HPSC Engineering course‹› 0:00 studied

HPSC Engineering · Module 65 · standard depth · 20 min

Water in soil: permeability, effective stress and seepage

What HPSC Engineering asks from this chapter

Syllabus lines this chapter covers:

  • Subject Knowledge Test (Civil) · Engineering mechanics, solid mechanics, structural analysis, construction materials and management

HPSC Engineering questions from this module are being sorted into our bank.

Cut to standard depth for HPSC Engineering. 3 deeper paragraphs are in the full chapter.

Start from zero: soil is full of water paths

Dig a pit near a river in Bihar during the monsoon and it soon fills with water. Dig the same pit in stiff clay and it can stay nearly dry for days. Both soils lie below the water table, yet water moves through one in minutes and through the other in months.

The difference lies in the pores. Every soil is a heap of grains with tiny connected gaps between them, and water travels through these gaps like traffic through narrow lanes. Wide lanes between sand grains carry water easily, while the very fine lanes between clay plates almost block it.

Water in soil matters for two reasons. First, it moves: it seeps under dams, into excavations and out of canals. Second, it pushes: water in the pores is under pressure, and that pressure takes away part of the load that the grains would otherwise carry.

This chapter builds both ideas step by step. We start with what drives water through soil, then measure how easily it flows, and finally see how water pressure decides the real stress between the grains.

Why it matters. Most failures of earth structures involve water. Slopes slide after heavy rain, excavations flood and heave, and dams fail through piping. An engineer who can calculate seepage and water pressure can prevent all three.

What drives the flow: head and Darcy's law

Imagine a thin tube, called a standpipe, pushed into the ground at a point. Water rises in it to some level. The height of water above the tip is the pressure head, hw = u/γw. The height of the tip above a chosen datum is the elevation head, hz.

Their sum is the total head, h = hz + hw. Strictly, total head also has a velocity head, but water moves so slowly in soil that this part is ignored. Water always flows from a point of higher total head to a point of lower total head, not simply from high pressure to low pressure.

The loss of head per unit length of flow path is the hydraulic gradient, i = Δh/L. Experiments on sand filters showed that for steady, slow flow through saturated soil, the velocity is directly proportional to this gradient. This is Darcy's law:

v = k i, and the flow rate is q = k i A.

Here k is the coefficient of permeability, with units of velocity such as cm/s or m/s. A is the full cross-section of the soil, including grains, so v is only an average, called the superficial or Darcy velocity.

Water actually passes only through the pores, which occupy a fraction n of the area. So the true seepage velocity in the pores is vs = v/n, which is always larger than v. In reality the water also follows a winding path, which makes its real speed even higher.

Why it matters. Darcy's law turns a messy maze of pores into one number, k. Once k is known, the engineer can estimate how much water will enter a trench, leak from a canal or pass under a dam.

How permeable are different soils?

Permeability varies more than almost any other property in engineering. Typical values of k in cm/s are:

Soil k (cm/s)
Gravel about 1
Coarse sand 1 to 10⁻¹
Medium sand 10⁻¹ to 10⁻²
Fine sand 10⁻² to 10⁻³
Silty sand 10⁻³ to 10⁻⁴
Silt about 10⁻⁵
Clay 10⁻⁷ to 10⁻⁹

So a clean sand can be about a million times more permeable than a clay. A small amount of fines mixed into a sand can cut its permeability sharply, because the fine grains block the larger pores.

Several factors control k. The most important is pore size, which depends on the grain size, the grading and the density of packing. That is why denser soil, with a lower void ratio, is less permeable, and why compaction reduces k. The soil structure matters too, as compacted clays show, and layered deposits pass water more easily along the layers than across them.

For clean filter sands, Hazen gave a simple estimate, k ≈ 100 D10² cm/s, with D10 in cm. The effective size D10 controls flow because the finest tenth of the grains fills the pores of the rest. For example, a sand with D10 = 0.02 cm gives k ≈ 100 × 0.0004 = 0.04 cm/s.

Why it matters. The k value decides the whole drainage design. A sand around a well screen must let water in freely, while the clay core of a dam must hold water back. Choosing materials with the right permeability is a basic design step.

Measuring permeability in the lab and the field

The constant head test suits coarse soils, where flow is fast enough to collect and measure accurately. Water passes through a sample of length L and area A under a head difference h kept constant. If a volume Q is collected in time t, then q = Q/t and Darcy's law gives k = qL/(Ah).

The falling head test suits fine soils, where the flow is too small to collect in a reasonable time. A narrow standpipe of area a feeds water to the sample, and the water level in it falls from h1 at time t1 to h2 at time t2. Matching the flow out of the standpipe with the flow through the sample and integrating gives:

k = 2.303 (aL/At) log₁₀(h1/h2), where t = t2 − t1.

The narrow standpipe makes a small flow show up as a clearly measurable drop in level. That is why the test works for silts and clays.

