Irrig Sci (2010) 28:173–179

DOI 10.1007/s00271-009-0172-1

ORIGINAL PAPER

Flow over rectangular sharp-crested weirs

Sara Bagheri Æ Manouchehr Heidarpour

Received: 4 March 2009 / Accepted: 27 July 2009 / Published online: 13 August 2009
Ó Springer-Verlag 2009

Abstract Sharp-crested weirs are the simplest form of q Discharge per unit width (m2/s)
over-flow spillway that commonly used to determine the Q Weir discharge (m3/s)
flow rate in hydraulic laboratories, industry and irrigation r Radius of the streamline at a given point (m)
systems, where highly accurate discharge measurements Rb Radius of streamline curvature at lower nappe of
are needed. In this study, the experimental upper and lower profile in the distance of OB (m)
nappe profiles in rectangular sharp-crested weirs are fitted Rs Radius of streamline curvature at upper nappe of
by quadratic and cubic equations, respectively. In addition, profile in the distance of OB (m)
free-vortex theory is used to simulate flow over this kind of V Velocity of the streamline at a given point (m/s)
weirs and determine discharge coefficient. Physical models V0 Upstream velocity (m/s)
of sharp-crested weirs with various widths and heights Vb Lower nappe velocity at the section of maximum
were considered. The proposed method agrees well with elevation of lower nappe (m/s)
the experimental observations. Also, the experimental data x Cartesian coordinate in direction of flow (m)
indicate that the suggested equation presents reasonable y Cartesian coordinate perpendicular to flow (m)
results for the range of 0 \ h/P \ 9. Y Flow depth at the section of maximum elevation of
lower nappe (m)
List of symbols
B Channel width (m)
b Weir width (m)
C Constant value in free-vortex theory (m2/s)
Cd Discharge coefficient of the weir (-) Introduction
g Ratio of weight to mass (m/s2)
h Upstream head (m) Flow measurement structure is generally designed to act as
H Total head upstream of the weir (m) a control in the channel to provide a unique relationship
k Correction coefficient (-) between the upstream head and the discharge. A weir is
L Thickness of weir crest (m) one of the most important structures that provide for this
P Weir height (m) purpose. The different types of weirs are defined as broad-
crested weir, short-crested weir and sharp-crested weir.
Classified under the term sharp-crested or thin-plate weirs
Communicated by J. Ayars. are those overflow structures whose upstream head (h) to
length of crest in the direction of flow (L) is greater than
S. Bagheri (&)
M. Heidarpour
about h/L [ 15 (Bos 1989). In general, sharp-crested weirs
Department of Water Engineering, Agricultural College,
Isfahan University of Technology, 84156-83111 Isfahan, Iran are used in hydraulic laboratories, industry and irrigation
e-mail: [email protected] pilot schemes where highly accurate discharge measure-
M. Heidarpour ments are required. Rectangular sharp-crested weirs are the
e-mail: [email protected] simplest forms of weirs that comprise the following two

123
174 Irrig Sci (2010) 28:173–179

Fig. 1 End view of rectangular B

B=b
sharp-crested weirs; a
suppressed and b contracted b

h h

P
P

(a) (b)

