using h = 1, underestimates considerably the velocity for air concentrations greater
than 0.7, as shown in Matos (2000b).
Downloaded from ascelibrary.org by University of California, San Diego on 04/12/15. Copyright ASCE. For personal use only; all rights reserved.

Velocity measurements nntl remarks

In Figures 6 and 7, typical velocity profiles obtained at the CSU and LNEC stepped chutes are
plotted along with the air concentration distributions at the same location. Sensitivity tests
carried out at LNEC showed that for the velocities experienced in the chute (i.e., V < 6 m/s),
even slight modifications in the back-flushing flow rate in air affected the measured velocity
(Matos, 1999). In addition, a constant back-flushing flow could not be completely assured in
successive tests. Therefore, an almost zero flow rate was adopted in LNEC experiments.
Eq. (1) was then applied considering the theoretical value for the tapping coefficient (h = 1).

2.0 1
Backflushing flow out of the Pitot tube
~ J.
1.6 4 A/ ~
r--- Viparelli (1953):
,
I tGoretica1 value
/ 0 Lai (1971)

I X Frizell et al. (1991:

I i A 1st. data series


0.4 h = 1 (theorLtica1 value) (Frizell. 1996)
i ------
(), 0 A---- ~~_- ~~~~
~..~ ~~~~:

0.2 0.4 0.6 0.8 I

Figure 5. Tapping coefficient of the Pitot tube in air-water mixtures.

For low mean air concentrations (C < 0.4), the velocity profiles shown in Figures 6 and 7
exhibit an identical trend, i.e., the velocity increases countinuosly with the distance to the
pseudo-bottom. However, for C > 0.4 to 0.5 (Figure 6) or C > 0.7 to 0.8 (Figure 7), sharp
modifications in the velocity profiles are observed, either corresponding to a sudden decrease
or increase in the velocity. The lower limit values of C at the CSU flume (Figure 6) may
possibly be attributed to the back-flushing flow rate out of the Pitot tube ports in air.

In light of the results contained in Figures 4 and 5, it may be concluded that the velocity
measurements in those flow regions are not accurate. It is interesting to note the similar trend
of the velocity data presented by Chamani and Rajaratam (1999), which were also obtained
with a back-flushing Pitot tube.
Copyright ASCE 2004 Water Resources 2000
Building Partnerships
It should also be mentioned that the structure of the flow in the laboratory testing is not
expected to be similar to that occurring in the self-aerated flow for high values of the local air
concentration. In actual fact, it is not physically possible to obtain homogeneous air-water
mixtures for very high air concentrations (i.e., for C greater than about 0.75) as mentioned in
Falvey (1980) and also experienced by Frizell and others in their calibration tests. In addition,
only part of the total air is entrained in the skimming flow for high values of C, the remaining
being moved between the highly irregular wavy surface. It is believed that the inexactness of
Eq. (1) might also be due to the high frequency of the free-surface waves when compared to
Downloaded from ascelibrary.org by University of California, San Diego on 04/12/15. Copyright ASCE. For personal use only; all rights reserved.

the time response of the Pitot tube, and possibly to the misalignment between the flow
direction and the pseudo-bottom, close to the “free-surface”.

The above reasoning is also in conformity with the velocity profiles presented by Boes and
Hager (1998) which show that the velocity increases continuously up to the depth defined as
YSO.In their study, Boes and Hager used a fiber-optical probe to measure the velocity of the
air-water interface (which equals the water velocity for y < YgO, assuming no slip between air
and water phases) and it was found that all the data collapsed on a power law curve similar to
that observed in conventional chutes (e.g., by Cain and Wood, 1981).

The shape of the velocity distribution seem however affected by the large flow resistance, as
suggested by Chanson ( 1994, pp. 99- 100) and also obtained experimentally by Matos (1999).

0.4 7 Y 6-4 unreliable


velocity data ,, -4 / ..
,,,/:= -_/ ’ ,,,..A’
\.‘\ 7 -, I*-..- _’ ,.’..’
0.3 j “\ A., -- C(s=5.1)
9
m-x V(s=5.1)
- ~- -C(s=31.1)
0.2 1
PA-- V(s=31.1)
~~ -. c (s = 39.8)
--o---- v (s = 39.8)

Figure 6. Air concentration and velocity distributions on the CSU large outdoor stepped chute.
(qw = 2.6 m3/s/m, dJh = 13.8)

Copyright ASCE 2004 Water Resources 2000


Building Partnerships
6 -; y(m) .--.__ \.
.X
unreliable ‘X
velocity data ~----- C(s=6.1)
_._._._._._._._._._.-.-.-.~-$._._._.
l--X-- V(s=6.1)
--c (s=33.7)
-A- - V (s = 33.7)
Downloaded from ascelibrary.org by University of California, San Diego on 04/12/15. Copyright ASCE. For personal use only; all rights reserved.

_______C (s = 82.7)
-E+- V (s = 82.7)

1.0 t-C

3 4 5 6 7 4--- v (m/s)
1 2 8

Figure 7. Air concentration and velocity distributions on the LNEC stepped chute.
(q, = 0.08 m2/s, de/h = 1.1)

Conclusion

The operation and calibration of the air concentration and velocity probes developed by
Reclamation are briefly reviewed in the present paper.

Air concentration and velocity data obtained at the CSU large outdoor flume, and at the LNEC
stepped chute, are also analysed and compared with that obtained by other researchers.

The experimental results indicate that self-aerated skimming flow over stepped spillways
behave similarly to self-aerated flow on smooth chutes. In particular. Wood’s (1985) turbulent
diffusion model of the air bubbles within the air-water mixture fits well the air concentration
distribution in the skimming flow. The shape of the velocity distribution, however, seems
affected by the large flow resistance.

