Vous êtes sur la page 1sur 8


An experimental and numerical study on hydraulic

characteristics and theoretical equations of
circular weirs
Benyamin Naghavi, Kazem Esmaili, Jafar Yazdi, and Fatemeh Koorosh Vahid
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Oklahoma Libraries on 11/29/11

Abstract: Due to streamline curvature, non-hydrostatic pressure distribution, and nappe adherence on the weir wall, the dis-
charge coefficient of weirs is less than 1.0. However, in cylindrical weirs, the discharge coefficient extends to values more
than 1.0 due to the different pressures and velocity distributions. In this study, pressure and velocity distributions of different
circular weirs are simulated by computational fluid dynamics (CFD) and Fluent software and compared with experimental
results. The numerical results of CFD show a significant correlation with measured data from manometers and a laser Dop-
pler velocimeter (LDV). Experimental observations show that the critical flow depth forms before and after the crest for
greater and smaller discharge coefficients, respectively. Nappe separation depends on overflow discharge and will shift to
the downstream face of the cylinder in high discharges. To recognize the location of critical flow and nappe separation, ana-
lytical formulations are proposed depending on weir size and inflow conditions. The results show good agreement between
the analytical predictions and experimental observations.
Key words: laboratory model, CFD, critical depth location.
Résumé : En raison de la courbe dans la ligne d’écoulement, de la distribution non hydrostatique de la pression et de l’ad-
hérence de la lame déversante sur le mur du déversoir, le coefficient de débit des déversoirs est inférieur à 1,0. Toutefois,
For personal use only.

dans les déversoirs cylindriques, le coefficient de débit atteint des valeurs supérieures à 1,0 en raison des pressions et des ré-
partitions des vitesses différentes. La présente étude simule les répartitions des pressions et des vitesses de différents déver-
soirs circulaires en utilisant Fluent, un logiciel de dynamique numérique des fluides, et les compare aux résultats
expérimentaux. Les résultats numériques de la dynamique numérique des fluides indiquent une corrélation importante avec
les données mesurées à partir de manomètres et d’un vélocimètre Doppler à laser. Les constatations des expériences mon-
trent que la plus profondeur critique d’écoulement se forme respectivement avant et après la crête pour des profondeurs cri-
tique respectivement supérieurs et inférieurs. La séparation de la lame d’eau dépend de la surverse et se déplacera vers la
face aval du cylindre lors de grands débits. Afin de reconnaître l’emplacement de l’écoulement critique et de la séparation
de la lame d’eau, des formules analytiques sont proposées selon la dimension du déversoir et les conditions du débit entrant.
Les résultats montrent une bonne corrélation entre les prévisions analytiques et les observations expérimentales.
Mots‐clés : modèle de laboratoire, dynamique numérique des fluides, emplacement de la profondeur critique.
[Traduit par la Rédaction]

1. Introduction et al. (1992), Ramamurthy and Vo (1993a, 1993b), Chanson

and Montes (1998), Chanson (2006), and Heidarpour and
Circular-crested or cylindrical weirs are one of the oldest Chamani (2006), show that the discharge coefficient, Cd, is
hydraulic structures for discharge measurement. The advan- primarily a function of upstream head to crest radius ratio,
tages of the circular weirs include their stable overflow pat- and is close to and usually greater than 1.0. Studies on
tern compared to sharp-crested weirs, their ease in passing nappe suction and nappe ventilation show the nappe suction
floating debris, their simplicity of design compared to the prevents flow separation, which subsequently raises coeffi-
ogee crest design, and their associated lower cost (Chanson cients by 15% to 20% (Escande and Sananes 1959). As
and Montes 1997). These advantages lead to abundant appli- water passes over the dam crest, the nappe free-surface re-
cation of circular crested weirs for water level control in irri- mains smooth and clear, and the falling nappe adheres to
gation systems (Heidarpour et al. 2008). Much research on the weir face. This phenomenon (i.e., the two-dimensional
circular weirs, including research by Verwoerd (1941), Sar- nappe flowing past the convex curved wall and adhering to
ginson (1972), Escande and Sananes (1959), Cassidy the wall) is called a Coanda effect. The resulting Coanda
(1965), Hager (1985, 1993, 1995), Bos (1989), Ramamurthy force, induced by convex curvature of the wall and suction
Received 2 February 2011. Revision accepted 17 September 2011. Published at www.nrcresearchpress.com/cjce on XX November 2011.
B. Naghavi, K. Esmaili, and F.K. Vahid. Department of Water Engineering, Ferdowsi University of Mashhad, Iran.
J. Yazdi. Water Research Institute, Tehran, Iran.
Corresponding author: Benyamin Naghavi (e-mail: ben.naghavi@gmail.com).
Written discussion of this article is welcomed and will be received by the Editor until 30 April 2012.

