GeoTerva Open the module

Theory · Sheet-Pile Walls

Cantilever sheet pile in clay: undrained pressures, the tension crack, embedment and moment

How a sheet-pile wall in clay finds its depth and its largest moment from the undrained strength, and how the base is checked against heave, with every figure drawn from the same numbers the module uses.

The module runs online, in your browser, on a phone, a tablet or a PC. Nothing to install, no dongle.

The Sheet pile in clay module on a phone: the wall in clay, the net pressure diagram, the embedment and the moment after Analyze
On this page · 17 chapters
  1. 1Sheet-pile walls and this module
  2. 2How a cantilever wall moves
  3. 3Active pressure of the retained clay
  4. 4The net pressure in the clay
  5. 5Equilibrium and the depth of embedment in clay
  6. 6The depth of embedment
  7. 7Water at different levels on the two faces
  8. 8Verifying the sheet pile to EN 1993-1-1
  9. 9Verifying the sheet pile to AISC 360-22
  10. 10The anchored wall
  11. 11The seismic situation
  12. 12The displacement of the wall
  13. 13The base of the excavation: bottom heave
  14. 14A worked example, by hand and by the module
  15. 15The opening wall, verified to EN 1993-1-1
  16. 16What the module assumes
  17. 17References
By the GeoTerva engineering teamUpdated 2026-10-1826 min readFree, no sign-up

A sheet-pile wall in clay holds itself up by the undrained strength of the clay it is driven into. This page follows the source for walls penetrating clay — the undrained pressures, the tension crack of the retained clay, the constant net pressure below the dredge line, the equilibrium of the wall and the quadratic for its depth, the maximum moment — and then what the module adds: layers, water at different levels, anchors, loads, the earthquake and the displacement.

First slide of the lecture slides: Sheet pile in clayPDF

Lecture slides

Ready-made slides for lectures and self-study

The mechanism, the derivation, the equations and the figures of this page, in teaching order. Every slide links back to its chapter.

37 slides · PDF · 1.6 MB

1Sheet-pile walls and this modulewhat a sheet-pile wall is, and which one is designed here

A sheet-pile wall is a row of interlocking piles driven into the ground to retain soil or water — a waterfront quay, a cofferdam, the side of an excavation. Timber, precast concrete and steel have all been used; steel sections dominate because they are light to handle, drive well and are re-usable. The steel piles come in two shapes, the Z-section and the U-section, and a manufacturer's table gives each one's section modulus per metre of wall — the number a design ends up looking the table up with.

Two ways of building the wall lead to the same structure: in a backfilled structure the piles are driven first and the soil is placed behind them; in a dredged structure the piles are driven into the existing ground and the soil in front of them is then removed down to the dredge line. Either way the wall ends up retaining a height L of soil above the dredge line, and holding itself up by its embedment D below it (the figure below).

water tableretained clayclay below the dredge linedredge lineLD
A cantilever sheet-pile wall: the retained soil behind it, the dredge line in front, the height L above the dredge line and the embedment D below it. The water may stand on both sides.

A cantilever wall has no other support: it stands like a fence post, fixed only by the passive resistance of the soil it is driven into. That makes it economical for moderate heights — as a rule up to about 6 m above the dredge line — beyond which an anchored wall, tied back near its top, is used instead.

This module designs a wall in clay: the profile in layers of undrained clay down the whole wall, each with its unit weights and its undrained shear strength cu (φ = 0). Above the dredge line the layers are retained, below it they act on both faces of the wall. The water may stand at the same level on both faces or at different levels; anchors, surcharges and the earthquake are taken as in the module in sand. A wall with a sand backfill is designed with the module in sand.

2How a cantilever wall movesrotation about a point near the toe, and the three zones of pressure

Under the push of the retained soil the wall rotates as a rigid body about a point O a little above its toe (the figure below). The rotation sets the pressures:

  • Zone A, above the dredge line: only the retained soil is there, pushing with its active pressure.
  • Zone B, from the dredge line down to O: the wall moves toward the excavation, so the retained side is active and the soil in front, being pushed, resists with its passive pressure.
  • Zone C, below O: the wall moves the other way, and the sides swap — passive behind, active in front.
OA — activeB — active behindpassive in frontC — passive behinddredge line
The cantilever wall rotating about O. Above the dredge line the retained soil is active; between the dredge line and O the soil in front resists passively; below O the sides swap.

