Diversion Works

Design of Hydraulic Structures

(CVL381)
Saumava Dey
Assistant Professor
Department of Civil Engineering
Block V, Room No. 304
Indian Institute of Technology Delhi
Email: [email protected]
 Theory of Seepage Flow
From the Continuity Equation:
𝜕𝑢 𝜕𝑣 𝜕𝑤
+ + =0
𝜕𝑥 𝜕𝑦 𝜕𝑧
From Darcy’s Law:
𝜕ℎ 𝜕ℎ 𝜕ℎ
𝑢= −𝐾 , 𝑣= −𝐾 , 𝑤= −𝐾
𝜕𝑥 𝜕𝑦 𝜕𝑧

Putting the values of u, v and w in the

continuity equation, we get: Assumptions:
𝜕2 ℎ 𝜕2 ℎ 𝜕2 ℎ ➢ The soil is homogeneous and isotropic.
𝐾 2 +𝐾 2 + 𝐾 2 =0
𝜕𝑥 𝜕𝑦 𝜕𝑧
➢ The voids are completely filled with water.
➢ No consolidation or expansion of the soil takes place.
∇2 ℎ = 0 (Laplace Equation) ➢ The soil and water are incompressible.
➢ The flow is steady and obeys Darcy’s law.
Dr. Saumava Dey, IIT Delhi
 Flownet Diagram
When the subsurface equipotential lines and
the seepage flowlines (streamlines) are
plotted, a flownet is constructed.

Let there be 𝑵𝒇 flowlines underneath the

structure. Then,
𝑞 = 𝑁𝑓 ∆𝑞
From Darcy’s law,
Δℎ
∆𝑞 = 𝐾 ∆𝑛
Δ𝑠
Let 𝑵𝒅 be the number of potential drops
between the u/s and d/s water levels. Assuming ∆𝑛 ≈ ∆𝑠,
Then, Seepage discharge underneath the structure:
𝐻𝑑𝑖𝑓𝑓 𝑁𝑓
∆ℎ = 𝑞 = 𝐾𝐻𝑑𝑖𝑓𝑓
𝑁𝑑 𝑁𝑑
Dr. Saumava Dey, IIT Delhi
 Determination of Seepage Pressure underneath
the structure

1. Trial and Error OR Graphical method

2. Mathematical solution of the Laplace Equation
3. Khosla’s method of independent variables
4. Method of electrical analogy
5. Method of relaxation

Dr. Saumava Dey, IIT Delhi

 Weaver’s Solution of the Laplace Equation
For 2D seepage flow occurring under a
straight horizontal floor:
Streamlines: Confocal ellipses with AB
as the major axis and the foci at A and
B.
𝑥2 𝑦2
2+ 2 =1
𝑏 𝑏
cosh 𝑢 sinh 𝑢
2 2
Equipotential lines: Confocal hyperbolas
𝑢: Stream function 𝑥2 𝑦2
2− 2 =1
𝑏 𝑏
1 𝑢 cos 𝑣 sin 𝑣
2 2
cosh 𝑢 = 𝑒 + 𝑒 −𝑢
2 ℎ
1 𝑣: Pressure function - 𝑣 = 𝜋Φ = 𝜋
𝐻
sinh 𝑢 = 𝑒 𝑢 − 𝑒 −𝑢
2
Φ: Pressure head ratio = Pressure at any point
Dr. Saumava Dey, IIT Delhi expressed as a ratio of the total head
 Weaver’s Solution of the Laplace Equation
Let P be any point within the 2D seepage flow domain.
In order to satisfy the above equations for streamlines and
equipotential lines, the following relationship is defined:
𝑏
𝑧 = cosh 𝑤
2
𝑧 = 𝑥 + 𝑖𝑦
𝑤 = 𝑢 + 𝑖𝑣
𝑏 Limiting Case: The first streamline is just
𝑥 + 𝑖𝑦 = cosh(𝑢 + 𝑖𝑣) beneath the outline of the floor AB.
2
𝑏 Hence, we can consider: 𝑢 = 0 at 𝑦 = 0.
𝑥 + 𝑖𝑦 = cosh 𝑢 cos 𝑣 + 𝑖 sinh 𝑢 sin 𝑣 𝑏 𝑏
Therefore, 𝑥 = cosh 0 cos 𝑣 = cos 𝑣
2 2 2
𝑏 −1
2𝑥
𝑥 = cosh 𝑢 cos 𝑣 𝑣 = 𝜋Φ = cos
2 𝑏
1 −1
2𝑥
𝑏 Φ = cos
𝑦 = sinh 𝑢 sin 𝑣 𝜋 𝑏
Dr. Saumava Dey, IIT Delhi 2
 Khosla’s method of independent variables
➢ Solution of the Laplace equation for seepage flow is feasible only for a simple
elementary profile of a weir section subjected to simple boundary conditions.
➢ Practically, barrage and weir sections conform to a composite profile with complex
boundary conditions.
➢ Khosla along with other investigators obtained solutions of the Laplace equation for
a number of simple profiles, using the method of independent variables.
➢ Solutions for these simple profiles are obtained in terms of the pressure head ratio 𝚽
at the key points.
➢ The key points are the junctions of the sheet piles with the floor.
➢ The values of 𝚽 calculated for the actual structure after breaking it down to these
simple profiles is valid for design purpose after corrections for the following are made:
1. Floor thickness
2. Mutual interference of the piles.
3. Slope of the floor.
Dr. Saumava Dey, IIT Delhi
 Khosla’s method of independent variables