Can. J. Civ. Eng. 38: 1327–1334 (2011) doi:10.1139/L11-092 Published by NRC Research Press
1328 Can. J. Civ. Eng. Vol. 38, 2011

pressure, acts on the wall surface in the same direction as the Fig. 1. Manometers joint on weir at angles 0° to 210°.
flow, which induces the nappe adherence (Chanson and
Montes 1997). Suction pressure and the resulting nappe ad-
herence cause the streamlines to become more curved and
the flow velocity to increase. Therefore, the discharge coeffi-
cient of a circular weir is greater than that of sharp and
broad-crested weirs. To predict the wall suction pressure,
Chanson (1996) proposed an analytical method in which a
pressure correction coefficient accounts for the non-hydro-
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Oklahoma Libraries on 11/29/11

static pressure distribution on the crest. The literature review

shows that the pressure distribution on a cylindrical weir dif-
fers from conventional weirs, and by increasing the upstream
head to crest radius ratio to more than 0.8, the pressure dis- To investigate the pressure distribution on the weir, eight
tribution on a cylindrical weir changes to non-hydrostatic manometers were installed on the weir wall at multiples of
(Bos 1989; Ramamurthy and Vo 1993a; Heidarpour et al. 30 degrees from 0° to 210° (Fig. 1). The water surface pro-
2008). The dimensionless velocity distribution is only a files were measured by a point gauge in 3 mm distances over
function of upstream flow to the radius of a circular weir, the weirs from upstream to downstream. The experimental er-
while the maximum crest velocity, the normalized crest pres- ror due to the apparatuses, i.e., point gauge and manometers,
sure, and the pressure correction coefficient are functions of were ±0.1 mm and ±0.1 cm, respectively.
the average upstream velocity and a velocity correction fac- To measure the horizontal and vertical components of ve-
tor (Heidarpour et al. 2008). Ramamurthy and Vo (1993b) locity, a 2D LDV was used.
used the expression of Dressler (1978) for the velocity pro-
file of a shallow irrotational flow over a curved bed to pre-
dict the velocity distribution over a cylindrical weir. 3. Numerical model
Heidarpour and Chamani (2006) present a method to predict Fluent was used to simulate weir flow. To measure the free
the velocity distribution over the crest of the cylindrical weir surface profile, the Hirt-Nichols’ VOF (volume of fluid) and
For personal use only.