The true distribution is curved. The source replaces it with a straight-line construction: the net pressure below the dredge line falls on one straight line through a zero point E — a point of the diagram, not the rotation point O — and the reversal at the toe is drawn as a block of height L5. Horizontal and moment equilibrium of that diagram fix the embedment.

The module draws exactly this construction (the Section view and the net-pressure figure of the Calculations tab): the active diagram A–C–D, the net line through E down to F, and the toe block F–G. The dashed extension F–H only fixes the ordinate at the toe; it is not a pressure area.

3Active pressure of the retained clayσv − 2cu, the tension crack zc, and P1 above the dredge line

Above the dredge line the retained clay pushes with its undrained active pressure — the vertical total stress less twice the undrained shear strength, Ka = 1 at φ = 0:

σa(z)=q+γz−2cu

Near the surface this is negative: the clay would pull on the wall. No such suction is credited — the pressure is zero down to the depth at which it vanishes, the depth of the tension crack, and grows linearly below it:

zc=2cu−qγ

With one clay above the dredge line the active diagram is then a triangle from zc to the dredge line, of resultant P1 acting a third of its height above the dredge line:

P1=12σa(L)(L−zc),z‾1=L−zc3
By default the crack is taken filled with water to its top — rain or surface water standing in it — and γw(z − zt) acts on the wall over the tension zone, zt its top, or the water table where it stands higher. Where the drainage of the site keeps it dry, the Design factors stage takes it dry, with a caution: no water in it above the water table, while the part of the crack below the water table holds the ground water, its pore pressure acting on the wall. The report prints the result of the other choice. Below the water table the total stress is the effective one plus the pore pressure; with the water at the same level on both faces the pore pressures cancel from the net pressure. With several layers above the dredge line the diagram changes at every boundary with its own cu and γ, and P1 and z̄1 are summed piece by piece.

4The net pressure in the clayundrained, φ = 0: σ6 = 4cu − (γL1 + γ′L2), σ7 at the toe

A clay loaded quickly, as a wall loads it, has no time to drain: it is analysed undrained, in total stress, with φ = 0 and its undrained shear strength cu. The Rankine coefficients are then both 1, and the lateral pressures on a face are the vertical total stress less or plus twice the cohesion:

σa=σv−2cu,σp=σv+2cu

Just below the dredge line the retained side pushes with the overburden it carries, γL1+γ′L2, less 2cu; the excavation side, with no overburden yet, resists with 2cu. Further down both faces gain the same overburden, so their difference does not change: the net pressure in the clay is constant, and it acts toward the retained soil:

σ6=4cu−(γL1+γ′L2)

It must be positive — a clay whose 4cu does not exceed the overburden at the dredge line cannot hold the wall, and the module stops there. The zero point E of the construction is the dredge line itself. Near the toe the wall moves the other way and the faces swap: the retained side resists with σv+2cu, the excavation side pushes with σv−2cu, and the reversed ordinate at the toe is

σ7=4cu+(γL1+γ′L2)
The total vertical stress is the effective one plus the pore pressure. With the water at the same level on both faces the pore pressures cancel from the net pressure, so the effective overburden of the retained clay appears in the expression for σ6 — γ above the water, γ′ below it; without a water table the overburden is γL. With the water at different levels the net water pressure is added. The pressures are not cut at zero below the dredge line: σ6 and σ7 are the source's unclipped differences.
A method decision. Kept below the dredge line where it is negative, the active pressure lets the retained face of a strong clay pull on the wall — the classical cohesive free-earth formulation of the source, and the one that gives σ6 and σ7. The embedment it gives can be very short and is sensitive to that choice. The module therefore offers the no-tension cutoff as well — the active pressure not taken below zero on either face — and prints the depth the other choice gives, whichever is used.
A clay above the dredge line, retained, pushes with σv − 2cu. Near the surface that is negative — the clay would pull on the wall. No such suction is credited: the pressure is taken as zero down to the depth at which σv = 2cu (the tension crack): no tensile soil stress is credited, and the pore-water pressure is treated independently, according to the groundwater levels. By default the crack is taken filled with water to its top — rain or surface water standing in it — and γw(z − zt) acts on the wall over the tension zone, zt its top, or the water table where it stands higher. Where the drainage of the site keeps it dry, the Design factors stage takes it dry, with a caution: no water in it above the water table, while the part of the crack below the water table holds the ground water, its pore pressure acting on the wall. The report prints the result of the other choice.