types, suppressed rectangular sharp-crested weirs and  0:7  1:46

0:611 þ 2:23 Bb  1 0:075  0:011 Bb  1
contracted rectangular sharp-crested weirs (Fig. 1). Cd ¼  0:7 þ  1:46
The conventional rectangular-notch, sharp-crested weir 1 þ 3:8 Bb  1 1 þ 4:8 Bb  1
equation is considered in the form: h
 ð4Þ
pffiffiffiffiffi P
Q ¼ 2=3Cd 2g  bh1:5 ð1Þ
Ramamurthy et al. (1987) applied the momentum principle
in which Q, the weir discharge; b, the weir width; h, the to estimate discharge coefficient for sharp-crested weirs
upstream head; g, gravitational acceleration and Cd, dis- and derived pressure and momentum coefficients
charge coefficient. Francis indicated that for suppressed experimentally. Swamee (1988) by combining Eqs. 2 and
rectangular sharp-crested weirs the effect of the lateral 3 and fitting the experimental data of Kandaswamy and
contractions could be considered by reducing b in Eq. 1 by Rouse (1957) represented the following full-range equation:
0.2h (Brater et al. 1996).    15 !0:1
Many studies have been done to determine the discharge 14:14P 10 h
Cd ¼ 1:06 þ ð5Þ
coefficient of sharp-crested weirs. Rehbock (1929) pro- 8:15P þ h hþP
posed the following equation for Cd without considering
viscous and capillary influence: Johnson (2000) studied flat-topped and sharp-crested weirs.
He proposed the relation between Cd and H/P in which H is
h
Cd ¼ 0:611 þ 0:08 : ð2Þ the total head upstream of the weir. Because of using the
P
velocity head in this method, a trial and error solution is
Equation 2 is valid for h/P B 5 in which P is the weir required to calculate Cd.
height. If the weir height is very high, the discharge The objective of the study was to determine a discharge
coefficient tends to p/(p ? 2) = 0.611. This value was coefficient for a sharp-crested weir based on the upper and
proposed by Kirchoff for contraction coefficient of a jet lower nappe profiles measured on a sharp-crested weir and
issuing without energy loss, and with negligible deflection on free-vortex theory. Experimental data sets and data from
by gravity, from a long rectangular slot in a large tank the literature were used to verify the resulting equation.
(Henderson 1966). Rouse (1936) proposed the following
equation for discharge coefficient in the range of h/P C 15:
  Theoretical consideration
P 1:5
Cd ¼ 1:06 1 þ ð3Þ
h
For a sharp-crested weir, the concept of critical flow is not
Kindsvater and Carter (1957) presented a solution for the applicable (Bos 1989). For the derivation of the head-dis-
weir discharge characteristics based on the experimental charge equation, it is assumed that sharp-crested weir
results and dimensional analysis. They provided a solution behaves similar to an orifice with a free-water surface.
for the discharge and a method of compensating for the Accordingly, the following assumptions are made:
influence of viscosity and surface tension. Strelkoff (1964)
1. the height of the water level above the weir crest is h,
solved the problem of two-dimensional irrotational flow
and there is no contraction;
over the sharp-crested weir using conformal mapping
2. mean flow velocity over the weir crest is almost
techniques, and presented numerical solutions that to relate
horizontal; and
the discharge coefficient with the values of h/P in the weir
3. the approaching velocity head V20/2 g is neglected.
range of 0 \ h/P \ 10. For a contracted weir with width of
b in a channel of width B, French (1986) obtained the Figure 2 illustrates the definition of sharp-crested weir
following equation: parameters.

123
Irrig Sci (2010) 28:173–179 175

Energy Line Table 1 Summary of model characteristics and test conditions

V 02
2g
E Model no. b (cm) b/B P (cm) Q (l/s) h/P
D Vs

h Y
1 32 1.000 10, 12, 15, 18 2.0–19.5 0.14–0.91
C Vb 2 24 0.750 8, 10, 14 1.8–15.3 0.17–0.89
O B F 3 20 0.625 8, 10, 14 1.6–12.4 0.23–0.92
A
Rb
4 16 0.500 8, 10, 14 1.8–8.5 0.25–1.00
P O' 5 12 0.375 8, 10, 14 1.3–6.4 0.36–1.03
6 8 0.250 8, 10, 14 0.8–4.1 0.35–1.04

