ETB/ETE/ETM/ETA 2102: Fluid Mechanics

Chapter 6: Open Channel Flow

1. At a certain location along the Rock River shown in Figure (a), the velocity (𝑉 ), of the flow is
a function of the depth(𝑦), of the river as indicated in Figure (b). A reasonable approximation
to these experimental results is,
2
𝑉 = 5𝑦 ⁄3 ______________(1)
where 𝑉 is in 𝑓𝑡/𝑠 and 𝑦 is in 𝑓𝑡. For what range of water depth will a surface wave on the
river be able to travel upstream?

2. Water flows up a 0.5 𝑓𝑡 tall ramp in a constant width, rectangular channel at a rate 𝑞 =
5.75 𝑓𝑡 2 ⁄𝑠 as is shown in Fig. (For now disregard the “bump.”). The upstream depth is
2.3 𝑓𝑡, and viscous effects are negligible. Determine the elevation of the water surface
downstream of the ramp, 𝑦2 + 𝑧2 .

3. Water flows in the canal of trapezoidal cross section shown in Fig. The bottom drops 1.4
ft per 1000 ft of length. The canal is lined with new finished concrete. Determine
i) The flowrate (when 𝑛 = 0.012) and
 ii) The Froude number for this flow.

For BG units, the flowrate have to be modified using a multification factor 𝑘 = 1.49.

4. Water flows in the channel shown in Figure (Question 3) at a rate of 𝑄 = 10.0 𝑚3 . For
this given flowrate, the Manning’s coefficient is 0.030. Determine the depth of the flow.
(Hint: The depth should be find out from a trial and error method and for the ease of finding,
check the value of depth for 1.0 m, 1.25 m, 1.5 m and 1.75 m).

5. Water flows in a round pipe of diameter 𝐷 at a depth 0 ≤ 𝑦 ≤ 𝐷 of as is shown in the

following Figure. The pipe is laid on a constant slope of and the Manning coefficient is 𝑛.
i) At what depth does the maximum flowrate occur?
ii) Show that for certain flowrates there are two depths possible with the same
flowrate. Explain this behavior.

6. Water flows along the drainage canal having the properties shown in Fig. E10.6a. The
1𝑓𝑡
bottom slope is 𝑆0 = ⁄500𝑓𝑡 = 0.002. Estimate the flowrate when the depth is 𝑦 =
0.8 𝑓𝑡 + 0.6 𝑓𝑡 = 1.4 𝑓𝑡. For BG units, the flowrate have to be modified using a
multification factor 𝑘 = 1.49.
7. Water flows uniformly in a rectangular channel of width 𝑏 and depth 𝑦. Determine the
aspect ratio, 𝑏/𝑦, for the best hydraulic cross section.

8. Water on the horizontal apron of the 100 𝑓𝑡 wide spillway shown in Figure has a depth of
0.60 𝑓𝑡 and a velocity of 18 𝑓𝑡/𝑠. Determine the depth 𝑦2 , after the jump, the Froude
numbers before and after the jump, 𝐹𝑟1 and 𝐹𝑟2 and the power dissipated, ℘𝑑 , within the
jump.

9. Water flows in a rectangular channel

of width 𝑏 = 2 𝑚 with flowrates
between 𝑄𝑚𝑖𝑛 = 0.02 𝑚3 ⁄𝑠 and
𝑄𝑚𝑎𝑥 = 0.06 𝑚3 ⁄𝑠. This flowrate is
to be measured by using either (a) a
rectangular sharp-crested weir, (b) a
triangular sharp-crested weir with 𝜃 =
90° (weir coefficient for triangular-
crested weirs can be taken from
relevant graph below), or (c) a broad-
crested weir (discharge coefficient for
broad-crested weirs can be taken from
relevant graph below). In all cases the
bottom of the flow area over the weir
is a distance 𝑃𝑤 = 1 𝑚 above the
channel bottom.
Plot a graph of 𝑄 = 𝑄(𝐻) for each
weir and comment on which weir
would be best for this application.
10. Water flows under the sluice gate shown in following Figure. The channel width is 𝑏 =
20 𝑓𝑡, the upstream depth is 𝑦1 = 6 𝑓𝑡 and the gate is 𝑎 = 1.0 𝑓𝑡 off the channel bottom.
Plot a graph of flowrate, 𝑄, as a function of 𝑦3 .
1. Consider, the river travels to the left with speed 𝑉, the surface wave travels upstream (to the
right) with speed √𝑔𝑦𝑐 relative to the water (not relative to the ground). Hence relative to the
stationary ground, the wave travels to the right with speed,