5Equilibrium and the depth of embedment in clayL4 from ΣH, the quadratic in D from ΣM, z′ and Mmax

Above the dredge line the active diagram of the retained soil has the resultant P1, acting a height z‾1 above the dredge line. Below it the net pressure σ6 acts toward the retained soil over the whole embedment D, and the toe block of height L4 swings the pressure to σ7 at the toe. Horizontal equilibrium gives the height of the toe block:

P1−σ6D+12L4(σ6+σ7)=0⇒L4=σ6D−P14cu

and the moments about the toe, with L4 substituted, a quadratic in D:

P1(D+z‾1)−σ6D22+4cu3L42=0
σ6(σ6−6cu)D2+P1(12cu−2σ6)D+P1(P1+12cuz‾1)=0

The shear is zero where the constant net pressure has cancelled P1, a depth z′ below the dredge line, and the moment is largest there:

z′=P1σ6,Mmax=P1(z′+z‾1)−σ6z′22
The module builds the same diagram as the module in sand — the pressures on both faces at every depth, the toe block from horizontal equilibrium, the depth from the moments about the toe — so it takes several layers, water at different levels, loads and anchors on the same footing. For the source's case its numbers are those of the quadratic for the depth and the maximum-moment equation; in the note its toe block is written L5 and the depth below E, here the whole embedment, L4.

6The depth of embedmentDtheory, and the two ways of adding a margin

The theoretical embedment is the depth at which the diagram is exactly in equilibrium:

Dtheory=L3+L4

At that depth the passive resistance is fully mobilised and there is no margin at all. The source offers two ways of adding one (the figure below):

  • Increase the depth: the diagram is built with the full Kp and the depth is increased by 20 to 30 percent, Ddesign=fDtheory with f=1.2 to 1.3. The pressures and the moment are those of the theoretical wall.
  • A factor of safety on Kp: the diagram is built from the start with Kp/FSp, FSp=1.5 to 2, and the depth that comes out is used as it is. Every ordinate below the dredge line changes, and the moment grows with them.
Increase D by a factor fDtheory​× ffull Kp in the diagram, the toe driven deeperFactor of safety on KpDtheory​Kp / FSa reduced Kp in the diagram, the depth used as found
The two routes to a margin on the embedment: a longer toe after the analysis, or a reduced passive coefficient before it.
The module offers both on the Design factors stage and states in the report which one was used. The total length of the piles is L + Ddesign.

A depth of your own. Instead of taking the design depth from the analysis, the depth of the sheet pile below the dredge line may be given. It is then verified against the depth the chosen route requires, and the factor of safety it mobilises on the passive pressure is found — the factor FSp,mob by which the passive pressure σp=σv+2cu can be divided before the diagram is only just in equilibrium at the given depth:

Dgiven≥Ddesign,req,D(σp/FSp,mob)=Dgiven

The total length, the displacement and the seismic check are worked with the given depth. For the internal forces two cases are worked: the required-depth design case, and the given-depth equilibrium case with the passive pressure divided by FSp,mob. The moment of the design case is not an upper bound of the given one — a longer wall in equilibrium with a partly mobilised passive resistance can bend more — so the sheet pile and the anchors are designed for the larger of the two; in the seismic situation the same is done when the given depth exceeds DE. The uniform mobilisation of the passive pressure is itself an idealisation.

For a wall penetrating clay the source increases the theoretical depth by 40 to 60 percent, f = 1.4 to 1.6. With the Kp/FS route the whole passive pressure of the clay is divided by FSp, (σv + 2cu) / FSp.

7Water at different levels on the two facesthe net water pressure added (beyond the source)

The source assumes the water stands at the same level on both faces. A dewatered excavation, or standing water in front of the wall, breaks that assumption: the hydrostatic pressures no longer cancel.