1.5 mm
Fig. 2 Definition of sharp-crested weir parameters Flow
90º

Assuming free-vortex motion at the distance of OB 60º

Sharp
(White 2003): Edge
C
V¼ ð6Þ
r
in which V and r are the velocity and radius of curvature of
the streamline at a given point, respectively, and C is
constant. The discharge passing over a unit length of the Fig. 3 Side view of the weir crest
weir crest can be written as:
ZRs ZRs were constructed in a smooth rectangular flume 0.32-m
C Rs
q¼ Vdr ¼ dr ¼ C ln ð7Þ wide, 0.35-m high, and 7.0-m long. The flume sidewalls
r Rb
Rb Rb were made of Plexiglas, with a metal bed. All of the models
were ventilated with a pipe that was installed aside of the
in which Rb and Rs are the radius of streamline curvature at
flume. To avoid the area of water surface draw-down, the
lower and upper nappe of profile in the distance of OB,
approach flow depth was measured 1.0 m upstream of the
respectively. Assuming streamlines are concentric at the
weir crest. The upstream head was measured with a point
distance of OB, the equation can be written as:
gage of ±0.01 mm reading accuracy, but because of free
Rs ¼ Rb þ Y ð8Þ surface fluctuations, the depth could be read only to the
where Y is the depth at the section of maximum elevation nearest millimeter. Each weir model was tested using 7–12
of lower nappe. From Eq. 6, the constant value of C can be flow rates and a total of 152 experiments were conducted.
given as: The discharge over the weir was measured using an elec-
tromagnetic flow meter with an accuracy of ±0.5%.
C ¼ Vb R b ð9Þ Measurements of velocity profiles upstream of the models
Substituting Eqs. 8 and 9 in Eq. 7, q can be expressed as were carried out to make sure that the approaching flow
follows: was fully developed.
Rb þ Y
q ¼ Vb Rb ln ð10Þ
Rb
Results

Experimental setup and procedure Dimensionless plot of all the measured upper and lower
flow nappe profiles over sharp-crested rectangular weir for
To determine the equations of flow nappe profiles and the different value of b/B is shown in Fig. 4, where y is the
discharge coefficient for rectangular sharp-crested weirs, vertical depth of flow at the any x, distance downstream
six types of models were made using transparent Plexiglas from the weir. It was found that the upper and lower nappe
with a thickness of 10 mm and different crest heights and profiles could be approximated very closely by quadratic
widths. Table 1 shows a summary of the model charac- and cubic equations, respectively. Using upper and lower
teristics and test conditions. The downstream edge of the flow profile’s equations the distances of OA, OB, AC and
notch was beveled as shown in Fig. 3. The weir models AE were obtained for each value of b/B. Considering these

123
176 Irrig Sci (2010) 28:173–179

Fig. 4 Upper and lower nappe

profiles over a sharp-crested
rectangular-notch weir

 
parameters, the mean radius of curvature of the streamline b
Y ¼ 0:839h exp 0:24 ð12Þ
along the distance of OB at lower nappe profiles (Rb) and B
flow depth at the section of maximum elevation of lower
nappe (Y) were found. More details were explained in
Substituting Eqs. 11 and 12 in Eq. 10:
Appendix. The analysis of the results showed that the non- !
 
dimensional mean radius of the streamline along the dis- b 3:884
tance of OB at the lower nappe profiles (Rb/h) and non- q ¼ Vb  0:216h exp 0:94 ln 1 þ
B expð1:18 Bb Þ
dimensional flow depth at the section of the maximum
ð13Þ
elevation of lower nappe (Y/h) have exponential relations
pffiffiffiffiffiffiffiffi
with the ratio of weir width to channel width (b/B) are where Vb ¼ 2gh:
represented in Fig. 5. Likewise, the following equations However, in a real flow because of gravity effects and
were found: head loss, the proposed method, which is based on the
  concentricity of streamline at the distance of OB begins to
b
Rb ¼ 0:216h exp 0:94 ð11Þ be violated. The concept of the concentricity and non-
B

123
Irrig Sci (2010) 28:173–179 177

 
h
k ¼ 0:188 þ 0:937 ð17Þ
P
Using Eq. 17, Eq. 16 can be rewritten as:
    !
b 0:73 Ph þ 3:64
Cd ¼ 0:324 exp 0:94 ln 1 þ   ð18Þ
B exp 1:18 Bb

