HPSC Engineering · Module 65 · standard depth · 16 min
Shear strength of soils: Mohr-Coulomb, shear tests and drainage
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: why soil fails by sliding
Pour dry sand from your fist onto a table. It forms a cone, and the sides always stand at about the same slope. Try to make the sides steeper and the grains simply slide down. The sand has not crushed; its grains have slid over each other.
This is how soil almost always fails. Soil grains are hard, so soil does not break by crushing like a weak brick. It fails when one mass of grains slides past another along a surface. The resistance that soil offers to this sliding is its shear strength.
Shear strength decides three of the most common design questions in civil engineering:
- the bearing capacity of a foundation, which fails when soil slides out from under it,
- the earth pressure on a retaining wall, which depends on how much the soil behind can hold itself up, and
- the stability of slopes, cuttings and embankments.
This chapter explains where shear strength comes from, how a simple straight-line law describes it, how it is measured, and why the same soil can show very different strength depending on whether water can drain while it is loaded.
Why it matters. Shear strength is the single most used soil property in foundation design. An error in it shows up directly as a failed footing, a toppled wall or a slipped embankment.
Friction and cohesion: the Mohr-Coulomb law
Shear strength comes from two sources. The first is friction. Like a brick sliding on a rough floor, grains resist sliding more when they are pressed together harder. So the frictional part of strength grows with the normal stress σ on the sliding surface, and its rate of growth is described by the angle of internal friction, φ.
The second source is cohesion, c. Some soils, mainly clays, seem to hold together even with no normal stress, as a lump of moist clay does in your hand. This part of strength does not depend on σ.
In the eighteenth century, Coulomb combined the two into a straight-line law. Later, Mohr linked it to the stresses acting inside the soil, so the law is called the Mohr-Coulomb criterion:
τf = c + σ tan φ.
Here τf is the shear strength on a plane, and σ is the normal stress on that plane. For clean sand, c is about zero, so strength is purely frictional. For a saturated clay loaded quickly, as we shall see, φ behaves as zero, so strength is purely cohesive.
Terzaghi showed that only grain contacts resist shear, because water has no shear strength. So the correct form uses effective stress:
τf = c′ + σ′ tan φ′.
The parameters c′ and φ′ are the effective stress parameters, and they are the true strength properties of the soil.
Why it matters. The law explains everyday failures. After heavy rain, pore pressure in a slope rises, σ′ falls, and the frictional strength drops. The slope that stood for years slides, although nothing was added to it except water.
The Mohr circle and the failure plane
Inside a loaded soil, the stresses differ from plane to plane. At any point there are always two planes at right angles on which there is no shear stress. These are the principal planes, and the normal stresses on them are the major and minor principal stresses, σ1 and σ3. In a laboratory cylinder, σ1 usually acts vertically and σ3 sideways.
Mohr found that if we plot normal stress against shear stress for every plane through a point, the points form a circle. Its centre is at (σ1 + σ3)/2 and its radius is (σ1 − σ3)/2. Three facts follow:
- The largest shear stress, equal to the radius (σ1 − σ3)/2, acts on planes at 45° to the principal planes.
- The soil fails when its Mohr circle just touches the Mohr-Coulomb line, called the failure envelope.
- The failure plane is where the circle touches the envelope. It is inclined at 45° + φ/2 to the major principal plane, not at 45°.
So the plane of maximum shear stress is not the plane of failure. Failure happens where the shear stress is largest compared with the strength available, and a plane at a steeper angle has less normal stress, so less friction.
From the geometry of the circle touching the envelope, the principal stresses at failure are linked by:
σ1 = σ3 tan²(45° + φ/2) + 2c tan(45° + φ/2).
A worked example. A soil has c′ = 10 kPa and φ′ = 30°, and is tested at σ′3 = 100 kPa. Since tan(60°) = 1.732 and tan²(60°) = 3, σ′1 at failure = 100 × 3 + 2 × 10 × 1.732 ≈ 335 kPa. The failure plane is at 45° + 15° = 60° to the horizontal, if σ1 acts vertically.
Try it yourself. The free Mohr circle tool on ujjwalinfra lets you enter σ1, σ3, c and φ and see the circle, the envelope and the failure plane move as you change them.
Why it matters. The angle 45° + φ/2 reappears in retaining wall and bearing capacity theory, where it sets the shape of the wedge of soil that slides. Understanding it once here makes those topics far easier.
Measuring shear strength
The direct shear test is the simplest. Soil is placed in a square box, usually 60 mm × 60 mm, split horizontally at mid-height. A normal load presses down on the lid, and the top half of the box is pushed sideways until the soil shears. Repeating the test at three or more normal loads and plotting shear stress at failure against normal stress gives c and φ directly.
The direct shear test is quick and suits sands and gravels, and large boxes can test coarse soils. But it has real limits:
- The failure plane is forced to be horizontal, which may not be the weakest plane.
- Failure starts at the edges and spreads inward, so stresses are not uniform.
- Pore pressure cannot be measured, and drainage cannot be fully controlled, so undrained results are not reliable.