as well as the discharge coefficient of the cylindrical weir, Youngs’ method are used (Hirt and Nichols 1981; Youngs
using potential flow past a cylindrical obstacle. Assuming 1982; Tadayon and Ramamurthy 2009). In modeling the free
an inviscid flow, the maximum velocity occurs at the edge surface between water and air, a transport equation can be
of the thin boundary layer and closer to the crest (Heidar- considered for the water phase:
pour et al. 2008).
In this research, computational fluid dynamics (CFD) is ½1 ðaw Þ þ rðaw vÞ ¼ 0
used to simulate the water surface profiles and the pressure @t
and velocity distributions on the cylindrical weirs. To vali- where aw is the volume fraction of water and v is flow velo-
date the fluid velocity field predicted by the CFD software, city. In this case, as the volume fraction for other phases can
a laser Doppler velocimeter (LDV) is used. Using an LDV, be inferred from the constraint, the transport equation should
which does not interfere with the flow, enables us to obtain only be solved:
localized flow measurements and record a wide range of ve-
locities. The CFD model’s results respecting the pressure dis- ½2 aa ¼ 1  aw
tribution are compared with experimental data from the
manometers and a new interpretation of pressure distribution This equation is solved in the entire domain and volume
on cylindrical weirs is obtained. Finally, methods for finding fraction computed for all cells. For each cell, aw = 1 when it
is full of water and aw = 0 when it is full of air. For cells
the location of critical depth and nappe separation are given,
that span the interface between air and water, aw is between
and some analytical equations are introduced.
0 and 1. The fluid properties in each cell are determined ac-
cording to the local volume fraction. For example, the density
2. Experimental works in each cell is as follows:
The experiments were performed on a rectangular flume of ½3 r ¼ aw rw þ ð1  aw Þra
10 m in length, 0.3 m in width, and 0.5 m in depth. The
overflow characteristics of cylindrical weirs were investigated where rw and ra are water and air densities, respectively.
for several configurations, i.e., six cylinder sizes with radius In this research, both structured and unstructured meshes
of 0.03 < R < 0.08 m, in supported and non-supported con- were used. Because of the complex geometry of circular
ditions. The height of weir supports were in a range of weirs, the area around the weir was meshed in unstructured
0.025 < P < 0.1 m. All cylinders were made of PVC pipes. form, and elsewhere the structured mesh was used. To define
In each run, and for all different discharges, the water surface water and air flow into the domain, two different inlets
profiles and the pressure and velocity distributions were should be used as boundary conditions for upstream flow.
measured. In all runs, the longitudinal slope of the flume To estimate the wall’s effect on flow, empirical wall func-
was constant and equal to 0.005. To simulate the flow in the tions known as standard wall functions were used. For strong
flume, a flow circulation system was set up, and discharges curvatures like spillways, applying the RNG k–3 model is
were measured with a prior calibrated rectangular sharp more appropriate (Tadayon and Ramamurthy 2009). Thus
crested weir (Hosseini 2003). the RNG k–3 turbulence model was used with standard wall

Published by NRC Research Press

Naghavi et al. 1329

functions (Yakhot and Orszag 1986, Yakhot and Smith Fig. 2. Comparison between experimental (Exp) and numerical
1992). Accordingly, for the upper boundary above the air (Num) water surface profiles in different discharges.
phase, symmetry conditions were considered, which enforce
a zero normal velocity and a zero shear stress. For down-
stream flow, gradients of all variables were assumed to be
zero, and since the outflow of the flume was assumed to be
a jet, pressure was considered to be zero, i.e., atmospheric
To complete the CFD modeling, the PRESTO pressure
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Oklahoma Libraries on 11/29/11

discretization, which showed the best convergence, was used.

The flow was simulated to be unsteady. Thus, the pressure–
velocity coupling scheme was achieved using the pressure-
implicit with splitting of operators (PISO) algorithm (Issa
Calculating an unsteady free surface profile requires small Fig. 3. Comparison between experimental (Exp) and numerical
initial time steps. Therefore, a time step ranging from 0.01 to (Num) pressure head in different discharges.
0.001 was selected. During calculations and simulations, the
convergence of the solution and the water surface profiles
was monitored. The convergence occurred whenever the nor-
malized residuals of each variable were equal to or less than
0.001. In this paper the free surface is defined by VOF = 0.5,
which is a common value for this purpose (Yazdi et al.
Computations were conducted by a Pentium IV system
with an AMD 3800+ processor. Each run of the numerical
model took 1 to 2 h, depending on model dimensions, mesh
For personal use only.

numbers, and flow characteristics.