The module then adds the net water pressure to the net effective pressure for the no-seepage hydraulic condition, and gives each face its own effective overburden:

unet(z)=γw(z−Dw)−γw(z−Dw,exc),each term≥0
p(z)=Ka,eσ′v,r(z)−Kp,eσ′v,e(z)+unet(z)

The net line is then piecewise, so its zero point, the resultant above it and the two equilibrium equations are solved on the diagram itself rather than by the closed form; the construction — zero point E, toe block L5 — is the same.

Seepage under the toe changes the effective stresses and can reduce the passive resistance; the wall loading here is that of the stated no-seepage condition, and the report says so.

8Verifying the sheet pile to EN 1993-1-1from S = M/σall to MEd ≤ W fy / γM0

The classical route sizes the pile with an allowable stress:

S=Mmaxσallthe classical route — 170 MPa for ASTM A-328, 210 MPa for A-572 and A-690

The module replaces it with a limit-state verification. The moment of the net-pressure diagram is a characteristic action effect; the partial factor γF raises it to a design value, and EN 1993-1-1 compares the design moment with the resistance of the section:

MEd=γFMmax
Wreq=MEdγM0fy
Mc,Rd=WfyγM0,η=MEdMc,Rd≤1

Wreq is the number to look the sheet-pile table up with. The pile given is then checked against it. The table prints two moduli per size, and the class of the section decides which may be used (the figure below): the elastic Wel holds the section to first yield and is valid up to class 3; the plastic Wpl lets it reach its full plastic moment, which is allowed for a class 1 or 2 section.

elastic: first yield, Mel​ = Wel​ fy​plastic: all at yield, Mpl​ = Wpl​ fy​fy​fy​
The stress across a sheet pile at its elastic and at its plastic moment. The ratio of the two moduli, the shape factor, is about 1.1 to 1.2 for Z- and U-piles.
The factors of the Design code stage (recommended values)
Factorwhat it acts onvalue
γFthe earth-pressure moment, a permanent unfavourable action1.35
γM0the resistance of a cross-section1.00
A National Annex may set other values; the module uses what is entered.
What is assumed: the modulus given is that of the whole wall (βB = 1 — true of Z-piles, and of U-piles whose interlocks on the neutral axis are crimped or welded); the shear of the wall is not verified; the depth of embedment keeps the margin of its own route and is not factored again. A utilisation of 0.95 or more is reported as “PASS — near capacity”.

9Verifying the sheet pile to AISC 360-22LRFD or ASD on the flexural strength Mn

The pile may instead be verified to AISC 360-22. Its nominal flexural strength is Mn=FyZ for a compact section — the plastic modulus — or FyS at first yield, and it is compared with the moment by one of two methods:

LRFD:Mu=γHMmax≤φbMn,φb=0.90
ASD:Ma=Mmax≤MnΩb,Ωb=1.67

The earth pressure is taken as the lateral earth pressure H: in the LRFD combinations it enters with γH=1.6 where it adds to the load effect, in the ASD combinations with 1.0.

AISC 360-22 is written for the ASTM steels it lists. An EN grade is used here by its yield strength.

10The anchored wallone row or several — free earth support, the hinge idealisation, the anchor reactions and the bond zones

Where the cantilever grows too long or bends too much, a row of anchors near the top of the wall holds it back. With one row the source designs the wall by the free earth support method: the anchor is a support at the depth za, and the embedment is only deep enough for the passive resistance to balance the moments about it — the toe is free to move, so there is no reversal of pressure at the toe.

ΣMA=∫0ztoep(z−za)dz=0,F=∫0ztoepdz

The net pressure p is the same diagram as for the cantilever — active above the zero point E, net passive below it — cut off at the toe. The first equation fixes the depth, the second the anchor force per metre of wall; over the spacing s and along the tendon axis it is the support reaction TLEM=Fs/cosα of each anchor. TLEM is a reaction of the design diagram — not the lock-off load P0 and not the force the tendon carries in service; the prestressed spring model of the displacement gives its own tendon force TSR, and the tendon resistance and the bond capacity are separate checks. The moment peaks where the shear below the anchor, V=∫pdz−F, is zero. For one sand the first equation is the cubic in L4; the module solves it numerically, so layers, water, surcharge and the seismic situation all follow.