Acknowlcclgntents

Both authors wish to thank Prof. Antonio Quintela, IST, and LNEC, in particular Eng. Matias
Ramos, and INAG - Portuguese Water Institution. for the personal, logistic and financial
support granted to the project. They would also like to thank Reclamation’s Research and
Technology Development program for their support. The assistance of Malin Jacobs, Electrical
Engineer, USBR. and Dr. Palma, Electrical and Electronic Group - LNEC, is gratefully
acknowledged.

References

BACHMEIER, G. (1988). Setup, culilvntiot~ atu’ use of LI measwitlg probe for determinatiorl
of the air corrcentratiorl itr a spilhrvy (‘lurk Diploma Dissertation, Institute of

Copyright ASCE 2004 Water Resources 2000


Building Partnerships
Hydromechanics, Karlsmhe University, Karlsruhe, Germany (Translated from the German,
USBR Translations).
BoES, R. M. and HAGER, W. H. (199s). “Fiber-optical experimentation in two-phase
cascade flow”. Proc, Irlt. RCC L)crm.sSemirlar, Ed. K. Hansen, Denver, EUA.
CAIN, P. (1978). Measuremetrb within self-aerated flow Ott a large spillwcry. Res. Rep. NO.
78-18, Univ. of Canterbury. Christchurch, New Zealand.
CAIN, P. and WOOD, I. R. (198 1). “Measurements of self-aerated flow on a spillway”. J. Of
Hyd. Div., ASCE, 107, HYll, Nov., pp. 1425-1444.
CHAMANI, M. R. and RAJARATNAM. N. (1999). “Characteristics of skimming flow over
Downloaded from ascelibrary.org by University of California, San Diego on 04/12/15. Copyright ASCE. For personal use only; all rights reserved.

stepped spillways”. J. of Hyd. Etqg., ASCE, Vol. 105, No. 4, pp. 361-368.
CHANSON, H. (1994). Hydraulic design of stepped cascades, channels, weirs and spdhvays.
Pergamon, UK.
FALVEY, H. T. (1980). Air-water flow it1 hydraulic structures. USBR Engrg. Monograph,
4 1, Denver, Colorado.
FRIZELL, K.H. (1996). Fir~al air cotlcetltrabotl probe calibration. Private notes, April.
FRIZELL, K.H., MATOS, J. and PINHEIRO, A. N. (2000). “Design of concrete stepped
overlay protection for embankment dams”, Proc. I&l. Workshop or1 Hydraulics of Stepped
Spillways (H.E. Minor and W. H. Hager eds.), Zurich, Switzerland, pp. 179- 186.
FRIZELL, K.H.; EHLER, D.G. and MEFFORD, B.W. (1994). “Developing air concentration
and velocity probes for measuring in highly-aerated, high-velocity flow”. Proc. Hyd. Engrg.
Cor$, ASCE, Buffalo, N.Y., pp. 265-277.
GASTON, M. (1995). Air Entrainmettt atrd etlergv dissipatiotl on a stepped block spilhvq.
M.Sc. thesis, Colorado State University, Fort Collins, Colorado, EUA.
JACOBS, M.L. (1997). Air Cotutlhuliotl Meter Electronics Package Manual. Project Notes
S450-98-O 1, U.S. Department of Interior, Bureau of Reclamation, Denver, EUA.
LAI, K. K. (197 1). Studies qf air etttraitunetlt Ott steep opetl chatmels. M.Eng. thesis, Univ. of
New South Wales, Sydney, Australia.
MATOS, J. (1999). Emr~lsiot~anzetr~o de ar e dis.sipa@io de etlergia do escoamento em
descarregadores em degraus (“Air entrainment and energy dissipation on stepped spillways”).
Research Report, IST, Lisbon (in Portuguese).
MATOS, J. (2000a). “Hydraulic design of stepped spillways over RCC dams”, Proc. Intl.
Workshop OH Hydraulics of Stepped ,Sl)ilhvay.s (H.E. Minor and W. H. Hager eds.), Zuich,
Switzerland, pp. 187- 194.
MATOS, J. (2000b). “Characteristics of skimming flow over stepped spillways - Discussion”.
Accepted for publication. .J. of Hyd. Etqtx., ASCE.
MATOS, J. and FRIZELL, K. H. (1997) “Air concentration measurements in highly turbulent
aerated flow”. Proc. 28th IAHR ~‘ott~~re.s.s. Theme D, Vol. 1, Ed. Sam S. Y. Wang and Torkild
Carstens, San Francisco, EUA, pp. 149- 154.
RUFF, J. F. and FRIZELL, K. H. (1994). “Air concentration measurements in highly-turbulent
flow on a steeply-sloping chute”. /?oc. Hyd EttFg. C’ot$, ASCE, Buffalo, N.Y.,
pp. 999- 1003.
VTPARELLI, M. (1953). “Flow in a flume with 1: 1 slope”. Proc. Mituzesota Itzt. Hydraulics
C‘onvetltiotz, IAE-IR/ASCE, Minneapolis. Minnesota, EUA, pp. 4 15-423.
WOOD, I. R. (1983). “Uniform flow region of self-aerated flow”. ./. of Hyd. Div., ASCE,
Vol. 109, No. 3, pp. 447-461.
WOOD, I.R. (19S5). “Air water flows”. Keynote Address, Proc. 21st IAHR Cotlpess,
Melbourne, Australia, Vol 6, pp. 18-29

Copyright ASCE 2004 Water Resources 2000


Building Partnerships

Anda mungkin juga menyukai