4. Results
Experimental results related to water surface profiles, pres- and the sub-atmospheric pressure increases the velocity. The
sure distributions, and streamwise velocity distributions for differences between the experimental and numeical pressure
flow over the circular crested weir were used to validate the distributions are due to severe pressure gradients on the cir-
numerical simulation predictions. For water surface profiles cular face. These pressure gradients cannot be predicted pre-
over the crest, Fig. 2 shows that there is a good agreement cisely by the K–3 turbulence model using standard wall
between CFD predictions and experimental results, and the functions (Tadayon and Ramamurthy 2009; Yazdi et al.
maximum difference is about 2.5%, which is mainly due to 2010). Therefore, a mean error of about 10% is seen in the
measurement inaccuracy in the experimental model and small results.
variations in the numerical model. The CFD model indicated The measured and predicted fluid velocity field is plotted
that the streamlines near the weir surface adhere to the weir in a contour form shown in Figs. 4 and 5, for horizontal and
wall, and thus the water surface profile follows the curvature vertical components, respectively. In this case, the average
shape of the weir. In the experimental model, the water sur- velocity of water is approximately 0.65 m/s. According to
face fluctuations are negligible before the weir crest, and the Fig. 4, a good agreement between the CFD simulation results
water surface slope is roughly constant. However, after the and the LDV data can be seen for the horizontal fluid veloc-
crest, depending on discharge values and water adherence to ity (0 < Vx < 1.5). The main differences are in high veloc-
the weir surface, the water surface slope changes. ities near the crest, where the CFD simulation is predicted
Both experimental and predicted pressure head distribu- higher than LDV data. In all points close to the weir surface
tions over the circular weir are shown in Fig. 3. For all dis- Vx = 0; in upper layers and towards the downstream flow, Vx
charge rates, the pressure head decreases with increasing is gradually increased, and in the region between 90° and
angle. However, the curves corresponding to all discharge 120°, it reaches maximum value. After this region, a decreas-
rates cross at an angle of 70°. That means that the pressure ing trend is formed.
head at an angle of 70° is independent of discharge rate. Be- For the vertical fluid velocity (Vy), the simulation also pro-
fore this region, higher discharges show higher pressure vides accurate predictions in most areas, especially in the up-
heads, and after that, the reverse is true, so that the lowest stream flow. Vy lies in the range between 0 and 0.8; on the
discharges show the highest pressure heads. Like the physical weir wall Vy = 0, and in upper layers it is increased. More-
model, the numerical model shows that the lowest pressure over, on the weir crest and in all upper layers of this region,
occurs at point 120° for the highest discharge, and as the dis- Vy = 0, while upstream and downstream of it, Vy extends to
charges decrease, the lowest pressure point shifts towards the maximum values (the region between 30° and 60° and also
downstream flow. The minimum pressure occurs in a region 120° and 150°, symmetrically). However, the upstream flow
after the crest, where the velocity head is at maximum. In is the inverse of the downstream; the vertical component of
this region, because of weir curvature, the pressure drops the velocity vector is in a positive direction.

Published by NRC Research Press

1330 Can. J. Civ. Eng. Vol. 38, 2011

Fig. 4. Horizontal fluid velocity, (a) CFD result vs. (b) LDV data.
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Oklahoma Libraries on 11/29/11
For personal use only.

5. Theoretical equations in cylindrical weirs generalized the Jaeger theory, assuming a linear law for the
radius of curvature of the streamlines along a normal depth,
5.1. Critical depth location and developed some potential flow equations to measure the
The critical flow principle is a useful approach for the hy- critical flow depth. He imposed a minimum specific energy
draulic analysis of round-crested weirs due to their single head for a given discharge to generalize the critical flow con-
head discharge relationships (Castro-Orgaz 2008). At the crit- dition for curved channel flows, and proposed a model based
ical point, the effects of the streamline curvature and slope on critical flow conditions developed by Hager (1985) to pre-
are significant because the velocity distribution is non-uni- dict the critical depth at the weir crest. According to Castro-
form and the pressure distribution is non-hydrostatic. These Orgaz (2008), in circular crested weirs, the flow is critical at
effects must be taken into account to obtain accurate results the crest (Fr = 1) for E/R around 0.5–0.6, where E is the spe-
(Hager 1985; Chanson 2006; Castro-Orgaz et al. 2008a). cific energy. For larger heads, the flow at the weir crest is
To calculate the critical flow over a round crested weir, supercritical, and for smaller heads, the flow at the crest is
different models are proposed. Castro-Orgaz et al. (2008b) subcritical. Chanson (2006) used the continuity and Bernoulli