Figure 8 presents variation of the calculated Cd with

Eq. 18 against Cd obtained from direct discharge
measurements. According to this figure, the error for a
test range of 0.14 B h/P B 1.04 is about ±3%.
The coefficient Cd in the proposed equation (Eq. 18) is
defined as a function of ratios h/P and b/B. For b/B = 1
Kandaswamy and Rouse (1957) remarked that Cd increases
up to h/P = 10 and then decreases with a further raise in h/
P. Figure 9 shows the variation of Cd with h/P for full-
Fig. 5 Variation of Rb/h and Y/h with b/B width weirs. The solid-line in this figure denotes a fitted
curve based on the calculated Cd with Eq. 18. Figure 9 also
concentricity of the streamlines has been represented in indicates a comparison of the results of the present model
Fig. 6. Defining a correction coefficient accounting for with the works of Kandaswamy and Rouse (1957),
non-concentricity streamline, Eq. 8 can be rewritten as: Strelkoff (1964) and Ramamurthy et al. (1987). From this
Rs ¼ Rb þ kY ð14Þ figure, it can be seen that application of Eq. 18 gives
acceptable results for the range of 0 \ h/P \ 9.
Considering Eq. 14, Eq. 13 can be rewritten as:
  !
b 3:884
q ¼ Vb  0:216h exp 0:94 ln 1 þ k  
B exp 1:18 Bb Conclusions
ð15Þ
pffiffiffiffiffiffiffiffi Equations of flow nappe profiles developed from experi-
Combining Eqs. 1 and 15 and considering Vb ¼ 2gh; mental data and free-vortex theory were used to determine
the discharge coefficient can be represented as follows: a relation for discharge coefficient in a sharp-crested rect-
  !
b 3:884 angular weir. Experimental data showed that upper and
Cd ¼ 0:324 exp 0:94 ln 1 þ k   ð16Þ lower nappe profiles could be fitted quite well by second-
B exp 1:18 Bb
and third-degree polynomial equations, respectively. The
More details for the development of Eq. 16 from Eqs. 1 results indicated that non-dimensional mean radius of
and 15 are presented in Appendix. Figure 7 reveals that the curvature of the streamline along the distance of OB at
best relation between the correction coefficient (k) and h/P lower nappe profiles (Rb/h) and non-dimensional flow
can be represented using a linear equation as follows depth at the section of maximum elevation of the lower
(R2 = 0.872): nappe (Y/h) have exponential relations with the ratio of

Fig. 6 Sketch of the a y y

concentric streamlines,
b non-concentric streamlines

x R3 x
R3
R2 R2
R1 R1

(a) (b)