2
= 𝑐 − 𝑉 = √𝑔𝑦𝑐 − 5𝑦 ⁄3
𝑓𝑡 1 2
= (132.2 2 𝑦) ⁄2 − 5𝑦 ⁄3
𝑠
To travel the wave to upstream, 𝑐 − 𝑉 > 0
Then from the above equation,
𝑓𝑡 1 2
(32.2 2 𝑦) ⁄2 − 5𝑦 ⁄3 > 0
𝑠
𝑓𝑡 1 2
(32.2 2 𝑦) ⁄2 > 5𝑦 ⁄3
𝑠
𝟐, 𝟏𝟒 𝒇𝒕 > 𝒚

Thus, if the river depth is less than 2.14 ft, its velocity is less than the wave speed and the wave
can travel upstream against the current.

Considering the Froude No (𝐹𝑟), if a wave is to travel upstream, the flow must be subcritical
𝑉
(𝐹𝑟 = < 1). For this flow,
𝑐

2
𝑉 5𝑦 ⁄3
𝐹𝑟 = = 1 <1
𝑐 (𝑔𝑦𝑐 ) ⁄2
1 𝑓𝑡 1
5𝑦 ⁄6 (32.2 2 ) ⁄2 < 1
𝑠
1⁄
0.881𝑦 6 < 1
𝒚 < 𝟐, 𝟏𝟒 𝒇𝒕

Then, only when the flow is subcritical, the wave can travel upstream.

2. Applying the Bernoulli equation,

𝑉12 𝑉22
𝑦1 + + 𝑧1 = 𝑦2 + + 𝑧2
2𝑔 2𝑔
Then,
(𝑧1 −𝑧2 )
𝑆0 = and ℎ𝐿 = 0
𝑙
𝑧1 = 0, 𝑧2 = 0.5 𝑓𝑡, 𝑦1 = 2.3 𝑓𝑡, 𝑉1 = 2.5 𝑓𝑡/𝑠
The Bernoulli equation becomes,
 𝑉22
1.9 = 𝑦2 + _____(1)
64.4
Then, applying the Continuity equation,
𝑦1 𝑉1 = 𝑦2 𝑉2
𝑓𝑡 2⁄
5.75 𝑠 = 𝑦2 𝑉2 _____(2)
From equations (1) and (2),
𝑦23 − 1.9𝑦22 + 0.513 = 0_____(3)
Then getting the solutions of equation (3),
𝑦2 = 1.72 𝑓𝑡 𝑜𝑟 𝑦2 = 0.638 𝑓𝑡 𝑜𝑟 𝑦2 = −0.466 𝑓𝑡
Only considering the two realistic answers for 𝑦2 + 𝑧2 ,
𝑦2 + 𝑧2 = 1.72 + 0.5 = 2.22 𝑓𝑡
Or
𝑦2 + 𝑧2 = 0.638 + 0.5 = 1.14 𝑓𝑡
Considering the specific energy at upstream,
𝑉12 5.752
𝐸1 = 𝑦1 + = 2.3 + = 2.40 𝑓𝑡
2𝑔 2𝑔 (2.3)2
𝑦1 = 2.3 𝑓𝑡
Considering the critical energy,