Laboratory samples are small and may be disturbed. Field pumping tests measure the bulk permeability of a whole stratum. Water is pumped from a well at a steady rate q, and the water levels h1 and h2 are read in two observation wells at distances r1 and r2.

  • For an unconfined (open) aquifer, k = q ln(r2/r1) / [π(h2² − h1²)].
  • For a confined aquifer of thickness D, k = q ln(r2/r1) / [2πD(h2 − h1)].

Why it matters. Dewatering a deep foundation pit or designing a tube well depends on k for the whole aquifer, not a small sample. Field tests cost more, but they prevent large errors in pump sizing.

Layered soils: along the layers and across them

River deposits are rarely uniform. They usually come in horizontal layers of sand, silt and clay, each with its own permeability. To use Darcy's law, we replace the layers with one equivalent k for each direction.

For flow along the layers (horizontal), every layer has the same gradient, and the total flow is the sum of the flows in all layers:

kH = (k1H1 + k2H2 + k3H3) / (H1 + H2 + H3).

For flow across the layers (vertical), the same flow passes through every layer, and the head losses add up:

kV = (H1 + H2 + H3) / (H1/k1 + H2/k2 + H3/k3).

Think of the layers as lanes on a road. Traffic along the road uses the fastest lane, so kH is controlled by the most permeable layer. Traffic crossing the layers must pass the slowest one, so kV is controlled by the least permeable layer. As a result, kH is always greater than kV.

Why it matters. A thin clay band can make a whole sand deposit almost watertight in the vertical direction, while a thin sand seam can drain a thick clay quickly along its length. Missing either layer during site investigation leads to badly wrong predictions of seepage and settlement time.

Effective stress: the stress the grains really feel

At any depth, the total stress σ is the weight of everything above a unit area: soil, water and any surface load. Below the water table, the water in the pores is under a pressure u = γw × (depth below the water table), when there is no flow.

Karl Terzaghi, the founder of modern soil mechanics, saw that the water pressure pushes the grains apart and so carries part of the total stress. Only the remainder is carried through the grain-to-grain contacts. He defined this part as the effective stress:

σ′ = σ − u.

His principle has a second half, which is the important one. All measurable effects of a change in stress, such as compression and a change in shear strength, are caused only by a change in effective stress. Water cannot resist shear, so a soil's strength comes from the grains pressing on each other.

Effective stress cannot be measured directly, only calculated from σ and u. It is also not the true contact stress at the tiny points where grains touch; it is the load carried by the grains spread over the whole area.

A worked profile. Sand lies from the surface to 6 m, with the water table at 2 m. Above the water table the unit weight is 18 kN/m³, and below it 20 kN/m³. At 6 m, σ = 2 × 18 + 4 × 20 = 116 kPa and u = 4 × 9.81 = 39.2 kPa, so σ′ = 116 − 39.2 = 76.8 kPa.

Changes in water level change the effective stress in a definite way:

  • If the water table falls, u drops, and σ′ rises. This is why heavy pumping of groundwater can make the ground settle over a wide area.
  • If the water table rises below ground, u increases, and σ′ falls, so soil becomes weaker.
  • If a flood rises above the ground surface, σ and u increase by the same amount, so σ′ in the ground does not change.

Above the water table, water rises into fine pores by capillarity, and this water is in tension. The pore pressure there is negative, so σ′ is greater than σ, which is why damp sand can stand in a vertical sandcastle wall while dry or flooded sand cannot.

Why it matters. Every strength and settlement calculation in soil mechanics uses effective stress. Slopes fail after rain because rising pore pressure lowers σ′, not because the soil suddenly became weaker material.

Seepage force and the quick condition

Flowing water drags on the grains it passes. When water flows downward, this drag presses the grains together and increases the effective stress. When water flows upward, the drag lifts the grains and reduces the effective stress.

If the upward gradient keeps increasing, a point comes when the effective stress becomes zero. The grains then float in the rising water, the soil loses all strength, and it behaves like a thick liquid. This is the quick condition, also called boiling. Quicksand is not a special kind of sand; it is ordinary fine sand under strong upward flow.

The gradient at which this happens is the critical hydraulic gradient:

icr = γ′/γw = (G − 1)/(1 + e).

For a typical sand with G = 2.65 and e = 0.65, icr = 1.65/1.65 = 1.0. So an upward gradient of about 1 is enough to make a sand quick. Clays rarely go quick because their cohesion holds the grains, and gravels rarely do because they need too much flow, so fine sands and coarse silts are most at risk.

Why it matters. When an excavation is dug below the water table in fine sand, water flows upward into its base. If the gradient approaches 1, the base boils and the pit can collapse. Engineers prevent this by lowering the water table outside the pit or by driving sheet piles deep enough to lengthen the flow path.