The triaxial test removes these limits. A cylindrical sample, about twice as long as its diameter, such as 38 mm × 76 mm, is wrapped in a rubber membrane and placed in a cell filled with water. The cell pressure applies σ3 all around. A ram then adds a vertical deviator stress, σ1 − σ3, until the sample fails.
Drainage lines allow water to leave or be held in, and a transducer measures the pore pressure.
The unconfined compression test is a special triaxial test with no cell pressure, σ3 = 0. A clay cylinder is simply squeezed, at about 0.5% to 2% strain per minute, until it fails or reaches 20% strain. The peak stress is the unconfined compressive strength, qu, and for a saturated clay the undrained cohesion is cu = qu/2. The test takes minutes and needs only a simple frame, so it is widely used for clays on site.
Why it matters. The test must match the question. A quick unconfined test gives a fair strength for a footing on soft clay built in a few weeks, but a long-term slope or a dam needs triaxial tests with drainage and pore pressure measured.
Drainage changes everything: UU, CU and CD
In a triaxial test, drainage can be allowed or stopped in each of the two stages: applying the cell pressure, and then shearing. This gives three standard tests:
- UU (unconsolidated undrained): no drainage in either stage. The test is quick.
- CU (consolidated undrained): the sample consolidates under the cell pressure, then is sheared without drainage, often with pore pressure measured.
- CD (consolidated drained): the sample consolidates, then is sheared so slowly that no excess pore pressure builds up.
In a UU test on saturated clay, raising the cell pressure raises the pore pressure by the same amount, so the effective stress does not change. Every sample therefore fails at the same deviator stress, and all total-stress Mohr circles have the same diameter. The envelope is horizontal: φu = 0, and τf = cu. For example, if a sample at σ3 = 300 kPa fails at a deviator stress of 100 kPa, then cu = 50 kPa, and an identical sample at σ3 = 400 kPa will fail at σ1 = 500 kPa.
For CU and CD tests, the total-stress circles grow with the cell pressure, giving inclined envelopes, and the total-stress c and φ differ from test to test. But when all results are plotted in effective stress, using the measured pore pressures, they fall on one envelope with c′ and φ′. Strength is a unique function of effective stress; the different total-stress results simply reflect different pore pressures.
Skempton expressed the pore pressure produced by a change in stress without drainage as:
Δu = B[Δσ3 + A(Δσ1 − Δσ3)].
B depends on saturation: it is 1 for a saturated soil and 0 for a dry one. A depends on the soil's history. It is high for normally consolidated clay, which tends to compress and build positive pore pressure, and low or negative for heavily over-consolidated clay, which tends to dilate.
How do engineers choose? A saturated clay loaded quickly, as at the end of building an embankment, has no time to drain, so its strength is cu from UU tests. Over the long term the pore pressures settle down, so effective-stress parameters from CD or CU tests with pore pressure are used. Staged loading and the sudden lowering of water in a reservoir are commonly checked with CU results.
Why it matters. The same clay can give cu = 30 kPa in a quick test and a much larger long-term strength. Using the wrong one either wastes money on oversized foundations or, worse, overestimates safety during construction, when soft clay is weakest.
Sands in shear: dense and loose
Sands drain quickly, so they are usually sheared under drained conditions, and their strength is described by φ′ with c′ about zero. Two things control their behaviour most: the relative density, ID, and the effective stress. A sand is usually called dense if ID is above 65% and loose if it is below 35%.
A dense sand shows a clear peak in shear stress at small strain, and then its strength falls. Its grains are tightly interlocked, so to slide they must ride up over one another, and the sample expands. This volume increase during shear is called dilation, and the extra work needed for it gives the dense sand its higher peak strength.
A loose sand has no peak. Its strength builds up gradually with strain, and the grains fall into the gaps between them, so the sample compresses.
At large strains, both samples reach the same final shear stress and the same final void ratio, whatever their starting density. The density at which a sand shears without any change in volume is a useful boundary between dilating and compressing behaviour.
Why it matters. A loose, saturated sand that tries to compress during fast loading, such as an earthquake, cannot drain in time. Its pore pressure rises, its effective stress falls, and it can lose almost all its strength. This is why loose sand fills are compacted or densified before heavy structures are built on them.
Quick revision
- Soil fails by sliding, and shear strength is the resistance to that sliding.
- Mohr-Coulomb: τf = c + σ tan φ, and in effective stress τf = c′ + σ′ tan φ′.
- Water has no shear strength, so strength depends on effective stress.
- The maximum shear stress acts at 45° to the principal planes, but the failure plane is at 45° + φ/2 to the major principal plane.
- At failure, σ1 = σ3 tan²(45° + φ/2) + 2c tan(45° + φ/2).
- The direct shear test forces a horizontal failure plane and cannot measure pore pressure; the triaxial test controls drainage and measures it.
- UU tests on saturated clay give φu = 0 and equal Mohr circles; the unconfined test gives cu = qu/2.
- In effective stress, UU, CU and CD results fall on one envelope.
- Skempton: Δu = B[Δσ3 + A(Δσ1 − Δσ3)], with B = 1 for saturated soil.
- Dense sand peaks and dilates; loose sand hardens gradually and compresses.