3𝑦𝑐 3 3 𝑞2 3 3 5.752
𝐸𝑚𝑖𝑛 = = √ = √ = 1.513 𝑓𝑡
2 2 𝑔 2 32.174

𝑦𝑐 = 1.01 𝑓𝑡
Considering the specific energy at downstream,
𝐸2 = 𝐸1 + (𝑧1 − 𝑧2 ) = 2.4 + 0 − 0.5 = 1.9 𝑓𝑡
As the flow propagation is only a ramp, the flow have not passed its minimum energy region (flow
is subcritical at both up and down streams). Thus, 2.3 < 𝑦2 < 1.01
Then the correct value for 𝑦2 + 𝑧2 is,
𝒚𝟐 + 𝒛𝟐 = 𝟏. 𝟕𝟐 + 𝟎. 𝟓 = 𝟐. 𝟐𝟐 𝒇𝒕
3.
(i) The flowrate in BG units,
2⁄ 1⁄
𝑅 3𝑆 2
𝑄 = 1.49𝐴 ℎ 0 _____(1)
𝑛
For a depth of 𝑦 = 5 𝑓𝑡 the flow area is,
5
𝐴 = 12 𝑓𝑡 (5 𝑓𝑡) + 5 𝑓𝑡 ( 𝑓𝑡) = 89.8 𝑓𝑡 2
tan 400
The wetted perimeter,
5
𝑃 = 12 𝑓𝑡 + 2 ( 𝑓𝑡) = 27.6 𝑓𝑡
sin 400
The hydraulic radius is determined to be,
𝐴
𝑅ℎ = = 3.25 𝑓𝑡
𝑃
The slope of the bottom is,
1.4 𝑓𝑡
𝑆0 = = 0.0014
1000 𝑓𝑡
Then, from eq (1),

𝑸 = 𝟗𝟏𝟓 𝒇𝒕𝟑

(ii) The Froude number based on the maximum depth for the flow can be determined from,
𝑉
𝐹𝑟 =
(𝑔𝑙)1/2
Corresponding to the average velocity,
𝑄 915
𝑉= = = 10.2 𝑓𝑡/𝑠
𝐴 89.8
𝑙 = 𝑦 = 5 𝑓𝑡
Then the Froude Number can be calculated as,
10.2
𝐹𝑟 =
(32.18 × 5)1/2

𝑭𝒓 = 𝟎. 𝟖𝟎𝟒
4. Consider the figure given in question 3, but the depth is unknown. Thus the hydraulic radius
and flow areas are also unknown.

Since the flowrate here is given by SI units, converting other BG values to SI units,
12 𝑓𝑡 ⟹ 12 × 0.3048 = 3.66 𝑚
Then the flow area is,
𝑦
𝐴 = 3.66𝑦 + 𝑦 ( ) = 1.19𝑦 2 + 3.66𝑦
tan 400
The wetted perimeter,
𝑦
𝑃 = 3.66 + 2 ( ) = 3.11𝑦 + 3.66
sin 400
The hydraulic radius is determined to be,
𝐴 (1.19𝑦 2 + 3.66𝑦)
𝑅ℎ = =
𝑃 (3.11𝑦 + 3.66)
Then, considering the flowrate,
2⁄ 1⁄
𝑅 3𝑆 2
𝑄=𝐴 ℎ 0
𝑛
2⁄
( 2 ) 3
1 1.19𝑦 + 3.66𝑦 1
10 = × (1.19𝑦 2 + 3.66𝑦) ( ) (0.0014) ⁄2
0.03 (3.11𝑦 + 3.66)

(1.19𝑦 2 + 3.66𝑦)5 − 515(3.11𝑦 + 3.66)2 = 0

Using the trial and error method, checking y for the given values of depth: 1.0 m, 1.25 m, 1.5
m and 1.75 m,
𝒚 = 𝟏. 𝟓 𝒎

5.
(i) The the flow area is,
𝐷2
𝐴= (𝜃 − 𝑠𝑖𝑛𝜃 )
8
The wetted perimeter,
𝐷𝜃
𝑃=
2
The hydraulic radius is determined to be,
 𝐴 𝐷(𝜃 − 𝑠𝑖𝑛𝜃 )
𝑅ℎ = =
𝑃 4𝜃
Then, considering the flowrate,
2⁄ 1⁄
𝑅ℎ 3 𝑆0 2
𝑄=𝐴
𝑛
2⁄
2
1 𝐷 𝐷(𝜃 − 𝑠𝑖𝑛𝜃 ) 3 1⁄2
𝑄 = ( (𝜃 − 𝑠𝑖𝑛𝜃 )) ( ) 𝑆0
𝑛 8 4𝜃

8 5⁄
1 1⁄ 𝐷 ⁄3 (𝜃 − 𝑠𝑖𝑛𝜃 ) 3
𝑄 = 𝑆0 2 2⁄ ( 2 ) ____(1)
𝑛 8 (4 ) 3 𝜃 ⁄3
The term 𝑦 can be written in the form of,
𝐷 𝐷 360 − 𝜃 𝐷 𝐷 𝜃 𝐷 𝐷 𝜃
𝑦 = ( ) + ( ) 𝑐𝑜𝑠 ( ) = ( ) + ( ) 𝑐𝑜𝑠 (180 − ) = ( ) − ( ) 𝑐𝑜𝑠 ( )
2 2 2 2 2 2 2 2 2
𝐷 𝜃
𝑦 = (1 − 𝑐𝑜𝑠 ( )) _____(2)
2 2
Equations (1) and (2) show that 𝑄 is a function of 𝑦. Then, draw the graph of flowrate (𝑄)
versus flow depth (𝑦).