Two-dimensional seepage and flow nets

Under a dam or around a sheet pile, water does not flow in a straight line. It curves downward, passes under the structure and rises on the far side. For steady flow in soil with the same k in all directions, combining Darcy's law with continuity gives the Laplace equation, which governs the head at every point.

The Laplace equation is usually solved by drawing a flow net. A flow net has two families of lines:

  • Flow lines show the paths that water particles follow.
  • Equipotential lines join points of equal total head.

The two families always cross at right angles, and they are drawn so that each figure is a curvilinear square, in which a circle can just be drawn. The space between two flow lines is a flow channel, and the head lost between two equipotential lines is the same everywhere.

If the net has Nf flow channels and Nd equal head drops, and the total head loss is H, then the seepage per unit length of the structure is:

q = k H (Nf/Nd).

A worked example. A sheet pile holds back a head of 6 m of water over a soil with k = 5 × 10⁻⁶ m/s. Its flow net has 4 flow channels and 12 head drops. The seepage is q = 5 × 10⁻⁶ × 6 × 4/12 = 1.0 × 10⁻⁵ m³/s per metre, or about 0.86 m³ per day per metre.

Two boundary rules help in drawing the net. A submerged permeable surface, such as the river bed upstream, is an equipotential line. The boundary between permeable soil and an impermeable layer or structure is a flow line.

Flow nets give three results that designers need:

  • the seepage loss under a dam or weir,
  • the uplift pressure along the base of a concrete structure, which tends to lift it, and
  • the exit gradient at the downstream toe, which decides the safety against piping.

Piping starts where the exit gradient approaches the critical gradient. Water carries away grains at the toe, a small channel forms and grows backward under the dam, and the dam can collapse suddenly. Many dams and weirs on sandy foundations have failed in this way.

Why it matters. Many barrages and weirs on the sandy plains of north India stand on permeable foundations. Their floors, cut-offs and filters are sized from exactly these ideas: limit the uplift and keep the exit gradient well below the critical value.

Quick revision

  • Water flows from higher to lower total head, where total head = elevation head + pressure head.
  • Darcy's law is v = ki, and the seepage velocity in the pores is vs = v/n.
  • Clean sand can be about a million times more permeable than clay, and k depends most on D10.
  • The constant head test is for coarse soils: k = qL/(Ah). The falling head test is for fine soils: k = 2.303 (aL/At) log₁₀(h1/h2).
  • In layered soil, kH is controlled by the most permeable layer and kV by the least, so kH > kV.
  • Effective stress σ′ = σ − u controls all compression and strength.
  • A falling water table raises σ′, a rising one lowers it, and a flood above ground leaves it unchanged.
  • Upward seepage reduces σ′, and the quick condition comes at icr = (G − 1)/(1 + e) ≈ 1.
  • Flow nets give q = kH Nf/Nd, the uplift pressure and the exit gradient that decides piping.

Check yourself

Exam-style questions on this chapter. Pick an answer to see the explanation. 0 of 14 answered

  1. Water flows through soil from a point of higher to a point of lower:

  2. If the Darcy velocity is 2 × 10⁻⁴ cm/s and porosity is 0.4, the seepage velocity is:

  3. The coefficient of permeability of a typical clay is of the order of:

  4. By Hazen's formula, a clean sand with D10 = 0.1 mm has k about:

  5. The falling head permeability test is most suitable for:

  6. Two layers of equal thickness have k = 10⁻³ and 10⁻⁵ cm/s. For flow across the layers, kV is about:

  7. For a horizontally layered soil deposit:

  8. Water table at 2 m; γ = 18 kN/m³ above and γsat = 20 kN/m³ below; γw = 9.81 kN/m³. The effective stress at 6 m depth is about:

  9. If flood water rises 3 m above the ground surface, the effective stress at a point in the saturated soil below:

  10. Lowering the groundwater table by pumping causes:

  11. For a sand with G = 2.68 and e = 0.68, the critical hydraulic gradient is:

  12. A flow net has 4 flow channels and 12 equipotential drops; head loss 6 m; k = 5 × 10⁻⁶ m/s. Seepage per metre length is:

  13. In a flow net, the boundary between a permeable soil and an impermeable layer is:

  14. Piping below a weir on sand is most likely when:

Previous year questions

Previous year questions for UPSC CSE will appear here once each one is checked word for word against the official question paper.

Frequently asked questions

What is Darcy's law?

For steady laminar flow in saturated soil, velocity is proportional to hydraulic gradient: v = ki, and the flow rate is q = kiA.

What is the critical hydraulic gradient?

It is the upward gradient that makes effective stress zero: icr = (G − 1)/(1 + e), which is about 1 for most sands.

Does a flood above ground change effective stress?

No. Total stress and pore pressure rise by the same amount, so the effective stress in the soil below stays the same.

How is seepage found from a flow net?

Seepage per unit length is q = kH × Nf/Nd, where Nf is the number of flow channels and Nd the number of head drops.