Chapter 5 Permeability and seepage

5.1 Darcys Law 5.2 Determination of coefficient of permeability in the laboratory 5.3 Determination of coefficient of permeability in the field 5.4 Effective coefficient of permeability for stratified soils 5.5 Factors affecting the coefficient of permeability 5.6 Electroosmosis 5.7 Equation of continuity Laplaces equation 5.8 Use of continuity equation for solution of simple flow problem 5.9 Flow nets 5.10 Hydraulic uplift force under a structure 5.11 Numerical analysis of seepage 5.12 Seepage force per unit volume of soil mass 5.13 Safety of hydraulic structures against piping 5.14 Filter design

5.1 Darcys Law

Total head elevation head+pressure head+velocity head

Development of Darcys law .

i: hydraulic gradient L: the distance between sections A and B h: loss of head v: discharge velocity k: coefficient of permeability q: the rate of seepage A: cross section of the soil

The coefficient of permeability of soils is generally expressed at a temperature of 20C.

the actual velocity of seepage through soil

v<vs
vs is the actual velocity of seepage through the soil; is discharge velocity.

Typical values of coefficient of permeability for various soils

Validity of Darcys Law

Reynolds number Rn For laminar flow
Rn

D Rn 1 v-discharge (superficial) velocity, cm/s;

D-average diameter of the soil particle, cm; -density of the fluid, g/cm3;

-coefficient of viscosity, g/(cms). for flow of water through all types of soil (sand, silt, and clay), the flow is laminar and Darcys law is valid. With coarse sands, gravels, and boulders, turbulent flow of water can be expected, and the hydraulic gradient can be given by the relation

Variation of v with i [Eqs. (5.12) and (5.13).] Hansbo (1960) Eq.(5.12) Eq.(5.13)

Discharge velocity-gradient relationship for four clays

(after Tavenas et al., 1983b) These tests were conducted using triaxial test equipment, and the results show that Darcys law is valid.

5.2 Determination of coefficient of permeability in the

laboratory
Constant-head test suitable for more permeable granular materials

Constant-head laboratory permeability test

Falling-head test

is more suitable for fine-grained soils

Falling-head test
h-head difference at any time t;
A-area of specimen; a-area of standpipe;

L-length of specimen.

Permeability from consolidation test Tv-time factor; Cv-coefficient of consolidation; H-length of average drainage path; t-time. w-unit weight of water; mv-volume coefficient of compressibility. e-change of void ratio for incremental loading; -incremental pressure applied; e-initial void ratio.
For 50% consolidation, T = 0197

5.3 Determination of K in the field

Pumping from wells pumping from a well at a constant rate and observing the steady-state
water table in nearby observation wells.

Determination of coefficient of permeability by pumping from wellsgravity well

the maximum radius of influence, R, for drawdown due to pumping can be given by (Kozeny (1933))

n = porosity R = radius of influence t = time during which discharge of water from well has been established h1= hw at r1= rw and h2=H at r2= R

h1=hw at r1=rw and h2=h at r2=r

Pumping from partially penetrating gravity wells

Artesian wells

Determination of coefficient of permeability by pumping from wells

confined aquifer

T is the thickness of the confined aquifier

Auger hole test extend to a depth of 10 times the diameter of the hole or to an impermeable layer; Water is pumped out of the hole, after which the rate of the rise of water with time is observed in several increments. rw=radius of the auger hole d =depth of the hole below the water table

S =shape factor for auger hole

dh/dt =rate of increase of water table at a depth h measured from the bottom of the hole

Application example

In the phreatic acquiferthere is a strip pit with length of L=500m, the maximum radius of influence R=70m, thickness of the acquifer is H=4.5m. h=0.5m, k=6.8m/d. Determine rate of flow into the pit from two sides.
Original water table Phreatic water

impermeable

Original water table Phreatic water

impermeable

Solution: Q=2kIA

=26.8m/d (4.5-0.5)/70(4.5+0.5)/2500m2
=971m3/d
I A

5.4 Effective coefficient of permeability for stratified soils

q = q1+q2+q3++qn q1= kh1iH1 q2= kh2iH2 q3= kh3iH3 q = kehiH

Flow in horizontal direction in stratified soil q is the flow rate through the stratified soil layers combined and q1 q2 q3is the rate of flow through soil layers 1, 2, 3 respectively.

v1,v2,v3 are the discharge velocities in layers 1, 2, 3respectively

Flow in vertical direction in stratified soil

The ratio of keh/kev for some soil varies from about 1.5 to 3.7. Casagrande and Poulos (1969) provided the ratio of keh/kev for a varved clay that varies from 4 to 40.

