GeoTerva Open the module

Theory · Braced Cuts

Braced cut in clay: apparent pressure, struts, wales and bottom heave

How the pressure on a braced cut in clay is found, and how the struts, the wales and the sheeting are designed to it, with the check of the base against bottom heave and 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.

Braced cut in clay · Results
The Braced cut in clay module after Analyze: the drawing and the results beside it
On this page · 17 chapters
  1. 1The braced cut and its members
  2. 2Why the pressure is not Rankine’s
  3. 3Apparent pressure from measured strut loads
  4. 4The envelopes for clay
  5. 5A surcharge, and water in clay
  6. 6The first strut and the tension crack
  7. 7Struts: from the envelope to the strut loads
  8. 8Sheet piles
  9. 9Wales
  10. 10The members as they are given
  11. 11Verifying the steel to EN 1993-1-1
  12. 12Verifying the steel to AISC 360-22
  13. 13The envelopes against real cuts
  14. 14The base of a cut in clay: bottom heave
  15. 15What the method assumes, and what it leaves out
  16. 16Worked example
  17. 17References
By the GeoTerva engineering teamUpdated 2026-10-1831 min readFree, no sign-up

A braced cut is held open by struts pressing on the sheeting from both sides, not by embedment below the base. This page follows the source section by section — the apparent-pressure envelopes, the averaging of several clays, the design of the struts, the sheet piles and the wales, the case studies and the bottom heave of the base — and then the verification of the steel members, with every figure drawn from the same numbers the module uses.

First slide of the lecture slides: Braced cut 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.

36 slides · PDF · 1.6 MB

1The braced cut and its membersSoldier beams or sheet piles, wales and struts.

Deep excavations with vertical faces — the basement of a building in a built-up street, a cut-and-cover tunnel — are held open by a temporary bracing system, which keeps the ground behind the face from failing and, with it, the settlement or the loss of bearing of the foundations nearby. Two arrangements are common. In the first, vertical soldier beams of steel or timber are driven before digging starts, and horizontal timber lagging is placed between them as the excavation goes down. In the second, interlocking sheet piles are driven first to form a continuous wall. In both, horizontal wales are fixed along the face at each strut level as soon as the excavation reaches it, and struts — horizontal steel members in compression — cross the cut from wale to wale.

ABCsectionsheetingwalestrut(a) soldier beams and lagging — planlagging · wale · struts(b) sheet piles — plans
Figure 1. The two common arrangements. In both the pressure of the ground is gathered by the wales at each level and carried across the cut by the struts, which are spaced s apart along the wall.

To choose the struts, the wales and the sheeting, the engineer needs the lateral pressure the bracing will receive. Theory gives the total thrust on the wall, but not how it is distributed down the wall: that depends on the soil, on the sequence and the speed of the work, on the crew and on the equipment. The design pressures are therefore empirical envelopes, drawn around the pressures back-figured from the strut loads measured on real cuts.

This module is for a cut with vertical walls braced by horizontal struts at any number of levels, in clay — one layer, or several clay layers averaged into one — with or without a surcharge beside the cut. Anchored and raking supports, the deflection of the wall and the settlement of the ground behind it are separate subjects, not part of this calculation.

The module draws the sheeting as sheet piles. For soldier beams and lagging the strut loads and the wale moments are the same; a soldier beam then takes the moment per metre of wall times its spacing, so its section modulus enters the module divided by that spacing.

2Why the pressure is not Rankine’sThe wall rotates about its top.

A cantilever retaining wall yields by rotating about its base, and Rankine’s active pressure — zero at the top, largest at the bottom — describes it well. A braced wall moves the other way. The first strut goes in while the cut is still shallow, and from then on the top of the wall hardly moves. Each deeper lift is dug before its strut is installed, and the time between the two grows with the depth, so the wall bends in progressively lower down: at the end of the excavation it has rotated about its top.

ABCapparentpressure σaRankine activeheld from the startyields as the cut deepensexcavationsheet piles · struts A–C · wales at each strut
Figure 2. The wall is held at the top from the start and yields below, so the pressure it receives, filled, is broad and nearly uniform instead of the dashed Rankine triangle.

Three things follow for the pressure. Near the top it stays close to the at-rest value, because the soil there has barely moved. Near the base it can fall below the Rankine active value, because the soil there has moved far past the state of plastic equilibrium. And the resultant thrust Pa acts higher up the wall than the third point of a triangle — at naH above the base, with na greater than 1/3.