Several rows. With more than one row the wall is statically indeterminate. The module makes it determinate with the structural idealisation of the hinge method: a hinge is put at every anchor below the uppermost, so that the wall between two anchors is a simple beam, and the part below the lowest anchor zn is designed by free earth support about it. The depth comes from the moments about the lowest anchor, each span gives the reactions at its two anchors, and the anchor force of a row is the sum of the reactions it receives:

∫znztoep(z−zn)dz=0,Rk(zk+1−zk)=∫zkzk+1p(zk+1−z)dz

The uppermost span also carries the cantilever above the first anchor. The module applies the idealisation to its limit-equilibrium net-pressure diagram, so the reactions and the moments are those of that diagram. With one row the method is exactly the free earth support method. The moment peaks at the zeros of the shear between and below the anchors, or at the uppermost anchor. The anchors pull the wall down by Tsinα each; that vertical component is reported, not carried into the check of the sheet pile.

The anchor holds the wall only if its bond zone lies in soil that stays in place when the active wedge moves. the module takes an unbonded length of at least 4.5 m (strand tendons; 3 m for bars) and the bond zone to start at least max(1.5m,H/5) behind the critical failure surface — for a cohesionless soil the Rankine plane rising from the dredge line at 45°+φ′/2:

xb=Lfcosα≥(L−zb)tan(45°−φ′2)+max(1.5m,H/5),zb=za+Lfsinα

In the displacement model the anchor is a prestressed spring: it starts with P0cosα/s per metre of wall and stiffens with ka=EtAtcos2α/(Lfs), the free length stretching as the wall moves.

The pull-out capacity of the bond zone and the strength of the tendon are not verified; the bond length is used for the place of the bond zone.
In clay the active wedge behind the wall is bounded by the plane at 45° (φ = 0), and the bond zones are placed behind it.

11The seismic situationpseudo-static Mononobe–Okabe coefficients, restrained water, Westergaard

In an earthquake the soil behind the wall and in front of it is shaken with it. The pseudo-static method replaces the shaking by constant inertia forces on the soil wedges — khW horizontally and kvW vertically — and keeps the limit-equilibrium construction of the static design. The engineer gives kh and kv from the governing seismic code; a common form is kh=αS/r and kv=±0.5kh.

Rotating the gravity of the wedge by the angle θ turns Coulomb's wedge into that of Mononobe and Okabe. For a vertical wall, a level surface and no wall friction — the same idealisation as the static Rankine design — the coefficients are

tanθ=kh1∓kv
KAE=cos2(φ′−θ)cos2θ[1+sinφ′sin(φ′−θ)cosθ]2,KPE=cos2(φ′−θ)cos2θ[1−sinφ′sin(φ′−θ)cosθ]2

At θ=0 they are exactly Ka and Kp. The pressures are (1∓kv)KAEσ′v and (1∓kv)KPEσ′v; both senses of kv are checked. With wall friction the Mononobe–Okabe passive coefficient becomes unsafe, which is why the log-spiral solutions exist; with δ=0 it is not, and the plane wedge is kept. When θ>φ′ there is no wedge in equilibrium at all and the module says so.

Water. A sand of ordinary permeability is dynamically impervious: its pore water moves with it. Below the water level the submerged unit weight carries the pressure while the inertia acts on the saturated mass, so the angle grows:

tanθw=γsatγsat−γw·kh1∓kv

In a highly permeable soil (k>5·10−4 m/s) the pore water is free instead (Annex E.7): only the dry mass carries inertia, tanθw=γd/(γsat−γw)·kh/(1∓kv), with the dry unit weight γd entered for each layer. The pore water of each face is then one free column from its water level down to the base of the wall, and adds the Westergaard pressure with h′ measured from the base, as E.7 measures it; the two faces shake in phase, one pressed and the other sucked, so their pressures act together toward the excavation. Carrying E.7 down the embedded part of a sheet pile is an extension of this module. The module offers both.

Water standing free in the excavation above the dredge line adds the hydrodynamic pressure of Westergaard (1933), taken in the unfavourable sense, as a suction on the excavation face:

pwd(z′)=78khγwhz′,Pwd=712khγwh2

The check. The net-pressure diagram is built with these coefficients exactly as the static one — the same layers, water and surcharge, the zero point E and the toe block — and its two equilibrium equations give the embedment DE. The seismic situation is accidental, so KPE carries no factor: the wall as designed holds if DE≤Ddesign. For a cantilever the seismic pressures are distributed like the static ones, in proportion to the effective overburden: it is free to rotate about its toe, and the dynamic increment then acts at the point of the static thrust.