5.5 Factors affecting the coefficient of permeability

1) Shape and size of the soil particles;
2) Void ratio; 3) Degree of saturation;

4) Composition of soil particles.

depends on the thickness of water held to the soil particles, the cation exchange capacity, valence of the cations... ; permeability decreases with increasing thickness of the diffuse double layer. 5) Soil structure; Fine-grained soils with a flocculated structure have a higher coefficient of permeability than those with a dispersed structure. 6) Viscosity of the permeant; 7) Density and concentration of the permeant.

5.6 Electroosmosis
the rate of seepage through clay soils is very small compared to that in granular soils, but the drainage can be increased by application of an external electric current. the exchangeable nature of the adsorbed cations in clay particles and the dipolar nature of the water molecules. Electroosmosis was first used by L. Casagrande in 1937 for soil stabilization Principles of electroosmosis in Germany.

e-flow velocity due to applied voltage E; D-dielectric constant; -zeta potential; -viscosity; L-electrode spacing.

Helmholtz-Smoluchowski theory for electroosmosis

5.7 Equation of continuity Laplaces equation

Laplaces equation of continuity on which flow net is based on.

Derivation of continuity equation

For steady flow through an incompressible medium, the flow entering

the elementary prism is equal to the flow leaving the elementary prism.

For two-dimensional flow in the xz plane

If the soil is isotropic with respect to permeability

Laplaces equation

Potential and stream functions

(x, z) and (x, z)

Consider a function (x, z) such that Laplaces equation

If differentiate above eqations with respect to x and z respectively and substitute in Laplaces equation, we get

Since x and z can be varied independently, or If h(x, z) is a constant equal to h1, above Eq. represents a curve in the xz plane. For this curve, will have a constant value. This is an equipotential

line. we can get a number of equipotential lines along which h=h1, h2, h3respectively. The slope along an equipotential line can now be derived:

Let

be a function such that

Note that the slope (dz/dx) is in the same direction as the resultant velocity.

we can see that at a given point (x, z) the equipotential line and the flow line are orthogonal.
The functions (x, z) and (x, z) are called the potential function and the stream function, respectively.

5.8 Use of continuity equation for solution of simple flow problem Boundary condition

One-directional flow through two layers of soil

where kA and kB are the coefficients of permeability of soils A and B,

respectively, and A is the area of cross-section of soil perpendicular to the direction of flow.

5.9 Flow nets

A set of flow lines and equipotential lines is called a flow net. A flow line is a line along which a water particle will travel. An equipotential line is a line joining the points that show the same piezometric elevation. The flow lines and the equipotential lines are drawn by trial and error. It must be remembered that the flow lines intersect the equipotential lines

at right angles. The flow and equipotential lines are usually drawn in
such a way that the flow elements are approximately squares. Drawing a flow net is time consuming and tedious because of the trialand-error process involved. Once a satisfactory flow net has been drawn, it can be traced out.

Boundary conditions: 1. AB is an equipotential line 2. EF is an equipotential line 3. BCDE (i.e., the sides of the sheet pile) is a flow line 4. GH is a flow line

Flow net around a single row of sheet pile structures

Flow through a flow channel

A flow channel is the strip located between two adjacent flow lines.

If the flow elements are drawn as squares

h = potential drop = drop in piezometric elevation between two

consecutive equipotential lines h = total hydraulic head = difference in elevation of water between the upstream and downstream side Nd= number of potential drops

If there are Nf flow channels in a flow net, the rate of seepage per unit

length of the hydraulic structure is

In the case the width-to-length ratio of the flow nets be a constant,

Example-Calculation of seepage from a flow net under a hydraulic structure For the flow net shown in the figure below: a) How high would water rise if a piezometer is placed at (i) A; (ii) B; (iii) C? b) If k=0.01mm/s, determine the seepage loss of the dam in m3/daym.

Solution The maximum hydraulic head h is 10 m. In the Figure, Nd=12, h= h/Nd= 10/12 = 0.833.
Part a(i): To reach A, water must go through three potential drops. So

head lost is equal to 30.833 = 2.5m. Hence the elevation of the water
level in the piezometer at A will be 102.5 = 7.5m above the ground surface. Part a(ii): The water level in the piezometer above the ground level is 1050.833 = 5.84m. Part a(iii): Points A and C are located on the same equipotential line. So water in a piezometer at C will rise to the same elevation as at A, i.e., 7.5m above the ground surface.

Part b: The seepage loss is given by q = kh(Nf/Nd). From the Figure ,

Nf = 5 and Nd= 12. Since

5.10 Hydraulic uplift force under a structure

Pressure head at D = 13.342(10/12)=11.67m

Pressure head at E
= (10+3.34)3(10/12)=10.84m Pressure head at F

= (10+1.67)3.5(10/12)=8.75m
Pressure head at G=4.56m; Pressure head at H=5.84m; Pressure head at I=5m.