123
178 Irrig Sci (2010) 28:173–179

Fig. 7 Variation of k with h/P

Fig. 9 Variation of Cd with h/P

According to the Fig. 4 (b/B = 1), it is found that the

lower and upper nappe profiles are approximated very
closely by Eqs. 19 and 20, respectively.
y x 3 x 2 x
¼ 1:318 2:205 þ 0:865 þ 0:013 ð19Þ
h h h h
y  x 2 x
¼ 0:251 0:239 þ 0:846 ð20Þ
h h h
These equations have been indicated in Fig. 4 for the
others of b/B (i.e. 0.75, 0.625, 0.5, 0.375 and 0.25).
Equating Eq. 19 to zero will give the intersection of the
lower nappe with the x axis (point B in Fig. 2) and equating
derivation of Eq. 19 to zero will determine the location of
maximum height of the lower nappe (point C in Fig. 2)
Fig. 8 Variation of calculated Cd with the measured Cd
y x 3 x 2 x
¼ 1:318 2:205 þ 0:865 þ 0:013
weir width to channel width (b/B). Moreover, using free- h h h h
x
vortex theory, an equation was determined for the dis- ¼ 0 ! ¼ 0:667 ð21Þ
h
charge coefficient of the rectangular sharp-crested weir. y 0 x 2 x
Analyzing the experimental data showed that the ratio of ¼ 1:318  3 2:205  2 þ 0:865
head over weir crest to weir height (h/P) can be applied to h h h
x
obtain a correction coefficient (k) for estimation of the ¼ 0 ! ¼ 0:254 ð22Þ
h
discharge coefficient (Cd). Further, comparison between
suggested equation (Eq. 18) and past and present data Substitution x/h = 0.254 in Eqs. 19 and 20, y/h = 0.112
indicates (and indeed support) that Eq. 18 can reasonably for lower nappe and y/h = 0.769 for upper nappe were
predict the coefficient Cd for the range of 0 \ h/P \ 9. obtained. Considering Fig. 2, the equation can be written as:
This equation begins to drift away from real data for h/ OB ¼ 0:667h
P C 9. AC ¼ 0:112h ð23Þ
Y ¼ 0:769h  0:112h ¼ 0:657h
Appendix in O0 AB triangular:
 2  2  
Equations 11 and 12 will be obtained for b/B = 1 as Rb Rb 0:667 2
¼  0:112 þ ! Rb ¼ 0:553h ð24Þ
follows: h h 2

123
Irrig Sci (2010) 28:173–179 179

Also, Eq. 16 has been obtained from Eqs. 1 and 15 in References

the following way:
Equation 1 can be rewritten as: Bos MG (1989) Discharge measurement structures. International
Institute for Land Reclamation and Improvement, ILRI,
2 pffiffiffiffiffi Q 2 pffiffiffiffiffiffiffiffi Wageningen
Q ¼ Cd 2g  bh1:5 ) q ¼ ¼ Cd 2gh  h ð25Þ Brater EF, King HW, Lindell JE, Wei CY (1996) Handbook of
3 b 3
pffiffiffiffiffiffiffiffi hydraulics, 7th edn. McGraw-Hill, New York, pp 5–13
Substituting Vb ¼ 2gh in Eq. 15: French RH (1986) Open channel hydraulics. McGraw-Hill, New York
pffiffiffiffiffiffiffiffi Henderson FM (1966) Open channel flow. Macmillian, New York
q ¼ 2gh  0:216 ! Johnson MC (2000) Discharge coefficient analysis for flat-topped and
  sharp-crested weirs. J Irrig Sci 19:133–137
b 3:884
 h exp 0:94 ln 1 þ k   ð26Þ Kandaswamy PK, Rouse H (1957) Characteristics of flow over
B exp 1:18 Bb terminal weirs and sills. J Hydraulics Div ASCE 83(4):1–13
Kindsvater CE, Carter RW (1957) Discharge characteristics of
Equating Eq. 25 with 26, Cd can be given as follows: rectangular thin-plate weirs. J Hydraulics Div ASCE 83(6):1–36
Ramamurthy AS, Tim US, Rao MVJ (1987) Flow over sharp-crested
2 pffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffi
Cd 2gh  h ¼ 2gh  0:216 plate weirs. J Irrig Drainage Eng ASCE 113(2):163–172
3   ! Rehbock T (1929) Discussion of precise weir measurements by EW
b 3:884 Schoder and KB Turner. Trans ASCE 93:1143–1162
 h exp 0:94 ln 1 þ k   Rouse H (1936) Discharge characteristics of the free overfall. Civil
B exp 1:18 Bb
Eng ASCE 6(4):257–260
ð27Þ Strelkoff T (1964) Solution of highly curvilinear gravity flows. J Eng
! Mech Div ASCE 90(3):195–221
  Swamee PK (1988) Generalized rectangular weir equation. J
b 3:884
Cd ¼ 0:324 exp 0:94 ln 1 þ k   ð28Þ Hydraulic Eng ASCE 114(8):945–949
B exp 1:18 Bb White FM (2003) Fluid mechanics, 5th edn. McGraw-Hill, New York

123