Rischio
Start free

Rankine Earth Pressure Theory vs Coulomb Wedge Theory

Rankine's earth pressure theory and Coulomb's wedge theory: their assumptions, Ka and Kp, when each one applies, and both worked on the same wall.

9 min readLéelo en español
34.1 kPaWall friction · δ′ = 17°Active thrustPa = 92.2 kN/m
In this article
  1. Rankine’s earth pressure theory
  2. Coulomb’s wedge theory
  3. Rankine vs Coulomb: the differences
  4. Which theory to use
  5. Worked example: Rankine and Coulomb on the same wall
  6. The example in Rischio
  7. Frequently asked questions
  8. References

Rankine’s and Coulomb’s are the two classical theories for the active and passive pressure of a soil on a retaining wall. Rankine (1857) works from the state of stress in the soil at failure. Coulomb (1776) works from the equilibrium of a wedge of soil sliding on a plane behind the wall. They agree in the simplest case and part ways once the wall has friction, a leaning back or a sloping backfill.

This guide sets out each theory with its assumptions and coefficients, compares them, says which one AASHTO LRFD applies to which wall, and works both on the same wall. What lateral earth pressure is, and the active, passive and at-rest states, are in lateral earth pressure.

Rankine’s earth pressure theory

William Rankine published his theory in 1857. It takes the soil behind the wall to be everywhere at failure, in a state of plastic equilibrium, and reads the pressure on the wall off that state of stress. It assumes that:

  • the back of the wall is vertical;
  • the backfill’s surface is level (an extension covers a uniform slope, below);
  • there is no friction between the wall and the backfill.

The active pressure at a depth with vertical effective stress σv′\sigma'_v is

σa′=σv′tan⁡2(45°−ϕ′2)−2c′tan⁡(45°−ϕ′2)=σv′Ka−2c′Ka\sigma'_a = \sigma'_v \tan^2\left(45° - \frac{\phi'}{2}\right) - 2c'\tan\left(45° - \frac{\phi'}{2}\right) = \sigma'_v K_a - 2c'\sqrt{K_a}

where:

  • ϕ′\phi' is the backfill’s effective friction angle;
  • c′c' is its effective cohesion;
  • Ka=tan⁡2(45°−ϕ′/2)K_a = \tan^2(45° - \phi'/2) is Rankine’s coefficient of active earth pressure.
−2c'√Kaσ'v Ka − 2c'√Kazc

With cohesion, the pressure is negative near the surface, where the soil would pull on the wall. Instead, it cracks away from the wall down to the depth where the pressure is zero:

zc=2c′γKaz_c = \frac{2c'}{\gamma\sqrt{K_a}}

The passive pressure is

σp′=σv′tan⁡2(45°+ϕ′2)+2c′tan⁡(45°+ϕ′2)=σv′Kp+2c′Kp\sigma'_p = \sigma'_v \tan^2\left(45° + \frac{\phi'}{2}\right) + 2c'\tan\left(45° + \frac{\phi'}{2}\right) = \sigma'_v K_p + 2c'\sqrt{K_p}

with Kp=tan⁡2(45°+ϕ′/2)K_p = \tan^2(45° + \phi'/2), Rankine’s coefficient of passive earth pressure. Here the cohesion adds: the passive diagram starts at 2c′Kp2c'\sqrt{K_p} at the surface, which raises the passive force.

2c'√Kpσ'v Kp + 2c'√Kp

Surcharges

A uniform surcharge qq over the whole backfill raises the vertical stress by qq at every depth, so it adds qKaq K_a to the active pressure, and qKpq K_p to the passive, down the full height of the wall:

σa′=(σv′+q)Ka−2c′Ka,σp′=(σv′+q)Kp+2c′Kp\sigma'_a = (\sigma'_v + q) K_a - 2c'\sqrt{K_a}, \qquad \sigma'_p = (\sigma'_v + q) K_p + 2c'\sqrt{K_p}
qq Kaσ'v Ka

A strip load, such as a footing or a road parallel to the wall, adds a pressure that changes with depth. It comes from the theory of elasticity, not from Rankine’s, and is the solution for a wall that doesn’t yield:

Δσh′=2qπ(β−sin⁡βcos⁡2α)\Delta\sigma'_h = \frac{2q}{\pi}\left(\beta - \sin\beta\cos 2\alpha\right)

where β\beta is the angle, in radians, that the strip subtends at the point of the wall, and α\alpha is the angle between the vertical and the line to the middle of the strip.

qβαΔσ'h

Sloping backfill

When the backfill rises at an angle α\alpha, Rankine’s coefficients become

Ka=cos⁡α cos⁡α−cos⁡2α−cos⁡2ϕ′cos⁡α+cos⁡2α−cos⁡2ϕ′,Kp=cos⁡α cos⁡α+cos⁡2α−cos⁡2ϕ′cos⁡α−cos⁡2α−cos⁡2ϕ′K_a = \cos\alpha\,\frac{\cos\alpha - \sqrt{\cos^2\alpha - \cos^2\phi'}}{\cos\alpha + \sqrt{\cos^2\alpha - \cos^2\phi'}}, \qquad K_p = \cos\alpha\,\frac{\cos\alpha + \sqrt{\cos^2\alpha - \cos^2\phi'}}{\cos\alpha - \sqrt{\cos^2\alpha - \cos^2\phi'}}

and the pressure acts parallel to the surface, inclined at α\alpha, with its resultant at H/3H/3. The slope can’t be steeper than ϕ′\phi': past it the root has no value, and the backfill wouldn’t stand on its own.

PaαH/3

Coulomb’s wedge theory

Charles-Augustin de Coulomb presented his theory in 1776. It looks at a wedge of soil behind the wall that slides on a plane through the wall’s heel, and finds the force on the wall from the equilibrium of the forces on the wedge: its weight WW, the reaction RR of the soil below the plane, at ϕ′\phi' from its normal, and the thrust of the wall, at δ′\delta' from the wall’s normal. The active thrust is the largest over all trial planes, and the passive the smallest.

βαWδ'Paφ'R

Coulomb’s theory takes into account:

  • the inclination of the backfill, α\alpha;
  • the inclination of the wall’s back, β\beta, measured from the horizontal at the heel, inside the wall: 90° for a vertical back, less when the back leans toward the front;
  • the friction between the backfill and the wall, δ′\delta'.

Its closed-form coefficients are for a cohesionless backfill, c′=0c' = 0. Since the theory balances forces on a wedge, it gives the magnitude and direction of the thrust and nothing about how the pressure is distributed. The distribution is assumed triangular, with the resultant at H/3H/3 (FHWA NHI-07-071, 3.2.4).

The active thrust per metre of wall is

Pa=12γH2KaP_a = \tfrac{1}{2}\gamma H^2 K_a

with γ\gamma the backfill’s unit weight, HH the wall’s height, and

Ka=sin⁡2(ϕ′+β)sin⁡2β sin⁡(β−δ′)[1+sin⁡(ϕ′+δ′)sin⁡(ϕ′−α)sin⁡(β−δ′)sin⁡(α+β)]2K_a = \frac{\sin^2(\phi' + \beta)}{\sin^2\beta\,\sin(\beta - \delta')\left[1 + \sqrt{\dfrac{\sin(\phi' + \delta')\sin(\phi' - \alpha)}{\sin(\beta - \delta')\sin(\alpha + \beta)}}\right]^2}

The passive thrust is Pp=12γH2KpP_p = \tfrac{1}{2}\gamma H^2 K_p, with

Kp=sin⁡2(β−ϕ′)sin⁡2β sin⁡(β+δ′)[1−sin⁡(ϕ′+δ′)sin⁡(ϕ′+α)sin⁡(β+δ′)sin⁡(α+β)]2K_p = \frac{\sin^2(\beta - \phi')}{\sin^2\beta\,\sin(\beta + \delta')\left[1 - \sqrt{\dfrac{\sin(\phi' + \delta')\sin(\phi' + \alpha)}{\sin(\beta + \delta')\sin(\alpha + \beta)}}\right]^2}

For the wall friction, AASHTO’s commentary gives typical values by material (Table C3.11.5.3-1) and notes that long practice has used δ′=0.67ϕ′\delta' = 0.67\phi' for soil against concrete and δ′=0.33ϕ′\delta' = 0.33\phi' for soil against steel (C3.11.5.3).