Nd=12; Nf=5

The pressure heads calculated above are plotted in Figure. Between

points F and G, the variation of pressure heads will be approximately

linear. The hydraulic uplift force per unit length of the dam, U, can now be calculated as

Flow nets in anisotropic material

In developing the procedure for plotting flow nets, we assumed that the permeable layer is isotropic, i.e., k(horizontal) =k(vertical)=k. Let us now consider the case of constructing flow nets for seepage through soils that show anisotropy with respect to permeability. For two-dimensional flow problems, we refer to Eq. (5.93):

The steps for construction of a flow net in an anisotropic medium are as follows:

Example

Construction of flow net under a dam

Solution: From the given data,

when the soil is anisotropic with respect to permeability, the flow and equipotential lines are not necessarily orthogonal.

5.11 Numerical analysis of seepage

approximate finite difference equations for solving seepage problems Laplaces equation

using Taylors series

Hydraulic heads for flow in a region

considering Darcys law, q = kiA.

In figure (a)

Numerical analysis of seepage

Since the total rate of flow into point 0 is equal to the total rate of flow out of point 0, qinqout= 0. Hence
Taking x =z,

If the point 0 is located on the boundary of a pervious and an impervious layer, as shown in Figure b,

or

When point 0 is located at the bottom of a piling (Figure c), the equation for the hydraulic head for flow continuity can be given by

Note that 2' and 2'' are two points at the same elevation on the opposite sides of the sheet pile with hydraulic heads of h2' and h2'' , respectively. For this condition we can obtain (for x=z), through a similar procedure to that above,

for the case of flow across the boundary of one homogeneous soil layer to another (fig. (d)),

if we replace soil 2 by soil 1, the replaced soil (i.e., soil 1) will have a hydraulic head of h4' in place of h4. For the velocity to remain the same,

or
Taking x = z,

5.12 Seepage force per unit volume of soil mass

two flow lines ab and cd

two equipotential lines ef and gh

Seepage force per unit volume of soil mass:

Seepage force determination

where i is the hydraulic gradient.

5.13 Safety of hydraulic structures against piping When upward seepage occurs and the hydraulic gradient i is equal to icr, piping or heaving originates in the soil mass:

(0.85~1.1)

so
Harza (1935) investigated the safety of hydraulic structures against piping. According to his work, the factor of safety against piping, FS, can be defined as 3~4

where iexit is the maximum exit gradient.

Flow net under a dam with a toe filter

Terzaghi (1922) conducted some model tests with a single row of sheet piles as shown in the Figure and found that the failure due to piping takes place within a distance of D/2 from the sheet piles (D is the depth of penetration of the sheet pile). Therefore, the stability of this type of structure can be determined by considering a soil prism on the downstream side of unit thickness and of section DD/2. Using the flow net, the hydraulic uplifting pressure can be determined as Failure due to piping for a single-row sheet pile structure where ha is the average hydraulic head at the base of the soil prism.

The submerged weight of the soil prism acting vertically downward can be given by

Hence the factor of safety against heave is

A factor of safety of about 4 is generally considered adequate.

Safety against piping under a dam

For structures other than a single row of sheet piles, Terzaghi (1943) recommended that the stability of several soil prisms of size D/2D'1 be investigated to find the minimum factor of safety. Note that 0 < D' D. However, Harr (1962) suggested that a factor of safety of 4 5 with D' = D should be sufficient for safe performance of the structure.

5.14 Filter design When seepage water flows from a soil with relatively fine grains into a coarser material, there is a danger that the fine soil particles may wash away into the coarse material. For the proper selection of the filter material, two conditions should be kept in mind. 1) The size of the voids in the filter material should be small enough to hold the larger particles of the protected material in place.

2) The filter material should have a high permeability to prevent buildup

of large seepage forces and hydrostatic pressures in the filters.

Terzaghi and Peck (1948) provided the following criteria.

D15F= diameter through which 15% of filter material will pass D15B= diameter through which 15% of soil to be protected will pass D85B= diameter through which 85% of soil to be protected will pass

Use of filter at the toe of an earth dam

Determination of grain-size distribution of soil filters using Eqs. (5.179) and (5.180).

Based on laboratory experimental results, several other filter design criteria have been suggested in the past. These are summarized in the Table.

clay clay

Confined aquifer

Confined aquifer

If H=15m,

M=5m, =20kN/m3, w=10kN/m3.

Q: Is that safe for the pit? If not, how can you solve this problem?

Assignment:
A flow net for a single row of sheet piles is given in Figure. Determine the factor of safety against piping . Assume '= 10.2kN/m3.