The anchored wall cannot rotate about its toe, and for such a wall the dynamic increment acts at mid-height. The retained side keeps its static Rankine pressure above the dredge line, and the increment of the active thrust over the retained height is spread uniformly over it:

ΔPAE=∫0L[(1∓kv)KAE−Ka]σ′vdz,ΔpAE=ΔPAEL

Below the dredge line the Mononobe–Okabe coefficients act on the overburden. The same anchored construction then gives DE and the seismic anchor forces TE=FEs/cosα. The active wedge of the earthquake is flatter than the static one — its plane rises at the Mononobe–Okabe angle αAE, 45°+φ′/2 at θ=0 — and the bond zones are checked behind it with the same margin; the tendon and the bond are not verified for TE.

The sheet pile is then verified under the seismic combination: in the seismic design situation the permanent actions and the seismic action enter with 1.0 and the variable ones with ψ2 — each surcharge takes the ψ2 entered with it, 1.0 for a permanent load — so MEd,E=Mmax,E of that diagram; in the LRFD and ASD routes the earth pressure H keeps 1.6 (LRFD) or 1.0 (ASD) and the seismic rise E enters with 1.0 or 0.7, Mu,E=1.6Mmax+(Mmax,E−Mmax) and Ma,E=Mmax+0.7(Mmax,E−Mmax), the two maxima combined on the safe side.

The displacement of the wall by subgrade reaction is computed for the static situation only.

The opening wall with kh=0.15 and kv=±0.075 — the water standing 2 m below the top on both sides, so 3 m of free water in the excavation:

DE=2.50m≤Ddesign=3.43m,Mmax,E=184.99kN·m/m(Mmax=181.91)

The static design holds in this earthquake: the seismic diagram needs 0.93 m less embedment than the design depth and 1.0 times the static moment.

Seismic situation — net pressure and bending momentNet pressure pBending moment Mdredge lineMmax,E​ = 184.99 kN·m/mMmax​ = 181.91 kN·m/mDE​ = 2.50 mz = 0— seismic (kᵥ downward) - - static designdepth z from the top of the wall ↓pressure axis on a square-root scalepseudo-static, GeoTerva undrained-clay approximation
The opening wall in the seismic situation: the net pressure and the moment (solid) beside the static design (dashed).
Mononobe–Okabe does not apply to an undrained clay: at φ = 0 its coefficients degenerate to 1. The clay is taken with the pseudo-static undrained wedge with its critical plane at 45°: σAE = (1 ∓ kv)(1 + tan θ) σ′v − 2cu and σPE = (1 ∓ kv)(1 − tan θ) σ′v + 2cu, tan θ = kh / (1 ∓ kv) — θw below the water, the pore water moving with the clay. This is a GeoTerva assumption, stated in the report.

12The displacement of the wallthe subgrade reaction method: an elastic beam on elasto-plastic springs

The limit-equilibrium construction of the previous sections gives the depth of embedment and the design moment, but not how far the wall moves: it describes the wall at the moment of failure, with the passive resistance fully mobilised. To estimate the displacement under working conditions the module uses the subgrade reaction method — a Winkler model, in which the wall is an elastic beam and the soil on each face a row of independent springs.

EId4ydz4=pR(z)−pE(z)+unet(z)

Here y is the displacement of the wall toward the excavation, EI its bending stiffness per metre of wall, and pR, pE the effective pressures of the retained and the excavation side. Each spring starts from the pressure at rest and changes in proportion to the displacement, never beyond the active or the passive pressure:

pR=K0σ′v,r−khy,Kaσ′v,r≤pR≤Kpσ′v,r
pE=K0σ′v,e+khy,Kaσ′v,e≤pE≤Kpσ′v,e
K0=1−sinφ′

The coefficient of horizontal subgrade reaction kh is the pressure per unit displacement. It is a parameter of the model rather than a property of the soil — it depends on the wall and on the loaded height as well — and the displacement depends on it: strongly while the springs are soft, hardly at all once the soil on both faces has reached its active and passive pressures, so the result is an estimate. For orientation, typical values for sand are: loose 4.8–16, medium dense 9.6–80, dense 64–128 MN/m³.