From the graph, it is clear that, the maximum flowrate, 𝑸𝒎𝒂𝒙 , does not occur when the
pipe is full. The 𝑸𝒎𝒂𝒙 occurs at 𝒚 = 𝟎. 𝟗𝟑𝟖𝑫.
The full flowrate occurs when,
𝑸𝒇𝒖𝒍𝒍 = 𝟎. 𝟗𝟐𝟗𝑸𝒎𝒂𝒙

(ii) Considering the graph, for any 0.929 < 𝑄⁄𝑄 < 1 there are two possible depths that
𝑚𝑎𝑥
give the same 𝑄.
The reason for this behavior can be seen by considering the gain in flow area, 𝐴, compared
to the increase in wetted perimeter, 𝑃, for 𝑦 ≈ 𝐷. The flow area increase for an increase in
𝑦 is very slight in this region, whereas the increase in wetted perimeter, and hence the
increase in shear force holding back the fluid, is relatively large. The net result is a decrease
in flowrate as the depth increases.
6. We divide the cross section into three subsections as is indicated in the Figure and write the
flowrate 𝑄 = 𝑄1 + 𝑄2 + 𝑄3 as where for each section,
2⁄ 1⁄
𝑅ℎ𝑖 3 𝑆0 2
𝑄𝑖 = 1.49𝐴𝑖
𝑛𝑖
The flow areas, wetted primers and hydraulic ratios are calculated using relevant equations and
given in the following table.
Section No (𝒊) 𝑨𝒊 (𝒇𝒕𝟐 ) 𝑷𝒊 (𝒇𝒕) 𝑹𝒉𝒊 𝒏𝒊
1 1.8 3.6 0.5 0.02
2 2.8 3.6 0.778 0.015
3 1.8 3.6 0.5 0.03

Thus the total flowrate is,

3 2⁄
1⁄ 𝑅 3
𝑄 = 𝑄1 + 𝑄2 + 𝑄3 = 1.49𝑆0 2 ∑ 𝐴𝑖 ℎ𝑖
𝑛𝑖
𝑖=1
2 2 2
1 0.5 ⁄3 0.778 ⁄3 0.5 ⁄3
𝑄 = 1.49 × 0.002 ⁄2 [(1.8 × ) + (2.8 × ) + (1.8 × )]
0.02 0.015 0.03
𝑸 = 𝟏𝟔. 𝟖 𝒇𝒕𝟐

7. Considering the flowrate (units are in SI units),

2⁄ 1⁄
𝑅 3𝑆 2
𝑄 = 𝐴 ℎ 0 _____(1)
𝑛
The area for a rectangular channel is given by,
𝐴 = 𝑏𝑦___________(2)
The wetted perimeter is given by,
𝑃 = 𝑏 + 2𝑦_________(3)
The hydraulic ratio from eqns (2) and (3),
𝐴 𝑏𝑦
𝑅ℎ = = ______(4)
𝑃 (𝑏 + 2𝑦)
The best hydraulic section is the one that gives the minimum 𝐴 for all 𝑦. So arranging eqn (4)
by using eqn (2),
𝐴 𝐴𝑦
𝑅ℎ = = ______(5)
𝐴 (𝐴 + 2𝑦 2 )
(𝑦 + 2𝑦)

Substituting eqn(5) in eqn (1),

 ⁄3 1 2
𝐴𝑦 ⁄
(( ) 𝑆0 2
𝐴 + 2𝑦 2 )
𝑄=𝐴
𝑛
Taking the power of 3/2 for whole eqn,
𝐴𝑦 3⁄
4
(( 2 ) 𝑆
3⁄ 3⁄ 𝐴 + 2𝑦 ) 0
𝑄 2=𝐴 2 3
𝑛 ⁄2