Published by NRC Research Press

Naghavi et al. 1331

Fig. 5. Vertical fluid velocity, (a) CFD result vs. (b) LDV data.
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Oklahoma Libraries on 11/29/11
For personal use only.

equations in terms of the critical flow depth and depth-aver- According to water surface profiles, the water surface
aged velocity, and proposed a new analytical expression of slopes around points 60° and 90° are the same and thus
the critical flow depth for ideal-fluid flows. He implied that
AB d
discharge coefficients larger than 1.0 may be obtained only ½4 tan a ¼ ¼ ) d ¼ AB  cos q
when the crest pressure distribution is less than hydrostatic. R Rcos q
Using the proposed methods for calculating the critical where AB = Hw – dcrest, Hw is the flow depth upstream, and d
depth in curved flow, a new method is introduced herein to is as follows:
identify the location of critical depth in a circular crested
weir. To determine the location of critical depth, a geometric ½5 d ¼ ðdc  dcrest Þ  Rð1  sin qÞ
relation is proposed herein. In Fig. 6, the lines CE and HI
represent the water depths at the crest (dcrest) and the critical Considering the equality of eq. [4] with eq. [5] and using
depth (dc), respectively. (dc – dcrest) = U, the following equation can be obtained:

Published by NRC Research Press

1332 Can. J. Civ. Eng. Vol. 38, 2011

Fig. 6. Geometrical parameters for critical depth and separation location.

Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Oklahoma Libraries on 11/29/11

Fig. 7. Critical depth location in different Cd. Fig. 8. Variation of Cd with dcrest/dc.
For personal use only.

over the cylindrical weirs in which the highest point is near-

½6 U  R ¼ AB  cos q  R  sin q est to the crest.
Moreover, the variation of Cd with the ratio of dcrest to dc
According to trigonometric functions and applying x = tan reported in Fig. 8, indicates an increase of dcrest/dc with de-
(q/2): creasing Cd. The higher Cd forms before the point 0.9 of
dcrest/dc. After this point, there are two decreasing branches
2x 1  x2 of Cd in which the higher one relates to 2 < Hw/R < 8 and
½7 sin q ¼ ; cos q ¼
1 þ x2 1 þ x2 the lower one relates to Hw/R < 2.
Comparison between Fig. 7 and Fig. 8 indicates that when
Accordingly, eq. [6] can change to a nonlinear equation. dcrest/dc are less and more than 0.9, the critical depths are lo-
The resulting equation for calculating the critical depth loca- cated before and after the crest, respectively. Higher values of
tion is as follows: Cd near the crest are the ones with lower dcrest/dc. The steady
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi! branch around point 120° in Fig. 7 expresses the lower
1 R þ ðABÞ2  U 2 þ 2UR
½8 q ¼ 2  tan branch in Fig. 8, hence for Hw/R < 2, the critical depth loca-
AB þ U  R tion is stable.

Figure 7 shows the variation of critical depth location q 5.2. Separation location
with discharge coefficient Cd. In this figure, two different To locate the separation point, an imaging technique is
trends are distinguishable, in regions before and after the used. For the sake of simplicity, we shall consider the anal-
crest, respectively. It is seen that the discharge coefficient is ogy with the overflow above a two-dimensional circular weir
always more than 1.0. Higher values of Cd form before the (Fig. 6). According to the images, it is seen that when the
crest (1.4 < Cd <1.6) and the highest ones are closest to the flow passes towards downstream, it separates because of the
crest. After the crest, again higher values of Cd are closer to flow inertia and weir curvature. Results of this study show
the crest, and these values decrease downstream of the crest the pressure decreases as the flow passes towards down-
(1.2 < Cd <1.4). Overall, a L trend forms for Cd variations stream. For the range of experimental discharges investigated

Published by NRC Research Press

Naghavi et al. 1333

Fig. 9. Separation location for different dcrest/R ratios.

Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Oklahoma Libraries on 11/29/11

herein, the minimum pressure occurs at point 120° for higher In experimental works, it is seen that in the region before
discharges, and for average and lower discharges, it occurs the crest, higher discharges show higher pressure heads
around points 150° and 180°, respectively. (Fig. 3), and thus according to eq. [10], Lcrest decrease leads
Suction pressure at the weir wall causes nappe adherence, to Cd increase. As it was mentioned before, when the dis-
and thus the pressure head decreases towards the downstream charge increases, the suction pressure gets closer to the crest.
flow (Chanson and Montes 1998). However, pressure in- Hence, dcrest decreases and the critical depth forms near the
creases again after the nappe separation, and as the discharge crest. For lower discharges, the effect of suction pressure on
decreases, the nappe separation location shifts towards down- dcrest is lower, and dc forms far from the crest. However, the
stream and farther away from the crest. Basic ideal-fluid flow pressure head is high enough that Cd > 1.4. For lower pres-
calculations provide an estimate of the wall pressure and sure heads in which Cd < 1.4, the value of dcrest/dc is greater
nappe separation location induced by cavitation on the down- than 0.9, and dc forms downstream of the crest. In this situa-
stream face of inflatable flexible membrane dams (Chanson tion, dcrest is nearly close to dc, and the water surface profile
1997): is uniform. Referring to Fig. 3, lower discharges show higher
For personal use only.

  pressure heads after the crest. Thus, dc moves downstream.

1 DP 2  dcrest However, for Cd < 1.2, in which Hw/R < 2, the critical depth
½9 cos 4sep ¼ 1   
sC P R is located around the point 120°. It seems that in this range
the head pressure effect is negligible.
where 4sep is separation location, sc is a critical cavitation
number at which cavitation starts to appear in the flow which
is typically about 0.5 (Knapp et al. 1970), DP = Patm – PV,
7. Conclusions
Patm is the atmospheric pressure, PV = Ps is the absolute The CFD model is used to predict the hydraulic character-
pressure at the cylinder surface, P = rw × g × R, rw is the istics of cylindrical weirs. Numerical simulations were in
water density, and g is the gravity acceleration. Other para- good agreement with experimental data, which confirmed
meters have been defined previously. The angle 4 is equal to the CFD capability in prediction of flow conditions on struc-
zero at the crest (Fig. 6). Figure 9 shows the variation of DP/ tures such as rubber dams. Moreover, some theoretical for-
P on the weir for different dcrest/R (sC = 0.5). The line DP/ mulations were derived to determine the critical depths and
P = 1 represents the atmospheric pressure where the nappe separation locations in cylindrical weirs. According to these
separation occurs. According to this figure, as dcrest/R in- equations, the critical depth points are a function of weir ra-
creases, the nappe separation location approaches the crest. dius and upstream flow depth. As the flow passes over the
In other words, for a given radius size, as the discharge in- weir, the pressures always decrease towards the downstream
creases, the separation point gets closer to the crest. In lower flow. However, depending on discharge values and after
discharges, there is a suction pressure around the point 210°, nappe separation, the pressure head increases. According to
and past this point, the atmospheric pressure leads to separa- analytical results, nappe separation usually occurs after the
tion. crest, and approaches it in higher discharges.