  • The wall has its design length, L+Ddesign, and both of its ends are free.
  • The pressures are characteristic: the full Rankine Kp, no factor of safety; the at-rest coefficient is that of Jaky (1944).
  • The wall is in place before the excavation, which is made in one stage; the construction sequence is not followed.
  • The equation is solved by beam finite elements, the springs iterated until each is either elastic or at its active or passive limit.
The displacement is reported as a value; it is not compared with a limit. Where the springs cannot hold the wall — every one of them yielded along the embedment — the model has no solution and the module says so; the limit-equilibrium design is unaffected.

The opening wall, with the design length 8.43 m, I = 21000 cm⁴/m, E = 210 000 MPa and kh = 15 MN/m³:

ytop=9.4mm,ydredge=2.7mm,ytoe=1.3mm

The largest moment of this model, 26.84 kN·m/m, differs from that of the limit-equilibrium construction (181.91 kN·m/m): the displacement model is an indicative response at working load — for instance it carries no water in a tension crack — and the design moment remains that of the limit-equilibrium construction.

Displacement and bending moment — subgrade reaction methodDisplacement yBending moment Mdredge lineytop​ = 9.4 mm2.7 mm1.3 mm← toward the excavationMmax​ = 26.84 kN·m/mz = 0toedepth z from the top of the wall ↓characteristic pressures · values indicative
The opening wall by the subgrade reaction method: its displacement (toward the excavation, to the left) and its bending moment.
In clay the spring starts from the total overburden, K0 = 1, and moves between the undrained active σv − 2cu (not below zero) and the passive σv + 2cu.

13The base of the excavation: bottom heavethe shear resistance along the failure surface against the weight beside the cut

A lateral equilibrium, however short the embedment it gives, says nothing about the base of the excavation. In clay the base can fail by bulging up: the clay under the excavation is squeezed in by the weight of the soil beside it, a bearing failure in reverse. The usual check for excavations in clay, as the braced cuts make it, compares the shear resistance along the failure surface with that weight:

FS=5.14cu(1+0.2B′′Lexc)+csideHB′σv(H)+q,B′=min(T,B2),B′′=2B′

B is the width of the excavation, Lexc its length in plan (a long excavation drops the shape term), T the depth of a hard stratum below the base — as the engineer enters it; a boundary between two clay layers is not one, and without T the failure surface is limited by the width alone — cu the clay under the base, cside the cohesion averaged over the retained height and σv(H) the total overburden beside the cut at the base level. The embedment of the sheet pile below the base is not credited — the safe side. The surcharge q is the uniform one plus the strip and point loads beside the excavation, which drive the heave too: each is turned into the uniform surcharge it puts on the ground within B′ next to the cut — the top of the soil block the heave pushes down. A strip load counts by its part within B′, spread over B′, qeq = qs b / B′; a point load within B′ is spread over a square B′ × B′, qeq = Q / B′², the length of the block along the wall taken equal to its width and its end shear not credited — a GeoTerva convention, on the safe side. A load beyond B′ does not bear on the block. With several layers the formula, derived for a homogeneous clay, is used in a GeoTerva layered-profile extension: the clay under the base for the bearing term, the cohesion averaged over the retained height for the side shear.

The overall stability of the wall with the ground around it — a deep slip surface passing below the toe and behind the anchors — is not part of this module; the report states that it has not been evaluated and needs its own verification.

14A worked example, by hand and by the modulea clay retained over a stiffer clay — every number checked

The worked example is a hand evaluation of the equations above: L = 5 m of clay (γ = 18 kN/m³, cu = 30 kPa) retained over a clay with cu = 70 kPa, no water, no surcharge — the tension crack taken dry, as the equations of the source take it.