𝑏
𝛼=
𝑑
1
𝜆 = 1 + 1 + 𝛼2
2

100 −1
𝜆−2 100 −1
𝜆−2
Φ𝐸 = cos Φ𝐶1 = 100 − cos
𝜋 𝜆 𝜋 𝜆
100 −1
𝜆−1 100 −1
𝜆−1
Φ𝐷 = cos Φ𝐷1 = 100 − cos
Dr. Saumava Dey, IIT Delhi
𝜋 𝜆 𝜋 𝜆
 Khosla’s method of independent variables
𝑏1
𝛼1 =
𝑑
𝑏2
𝛼2 =
𝑑
1
𝜆1 = 1 + 𝛼1 2 − 1 + 𝛼2 2
2
1
𝜆2 = 1 + 𝛼1 2 + 1 + 𝛼2 2
2
100 −1
𝜆1 + 1
Φ𝐶 = cos
𝜋 𝜆2
100 −1
𝜆1
Φ𝐷 = cos
𝜋 𝜆2
100 −1
𝜆1 − 1
Φ𝐸 = cos
𝜋 𝜆2
Dr. Saumava Dey, IIT Delhi
 Khosla’s method – Correction for Floor Thickness
The key points E/E1 and C/C1
correspond to the top of the floor.

The values of the pressure head ratio at

the corresponding key points E’/E1’
and C’/C1’ at the bottom of the floor is
calculated using linear interpolation
method.

Φ𝐸 −Φ𝐷
Φ𝐸′ = Φ𝐸 − 𝑡
𝑑

Φ𝐷 −Φ𝐶
Φ𝐶′ = Φ𝐶 + 𝑡
𝑑
Dr. Saumava Dey, IIT Delhi
 Khosla’s method – Correction for Mutual Interference of Piles
Correction C% for the key point C1a of sheet pile A for the interference of pile B:
𝑑1 𝑑1 +𝑑
𝐶 = 19
𝑏′ 𝑏

𝑏 ′ : Distance between two pile

lines A and B.
𝑑1 : Depth of the interfering
pile B (whose influence is to
be determined) below the level
at which the interference is
desired (i.e., below the top of
the neighbouring pile A).
𝑑: Depth of the pile A.
𝑏: Floor length.
Dr. Saumava Dey, IIT Delhi
 Khosla’s method – Correction for Mutual Interference of Piles

The correction C% is positive for key points upstream and negative for key points
downstream of the interfering pile.
The correction is calculated only for the key points of the adjacent pile towards the
interfering pile.

Key Points Interference Correction

of Pile Type
C1a (Pile A) Pile B +
E (Pile B) Pile A -
C (Pile B) Pile C +
EC (Pile C) Pile B -

Dr. Saumava Dey, IIT Delhi

 Khosla’s method – Correction for Slope of the Floor
➢ The correction for slope of the floor is applied to the pressures of the key point of
that pile which is fixed either at the beginning or at the end of the slope.
➢ Positive correction – downward slope.
➢ Negative correction – upward slope.

Correction at key point E of pile B:

𝑏𝑠
𝐶 = 𝐶𝑠 ′
𝑏

Slope (V:H) 1:1 1:2 1:3 1:4 1:5 1:6 1:7 1:8
Correction (Cs) in % 11.2 6.5 4.5 3.3 2.8 2.5 2.3 2.0
Dr. Saumava Dey, IIT Delhi
 Khosla’s method -Exit Gradient
Exit gradient is calculated as:
𝐻 1
𝐺𝐸 =
𝑑𝜋 𝜆
𝐻: Effective hydraulic head
𝑑: Depth of the downstream pile
𝑏 1
𝛼= , 𝜆= 1 + 1 + 𝛼2
𝑑 2

➢ If there is no d/s pile, 𝑑 = 0, and therefore, 𝐺𝐸 = ∞.

➢ It is therefore necessary to provide a vertical sheet pile at the d/s end of the
impervious floor.
➢ To prevent piping failure, the exit gradient is not allowed to exceed a critical value
depending on the soil type.
1 1
➢ The critical exit gradient for sand varies between and .
7 5
Dr. Saumava Dey, IIT Delhi
 Numerical Problem: Example 3
Using the Khosla’s method, calculate the residual seepage pressures at the key points
for the weir profile shown in the figure. Also calculate the value of the exit gradient.

Dr. Saumava Dey, IIT Delhi