6. Discussion
Bos, M.G. 1989. Discharge measurement structures, 3rd ed. Institute
To calculate the dimensionless critical depth in non- of Land Reclamation and Improvement ILRI, Wageningen,
hydrostatic pressure condition, Chanson (2006) gives an Netherlands.
expression as a function of the pressure correction coeffi- Cassidy, J.J. 1965. Irrotational flow over spillways of finite height.
cient Lcrest, momentum correction coefficient bcrest, and Journal of the Engineering Mechanics Division, 91(6): 155–173.
discharge coefficient Cd as follows: Castro-Orgaz, O. 2008. Curvilinear flow over round-crested weirs.
Journal of Hydraulic Research IAHR, 46(4): 543–547. doi:10.
½10 bcrest  Cd2  L2crest  1 3826/jhr.2008.3251.
Castro-Orgaz, O., Giraldez, J.V., and Ayuso, J.L. 2008a. Higher order
As bcrest is equal to or larger than 1.0, eq. [10] implies that critical flow condition in curved streamline flow. Journal of
discharge coefficients larger than 1.0 may be obtained only Hydraulic Research IAHR, 46(6): 849–853. doi:10.1080/
when the crest pressure distribution is less than hydrostatic 00221686.2008.9521931.
(Chanson 2006). Castro-Orgaz, O., Giraldez, J.V., and Ayuso, J.L. 2008b. Critical flow

Published by NRC Research Press

1334 Can. J. Civ. Eng. Vol. 38, 2011

over circular crested weirs. Journal of Hydraulic Engineering, 118(6): 988–994. doi:10.1061/(ASCE)0733-9437(1992)118:6
134(11): 1661–1664. doi:10.1061/(ASCE)0733-9429(2008) (988).
134:11(1661). Sarginson, E.J. 1972. The influence of surface tension on weir flow.
Chanson, H. 1996. Some hydraulic aspects during overflow above Journal of Hydraulic Research, 10(4): 431–446. doi:10.1080/
inflatable flexible memberane dam. Queensland (AU), Department 00221687209500034.
of Civil Engineering at the University of Queensland. Report No. Tadayon, R., and Ramamurthy, A.S. 2009. Turbulence modeling of
CH47/96. pp. 1–60. flows over circular spillways. Journal of Irrigation and Drainage
Chanson, H. 1997. A review of the overflow of inflatable flexible Engineering, 135(4): 493–498. doi:10.1061/(ASCE)IR.1943-
membrane dams. Journal of Australian Civil/Structure Engineer- 4774.0000012.
ing Transactions, 39(2–3): 107–116. Verwoerd, A.L. 1941. Calcul du débit de déversoirs dénoyés et noyés
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Oklahoma Libraries on 11/29/11

Chanson, H. 2006. Minimum specific energy and critical flow à crête arrondie. L’Ingénieur des Indes Néerlandaises, 8: 65–78.
conditions in open channels. Journal of Irrigation and Drainage [In French.]
Engineering, 132(5): 498–502. doi:10.1061/(ASCE)0733-9437 Yakhot, V., V, and Orszag, S.A. 1986. Renormalization-group
(2006)132:5(498). analysis of turbulence. Physical Review Letters, 57(14): 1722–
Chanson, H., and Montes, J.S. 1997. Overflow characteristics of 1724. doi:10.1103/PhysRevLett.57.1722. PMID:10033528.
circular weirs. Queensland (AU): Department of Civil Engineering Yakhot, V., and Smith, L.M. 1992. The renormalization group, the 3-
at the University of Queensland. Report No. CE154. pp. 1–98. expansion and derivation of turbulence models. Journal of
Chanson, H., and Montes, J.S. 1998. Overflow characteristics of Scientific Computing, 7(1): 35–61. doi:10.1007/BF01060210.
circular weir. Journal of Irrigation and Drainage Engineering, Yazdi, J., Sarkardeh, H., Azamathulla, H., and Ghani, A.A. 2010. 3D
124(3): 152–162. doi:10.1061/(ASCE)0733-9437(1998)124:3 simulation of flow around a single spur dike with free-surface
(152). flow. International Journal of River Basin Management, 8(1): 55–
Dressler, R.F. 1978. New nonlinear shallow flow equations with 62. doi:10.1080/15715121003715107.
curvature. Journal of Hydraulic Research, 16(3): 205–222. doi:10. Youngs, D.L. 1982. Time-dependent multi-material flow with large
1080/00221687809499617. fluid distortion. In Numerical methods for fluid dynamics. Edited
Escande, L., and Sananes, F. 1959. Etude des seuils déversants à fente by K.W. Morton and M.J. Baines. Academic Press, London.
aspiratrice. Houille Blanche, 14: 892–902. doi:10.1051/lhb/ pp. 273–285.
1959005. [In French.]
For personal use only.