zc=2×3018=3.333m,σa(L)=18×5−60=30kPa
P1=12×30×1.667=25.00kN/m,z‾1=0.5556m
σ6=4×70−90=190kPa,σ7=280+90=370kPa
−43700D2+11500D+12291.7=0⇒D=0.678m
z′=25190=0.1316m,Mmax=25×(0.1316+0.5556)−95×0.13162=15.53kN·m/m
The worked example — by hand and by the module
Quantityby handmodule
P1 (kN/m)25.0025.00
z‾1 (m)0.55560.5556
σ6 (kPa)190.00190.00
σ7 (kPa)370.00370.00
Dtheory (m)0.6780.678
L4 (m)0.37080.3708
z′ (m)0.13160.1316
Mmax (kN·m/m)15.5315.53
The module builds the pressures of both faces at every depth and solves the two equilibrium equations numerically; for this one-clay case its numbers are those of the closed form.
Net lateral pressure diagram on the cantilever wall5.000.67800.678dredge lineADBHGFpD​ = 30.00 kPaσ3​ = 190.00 kPaσ4​ = 370.00 kPa0.371P = 25.00 kN/mz̄ = 0.5556 m12Clay-1Clay-2pressure axis compressed (square-root scale) so the toe ordinates fit beside the active block
The net pressure diagram of the worked example as the module draws it: the active pressure of the retained clay below the tension crack, the constant net pressure σ₆ below the dredge line and the toe block to σ₇.

15The opening wall, verified to EN 1993-1-1MEd, Wreq, Mc,Rd and η for the default model

The wall the Model tab opens with is the wall of the worked example above with its tension crack filled with water to its top, the default of the module — Dtheory = 2.288 m and Mmax = 181.91 kN·m/m, against 0.678 m and 15.53 kN·m/m with the crack dry, with a sheet pile of steel S275 given by its elastic modulus Wel = 1400 cm³/m, verified to EN 1993-1-1 with γF=1.35 and γM0=1.00:

MEd=1.35×181.91=245.58kN·m/m
Wreq=245.58×1.00275000=893.0cm3/m
Mc,Rd=1400×10−6×2750001.00=385.00kN·m/m
η=245.58385.00=0.638≤1

The pile carries the design moment. The Calculations tab writes out the same chain for whatever wall is entered.

16What the module assumesthe limits of the calculation, in one place

  • Every layer is clay, analysed undrained in total stress with φ = 0 — the short-term condition. The long-term, drained condition (c′, φ′) is not checked here; for a wall that stays in place it can govern. A wall with a sand backfill is designed with the module in sand.
  • Rankine pressures without wall adhesion: Ka = Kp = 1, the active pressure σv − 2cu, the passive σv + 2cu; no suction above the dredge line (the tension crack, water-filled to its top by default or dry as chosen; below the water table it holds the ground water).
  • The source's simplified net-pressure diagram: the pressures of both faces at every depth, a toe block from horizontal equilibrium, the depth from the moments about the toe.
  • Water at the same level on both faces, or at different levels with the no-seepage condition; seepage is not modelled.
  • In the seismic situation the clay is taken with a pseudo-static undrained wedge with its critical plane at 45° — a GeoTerva assumption; Mononobe–Okabe does not apply at φ = 0.
  • The sheet pile is verified in bending on the modulus given; the shear of the wall, local buckling and corrosion loss are not verified. The displacement is estimated by the subgrade reaction method and reported without a limit.

17Referencesthe sources

  • Das, B.M. (2016). Principles of Foundation Engineering, 8th ed. Cengage Learning.
  • EN 1993-1-1:2005 — Design of steel structures, general rules.
  • ANSI/AISC 360-22 — Specification for Structural Steel Buildings.
  • Bowles, J.E. (1996). Foundation Analysis and Design, 5th ed. McGraw-Hill.
  • Jaky, J. (1944). The coefficient of earth pressure at rest. Journal of the Society of Hungarian Architects and Engineers, 7, 355–358.
  • Mononobe, N. and Matsuo, H. (1929). On the determination of earth pressure during earthquakes. World Engineering Congress, Tokyo. Okabe, S. (1926). General theory of earth pressure. Journal of the Japan Society of Civil Engineering, 12(1).
  • Westergaard, H.M. (1933). Water pressures on dams during earthquakes. Transactions of the American Society of Civil Engineers, 98, 418–433.

From theory to your own design

Design your own sheet-pile wall in clay in minutes

  • The undrained pressures, the tension crack, the embedment depth and the maximum moment, step by step
  • Water on both faces, anchors, the earthquake and the displacement of the wall
  • Bottom heave of the base, and the sheet pile checked to EN 1993-1-1 or AISC 360-22
  • A clean calculation report, saved as a PDF. Works on your phone and in your browser

More on sheet-pile walls