The magnitude of that thrust can be found by the general wedge theory. A trial failure surface behind the wall is taken as an arc of a logarithmic spiral, and the wedge it cuts off is balanced by moments about the pole of the spiral:

r=roeθtanφ′(1)

Every radius of this spiral meets the curve at φ′ to its normal, so the resultant of the normal and friction forces on the surface passes through the pole and drops out of the moment equation. What remains is the weight W of the wedge, the adhesion Ca = c′aH on the wall, the moment MC of the cohesion along the spiral, and the thrust P itself, inclined at δ′ to the normal of the wall:

P=WlW−Cala−MClP,MC=c′2tanφ′(r2−ro2)(2)

Repeating the trial for several surfaces gives the largest P, the active thrust Pa. For a granular soil it comes out some 15 to 20 per cent above Coulomb’s value for a similar wall. But the analysis still says nothing about how the pressure is distributed down the wall, and na depends on factors that belong to the soil and to the site. That is why braced cuts are designed from envelopes instead.

The module does not compute the wedge thrust. As the source recommends, the bracing is designed from the empirical envelopes of the sections that follow.

3Apparent pressure from measured strut loadsHow the envelopes were made.

The pressure in a braced cut changes from place to place, and every strut should be designed for the largest load it may meet. The envelopes are built for exactly that. On an instrumented cut the load in each strut is measured, and dividing it by the area of wall that strut serves — the horizontal spacing s times half the distance to the strut above plus half the distance to the strut below — turns it into an apparent pressure over that band:

σ1=P1s(d1+d22),σi=Pis(di2+di+12)(3)

with d1 the depth of the first strut and di the distances between the following ones. A diagram drawn around all the diagrams obtained this way, cut after cut, is an apparent-pressure envelope.

σ1​P1​ measuredσ2​P2​ measuredσ3​P3​ measuredσ4​P4​ measuredd2​/2 + d3​/2apparent pressureeach strut load spread over the band of wall it serves: s × (half the distance to each neighbour)
Figure 3. Measured strut loads turned into apparent pressures: each load spread over the band of wall its strut serves. The envelope of the design is drawn around many such diagrams.

Peck built envelopes in this way from the strut loads observed in the subway cuts of Berlin, Munich and New York for sand, and in the same manner for soft to medium and for stiff clays. An envelope is therefore not the pressure on the wall at any one time; it is drawn so that the load taken from it for a single strut is on the safe side of the loads that were measured.

Read backwards, this bookkeeping is the module’s Every strut hinge convention: under a uniform pressure each strut then takes the band between the midpoints to its neighbours. The source’s own convention, the default, splits the wall differently — see the section on strut loads.

4The envelopes for claySoft to medium or stiff, by γH/c.

For clay the shape of the envelope depends on the stability number γH/c, with c the undrained cohesion (φ = 0). A soft to medium clay, with γH/c above 4, is being worked well past its strength by the cut:

γHc>4:σa=max(γH[1−4cγH],0.3γH)(4)

Its envelope is zero at the surface, grows linearly over the top quarter of the height and keeps the plateau σa down to the base. The larger of the two expressions governs: when 4c/γH comes close to 1 the first becomes small, and 0.3γH takes over. A stiff clay stays closer to its elastic range:

γHc≤4:σa=0.2γH to 0.4γH(0.3γH on average)(5)

Its envelope grows over the top quarter, holds the plateau over the middle half and falls back to zero over the bottom quarter.

H/4H/4HσaSandσa = 0.65 γ H KaH/4H/4HσaSoft to medium clay (γH/c > 4)σa = max{ γH[1 − 4c/γH], 0.3 γH }H/4H/4HσaStiff clay (γH/c ≤ 4)σa = 0.2 to 0.4 γH (0.3 γH used)The plateau value σa is what the tool reports; the shape decides how it is shared between the struts.
Figure 4. Peck’s three envelopes. Soft to medium clay: zero at the surface, a plateau from H/4 down. Stiff clay: ramps over the outer quarters and a plateau across the middle half. The sand rectangle is drawn for comparison.

Where the envelopes apply

The source lists four conditions that come with the envelopes:

  1. they apply to excavations deeper than about 6 m (20 ft);
  2. they assume the water table below the bottom of the cut;
  3. sand is taken as drained, with zero pore-water pressure;
  4. clay is taken as undrained, and its pore-water pressure is not considered.
The module chooses the envelope from γH/c. The stiff-clay coefficient is an input on the Analysis settings stage, 0.3 by default — the average the source works with; 0.2 and 0.4 are the ends of its range, so 0.3 is a central estimate rather than an upper bound. A cut shallower than 6 m is calculated and noted, not refused.