Hager, W.H. 1985. Critical flow condition in open channel List of symbols
hydraulics. Acta Mechanica, 54(3-4): 157–179. doi:10.1007/
AB differences between upstream flow and water depths at
the weir crest, m
Hager, W.H. 1993. Abfluss über zylinderwehr. Wasser und Boden, Cd discharge coefficient
44(1): 9–14. [In German.] dc critical flow depth, m
Hager, W.H. 1995. Discussion to characteristics of circular crested dcrest water depths at the weir crest, m
weirs. Journal of Hydraulic Engineering, 120(12): 1494–1495. i specific energy, m
doi:10.1061/(ASCE)0733-9429(1994)120:12(1494). Fr Froude number
Heidarpour, M., and Chamani, M.R. 2006. Velocity distribution over g gravity acceleration, m/s2
cylindrical weirs. Journal of Hydraulic Research IAHR, 44(5): Hp pressure head, m
708–711. doi:10.1080/00221686.2006.9521719. Hw flow depth in the upstream of the weir crest, m
Heidarpour, M., Habili, J.M., and Haghiabi, A.H. 2008. Application P pressure head on the weir, Pa
of potential flow to circular-crested weir. Journal of Hydraulic Patm atmospheric pressure, Pa
Research IAHR, 46(5): 699–702. doi:10.3826/jhr.2008.3181. PV absolute pressure at the cylinder surface, Pa
Hirt, C.W., and Nichols, B.D. 1981. Volume of fluid (VOF) method R radius of cylindrical weir, m
for the dynamics of free boundaries. Journal of Computational U differences between critical depth and water depths at
Physics, 39(1): 201–225. doi:10.1016/0021-9991(81)90145-5. the weir crest, m
Hosseini, M. 2003. Investigation of efficiency and discharge Vx horizontal fluid velocity, m/s
coefficient of culverts, M.Sc. thesis, Ferdowsi University of Vy vertical fluid velocity, m/s
Mashhad, Iran. v flow velocity, m/s
aa volume fraction of air
Issa, R.I. 1986. Solution of the implicitly discretised fluid flow
aw volume fraction of water
equations by operator-splitting. Journal of Computational Physics,
bcrest momentum correction coefficient
62(1): 40–65. doi:10.1016/0021-9991(86)90099-9. DP differences between atmospheric and absolute pressures,
Knapp, R.T., Daily, J.W., and Hammit, F.G. 1970. Cavitation. Pa
McGraw-Hill Book Company, New York, N.Y. d flow depth in the critical depth location, m
Ramamurthy, A.S., and Vo, N.D. 1993a. Characteristics of circular- q critical depth location
crested weirs. Journal of Hydraulic Engineering, 119(9): 1055– Lcrest pressure correction coefficient
1062. doi:10.1061/(ASCE)0733-9429(1993)119:9(1055). r density of each cell in CFD model
Ramamurthy, A.S., and Vo, N.D. 1993b. Application of dressler ra air density, kg/m3
theory to weir flow. Journal of Applied Mechanics, 60(1): 163– rw water density, kg/m3
166. doi:10.1115/1.2900740. sC critical cavitation number (= 0.5)
Ramamurthy, A.S., Vo, N.D., and Vera, G. 1992. Momentum model 4sep separation location
of flow past weir. Journal of Irrigation and Drainage Engineering,

Published by NRC Research Press

Vous aimerez peut-être aussi