3⁄ 3⁄ 3⁄
4
Taking 𝑆0 ,𝑄 2 and 𝑛 2 as constants, 𝐾,
5
𝐴 ⁄2 𝑦
𝐾= ____(6)
(𝐴 + 2𝑦 2 )

5
𝐾(𝐴 + 2𝑦 2 ) = 𝐴 ⁄2 𝑦
To find minimum 𝐴 for all 𝑦, eqn (6) should be differentiate w.r.t. 𝑦,
5 3⁄ 𝑑𝐴 5 𝑑𝐴
𝐴 2 + 𝐴 ⁄2 = 𝐾 ( + 4𝑦) ______(7)
2 𝑑𝑦 𝑑𝑦
𝑑𝐴
When 𝑑𝑦 = 0, eqn (7) reduces to,
5⁄
𝐴 2 = 4𝐾𝑦
Then,
5
𝐴 ⁄2
𝐾 = _____(8)
4𝑦
From eqn (6) and (8),
5 5
𝐴 ⁄2 𝑦 𝐴 ⁄2
=
(𝐴 + 2𝑦 2 ) 4𝑦
4𝑦 = (𝐴 + 2𝑦 2 )
2

2𝑦 2 = 𝐴 = 𝑏𝑦
That is, the rectangle with the best hydraulic cross section is twice as wide as it is deep, or,
𝒃
=𝟐
𝒚

8. Conditions across the jump are determined by the upstream Froude number,
𝑉1 18 𝑓𝑡/𝑠
𝐹𝑟1 = = = 4.10
√𝑔𝑦1 √(32.2 𝑓𝑡/𝑠 2 )(0.6 𝑓𝑡)
𝑭𝒓𝟏 = 𝟒. 𝟏𝟎
𝐹𝑟1 > 1 thus, the upstream flow is supercritical, and it is possible to generate a hydraulic jump
as sketched.
 Considering the depth ratio across the jump as,
𝑦2 1 1
= (−1 + √1 + 8𝐹𝑟12 ) = (−1 + √1 + 8(4.12 ))
𝑦1 2 2
𝑦2
= 5.32
𝑦1
Then,
𝑦2
= 5.32
0.6
𝒚𝟐 = 𝟑. 𝟏𝟗 𝒇𝒕
Since the flowrate at any cross-sections is same,
𝑄1 = 𝑄2
𝑦1 𝑉1 0.6 𝑓𝑡 × 18 𝑓𝑡⁄𝑠
𝑉2 = =
𝑦2 3.9 𝑓𝑡
𝑽𝟐 = 𝟑. 𝟑𝟗 𝒇𝒕⁄𝒔
To determine the downstream Froude number,
𝑉2 3.39 𝑓𝑡/𝑠
𝐹𝑟2 = =
√𝑔𝑦2 √(32.2 𝑓𝑡/𝑠 2 )(3.19 𝑓𝑡)
𝑭𝒓𝟐 = 𝟎. 𝟑𝟑𝟒
To obtain the hydraulic jump, apply the energy equation for upstream and downstream,
𝑉12 𝑉22
𝑦1 + = 𝑦2 + + ℎ𝐿
2𝑔 2𝑔
182 3.392
0.6 + = 3.19 + + ℎ𝐿
(2 × 32.2) (2 × 32.2)
𝒉𝑳 = 𝟐. 𝟐𝟔 𝒇𝒕
The power dissipated ℘𝑑 ,
℘𝑑 = 𝑔𝑄ℎ𝐿 = 𝑔𝑏𝑦1 𝑉1 ℎ𝐿
𝑓𝑡
℘𝑑 = (62.4 𝑙𝑏/𝑓𝑡 3 ) × (100 𝑓𝑡) × (0.60 𝑓𝑡) × (18.0 2 ) × (2.26 𝑓𝑡)
𝑠
𝟓
℘𝒅 = 𝟏. 𝟓𝟐 × 𝟏𝟎 𝒇𝒕 ∙ 𝒍𝒃/𝒔

9. (a) For Sharp crested weir,

2 𝐻 3
𝑄= [0.611 + 0.075 ( )] √2𝑔 𝑏𝐻 ⁄2
3 𝑃𝑤
2 3 3
𝑄= [0.611 + 0.075𝐻]√(2 × 9.81) × 2𝐻 ⁄2 = 5.91 (0.611 + 0.075𝐻)𝐻 ⁄2
3
3
𝑄 = 5.91 (0.611 + 0.075𝐻)𝐻 ⁄2 _______(1)