Where the backfill is irregular, or a load sits on part of it, the closed form no longer applies, and the same equilibrium is solved for a series of trial wedges (AASHTO 3.11.5.3).

Rankine vs Coulomb: the differences

RankineCoulomb
BasisState of stress at failureEquilibrium of a sliding wedge
Wall’s backVerticalAny inclination
Wall frictionIgnoredIncluded, δ′\delta'
BackfillLevel, or a uniform slopeA uniform slope; irregular by trial wedges
CohesionIncluded (2c′K2c'\sqrt{K})Not in the closed form
ResultThe pressure at every depthThe thrust, its size and direction
Direction of the thrustParallel to the backfillAt δ′\delta' from the wall’s normal

On the coefficients, FHWA’s manual on earth retaining structures (NHI-07-071, 3.2.5) sums up the effect of wall friction:

  • Rankine can’t take wall friction into account, so it overestimates KaK_a slightly and underestimates KpK_p. That makes it conservative for most applications.
  • Coulomb includes wall friction, but its plane failure surface makes the passive results unreliable once δ′\delta' passes ϕ′/3\phi'/3. It underestimates KaK_a slightly and overestimates KpK_p, by a wide margin at higher ϕ′\phi', which is on the unsafe side. The real passive failure surface is curved, close to a log spiral.

With a vertical back, a level backfill and no wall friction, the two theories give the same pressure.

Which theory to use

AASHTO LRFD’s commentary (C3.11.5.3) decides by the wall’s geometry. Coulomb’s theory is needed where the back of the wall gets in the way of the full sliding surfaces Rankine assumes in the backfill. In general it applies to gravity, semigravity, prefabricated modular and MSE walls, and to concrete cantilever walls with short heels. For a cantilever wall with a long heel, either theory may be used, on the vertical plane through the heel, with the soil between the stem and that plane counted as part of the wall’s weight.

For the passive resistance in front of a wall, where Coulomb’s error grows with the wall friction, Rankine’s coefficient or a log-spiral solution is the safer choice.

Worked example: Rankine and Coulomb on the same wall

Find the active thrust on the wall in the figure with Rankine’s theory and with Coulomb’s.

10°Backfillγ = 18 kN/m³φ' = 30°c' = 0δ' = 20°3.00 m

Rankine’s theory

With the backfill rising at α=10°\alpha = 10°:

Ka=cos⁡10° cos⁡10°−cos⁡210°−cos⁡230°cos⁡10°+cos⁡210°−cos⁡230°=0.3495\begin{aligned} K_a &= \cos 10° \,\frac{\cos 10° - \sqrt{\cos^2 10° - \cos^2 30°}}{\cos 10° + \sqrt{\cos^2 10° - \cos^2 30°}} \\ &= 0.3495 \end{aligned}

Without a surcharge, the vertical stress at the base of the wall is

σv′=γH=18×3=54 kPa\sigma'_v = \gamma H = 18 \times 3 = 54 \text{ kPa}

and, the backfill having no cohesion, the active pressure there, parallel to the backfill, is

σa′=σv′Ka=54×0.3495=18.87 kPa\sigma'_a = \sigma'_v K_a = 54 \times 0.3495 = 18.87 \text{ kPa}

The thrust is the area of the triangle, acting at H/3H/3 above the base and inclined at 10°, like the backfill:

Pa=12σa′H=12×18.87×3=28.31 kN/mP_a = \tfrac{1}{2}\sigma'_a H = \tfrac{1}{2} \times 18.87 \times 3 = 28.31 \text{ kN/m}

Coulomb’s theory

The back is vertical, β=90°\beta = 90°, and the wall friction is δ′=23ϕ′=20°\delta' = \tfrac{2}{3}\phi' = 20°:

Ka=sin⁡2(30°+90°)sin⁡290° sin⁡(90°−20°)[1+sin⁡(30°+20°)sin⁡(30°−10°)sin⁡(90°−20°)sin⁡(10°+90°)]2=0.340\begin{aligned} K_a &= \frac{\sin^2(30° + 90°)}{\sin^2 90°\,\sin(90° - 20°)\left[1 + \sqrt{\dfrac{\sin(30° + 20°)\sin(30° - 10°)}{\sin(90° - 20°)\sin(10° + 90°)}}\right]^2} \\ &= 0.340 \end{aligned}

The thrust acts at H/3H/3, inclined at δ′=20°\delta' = 20° from the normal to the back:

Pa=12γH2Ka=12×18×32×0.340=27.54 kN/mP_a = \tfrac{1}{2}\gamma H^2 K_a = \tfrac{1}{2} \times 18 \times 3^2 \times 0.340 = 27.54 \text{ kN/m}

The two results

10°Pa = 28.31 kN/mRankine20°Pa = 27.54 kN/mCoulomb

The two theories give nearly the same thrust: 28.31 kN/m by Rankine and 27.54 kN/m by Coulomb, 3% less. The direction differs more. The wall friction tilts Coulomb’s thrust by 20°, so its horizontal component, the one that pushes the wall, is 27.54cos⁡20°=25.8827.54\cos 20° = 25.88 kN/m, against 28.31cos⁡10°=27.8828.31\cos 10° = 27.88 kN/m by Rankine.

If the wall friction were equal to the slope of the backfill, δ′=α=10°\delta' = \alpha = 10°, Coulomb’s KaK_a would be Rankine’s, 0.3495. On a vertical back, Rankine’s solution for a sloping backfill is Coulomb’s with δ′=α\delta' = \alpha.

The example in Rischio

Rischio’s retaining wall tool picks the method from the data, as AASHTO does. A wall without a heel, like this one, gets Coulomb’s pressure on its back, with a wall friction of half of ϕ′\phi', 15° here. That gives KaK_a = 0.343 and a thrust of 26.85 kN/m horizontal and 7.19 kN/m vertical. With the wall friction set to ⅔ of ϕ′\phi' (Wall friction), Coulomb gives the 27.54 kN/m above. With Rankine chosen in Earth pressure, it gives KaK_a = 0.3495 and a thrust of 27.88 kN/m horizontal and 4.92 kN/m vertical: the same 28.31 kN/m.

The free lateral earth pressure calculator opens on this example: the engine gives Rankine’s thrust, and the page works out Coulomb’s with δ′=20°\delta' = 20° beside it. Change the backfill’s slope or the wall friction to see where the two theories part. The rules Rischio follows for each wall are in Earth pressure.

Frequently asked questions

What is the main difference between Rankine’s and Coulomb’s theories?

Rankine works from the state of stress in a soil at failure, behind a smooth vertical wall, and gives the pressure at every depth. Coulomb works from the equilibrium of a wedge sliding on a plane, includes the wall friction and the inclination of the wall’s back, and gives the resultant thrust.

Which theory gives the larger active pressure?

Usually Rankine’s, since it ignores the wall friction. In the example, Rankine’s horizontal thrust is 27.88 kN/m and Coulomb’s, with δ′=20°\delta' = 20°, 25.88 kN/m.

Can Coulomb’s theory be used for passive pressure?

With care. Coulomb’s plane wedge overestimates the passive pressure once the wall friction passes about a third of ϕ′\phi', which is on the unsafe side. With wall friction, a log-spiral solution or Rankine’s coefficient is used instead.

Does Coulomb’s theory work for cohesive soils?

Its closed-form coefficients are for cohesionless backfill. A cohesive backfill is handled with Rankine’s equation and its cohesion term, or with a trial-wedge analysis that includes the cohesion on the sliding plane.

References

  • AASHTO (2024). LRFD Bridge Design Specifications, 10th edition. Article 3.11.5.3 and its commentary.
  • Das, B. M., & Sivakugan, N. (2019). Principles of Foundation Engineering.
  • FHWA (2008). Earth Retaining Structures, FHWA-NHI-07-071, Chapter 3.