(b) Similarly, for the triangular-crested weir,

8 𝜃 5
𝑄= 𝐶𝑤𝑡 𝑡𝑎𝑛 ( ) √2𝑔 𝐻 ⁄2
15 2
 8 5
𝑄= 𝐶𝑤𝑡 𝑡𝑎𝑛(45𝑜 )√2(9.81) 𝐻 ⁄2
15
5⁄
𝑄 = 2.36 𝐶𝑤𝑡 𝐻 2 ___________(2)

(c) Similarly, for the broad-crested weir,

3⁄
2 2
3⁄
𝑄 = 𝐶𝑤𝑏 𝑏√𝑔 ( ) 𝐻 2
3
1
𝐻 2 3
1+𝑃 2 ⁄2 3⁄
𝑤
𝑄 = 1.125 ( ) 𝑏√𝑔 ( ) 𝐻 2
𝐻 3
2+𝑃
𝑤
[ ]
Thus, with 𝑃𝑤 = 1 𝑚
1 3⁄
1+𝐻 2 2 2
3⁄
𝑄 = [1.125 ( ) ] (2)√9.81 ( ) 𝐻 2
2+𝐻 3

1
1 + 𝐻 2 3⁄
𝑄 = 3.84 ( ) 𝐻 2 _____(3)
2+𝐻
From eqns (1), (2) and (3), following graph can be drawn.

It appears as any of the three weirs would work well for the upper portion of the flowrate range.
Neither the rectangular nor the rectangular or the broad-crested weir would be very accurate
for small flowrates near 𝑄 = 𝑄𝑚𝑖𝑛 because of the small head, 𝐻, at these conditions.
The triangular weir, however, would allow reasonably large values of 𝐻 at the lowest
flowrates. The corresponding heads with 𝑄 = 𝑄𝑚𝑖𝑛 = 0.02 𝑚3 ⁄𝑠 for rectangular, triangular,
and broad-crested weirs are 0.0312, 0.182, and 0.0375 m, respectively.
In addition, for proper operation the broad-crested weir geometry is restricted to
𝐻
0.08 < 𝐿 < 0.50, where 𝐿𝑤 is the weir block length. From Eq. 3 with 𝑄𝑚𝑎𝑥 = 0.6 𝑚3 ⁄𝑠,
𝑤
𝐻𝑚𝑎𝑥 will be 0.349. Thus, to maintain proper critical flow conditions at the largest flowrate in
𝐻𝑚𝑎𝑥
the channel 𝐿𝑤 > 0.5 = 0.698 𝑚 . However, with 𝑄 = 𝑄𝑚𝑖𝑛 = 0.02 𝑚3 ⁄𝑠 , the 𝐻𝑚𝑖𝑛 will be
𝐻𝑚𝑖𝑛
0.0375 m. Thus, to ensure that frictional effects are not important, 𝐿𝑤 < 0.08 = 0.469 𝑚.
Clearly, these two constraints on the geometry of the weir block, 𝐿𝑤 , are incompatible.
A broad-crested weir will not function properly under the wide range of flowrates considered
in this question. The sharp-crested triangular weir would be the best of the three types
considered, provided the channel can handle the 𝑯𝒎𝒂𝒙 = 𝟎. 𝟕𝟏𝟗 𝒎 head.

10. Considering the flowrate per unit width in the sluice gate,
𝑞 = 𝐶𝑑 𝑎√2𝑔𝑦1
Then the total flowrate,
𝑄 = 𝑏𝑞 = 𝐶𝑑 𝑎𝑏√2𝑔𝑦1
𝑄 = 𝐶𝑑 1.0 𝑓𝑡 × 20 𝑓𝑡√2(32.2 𝑓𝑡⁄𝑠 2 )(60 𝑓𝑡)
𝑄 = 393𝐶𝑑
Then the graph can be drawn as follows.

For 𝑦3 < 3.2 𝑓𝑡 the flowrate is independent of 𝑦3 and the outflow is a free (not submerged)
outflow. For such cases the inertia of the water flowing under the gate is sufficient to produce
free outflow even with 𝑦3 > 𝑎.