Several clay layers

When the cut passes through several clays, the source averages them by thickness and then uses the clay envelopes with the averages:

cav=1H(c1H1+c2H2+…+cnHn),γav=1H(γ1H1+γ2H2+…+γnHn)(6)
Only the part of each layer inside the cut counts; the last layer is taken to continue down to the base.

5A surcharge, and water in clayΔp = q beside the envelope.

The envelopes were assembled from cuts without a surface load, so a surcharge q beside the cut is not folded into the envelope — that would put the empirical coefficients on it as well. It is carried as its own uniform pressure over the full height of the cut, with Ka = 1 for the undrained clay:

Δpq=Kaq=q(φ=0⇒Ka=1)(7)

The clay envelopes are undrained, total-stress constructions: the water in the clay is already inside its unit weight and its undrained strength. A water table above the base is therefore drawn and noted, not added as a separate load; what it changes is the weight of the soil — below the table each layer counts its saturated unit weight, so the unit weight of a layer the table cuts through is the thickness average of the two:

γ=γzw+γsat(H−zw)H(8)one clay with the water table at zw; with several layers each row of the average takes its own part of the cut
The surcharge also enters the base: it adds to the load on the clay beside the cut in the heave check, and it shortens the tension crack at the top.

6The first strut and the tension crackzc = 2c/γ in clay.

A clay face can stand unsupported down to the depth at which its tension crack closes. The source takes it from the Rankine active pressure with cohesion, which is zero at that depth:

σ′a=γzKa−2c′Ka=0⇒zc=2c′γKa(9)

For φ = 0, Ka = 1 and the depth becomes 2c/γ. In a braced cut in clay the first strut should be installed above it, so that the sheeting never relies on an open clay face deeper than the clay itself can hold.

zc=2cγ(q=0),zc=2c−qγ(with a surcharge q)(10)
Azc​ = 2c/γcrack closesfirst strut above zc​clay, φ = 0above the crack depth the face would be held by tension the clay does not have
Figure 5. The tension crack at the top of a clay face: the first strut goes in above the depth at which the crack closes.
The module checks the first strut against zc when the top layer is clay, and warns when it lies deeper. A surcharge shortens the crack: with the surcharge on the surface the pressure (γz + q) − 2c is zero at (2c − q)/γ.

7Struts: from the envelope to the strut loadsThe source’s steps; P = R s.

Struts are horizontal columns: they carry the pressure of the wall along their axis and are bent by their own weight. Their capacity depends on their slenderness, which intermediate vertical and horizontal supports reduce, and in wide cuts they may have to be spliced. In construction work the struts should be at least about 2.75 m (9 ft) apart vertically. In clay the first of them goes in above the tension-crack depth of the previous section.

For the loads, the source gives a simplified and conservative procedure, which designers vary in detail:

  1. Draw the pressure envelope and the strut levels. The sheeting is taken as hinged at every strut level except the top and the bottom ones; many designers instead take it as hinged at every level except the top.
  2. Split the wall at the hinges into simple cantilevers at the ends and simple beams between them, and find the reactions of each piece by statics — A, B1, B2, C1, C2 and D in a four-strut cut.
  3. The load of each strut is the sum of the reactions it receives, times the horizontal spacing s.
  4. With the load of each level and the bracing conditions known, choose the sections.
PA=As,PB=(B1+B2)s,PC=(C1+C2)s,PD=Ds(11)
ABCDA–BB–CC–DInterior strutsthe source — Auto starts hereABCDA–BB–CC–DDAll but the top"many designers"ABCDAA–BB–CC–DDEvery strutAuto, if a strut is in tension= hinge (zero moment)
Figure 6. The same wall split into determinate pieces in three ways. Left, the source’s convention: hinges at the interior struts. Middle, the alternative many designers use. Right, a hinge at every strut: a cantilever above the first, simple spans between, a cantilever below the last.

Every piece is statically determinate: two supports carry it, and the moment equation about one of them together with the balance of horizontal forces gives both reactions. A strut at a hinge belongs to the pieces on both sides, which is why it collects two reactions. The reactions are forces per unit length of the wall; the strut spacing turns them into strut loads. In the worked example at the end of this page they come out as A = 54.02, B1 = 45.21, B2 = 45.21, C = 54.02 kN/m.

The module applies these steps to any number of struts at any levels, and offers four conventions on the Analysis settings stage: Interior struts (the source’s), All but the top, Every strut, and Auto, the default, which keeps the source’s convention unless it would put a strut in tension and then hinges every strut. A strut can only push: one that still comes out in tension is slack, is dropped, and the wall is re-solved on the rest. The sum of the reactions is checked against the resultant of the loading in the calculation.

8Sheet pilesThe largest moment and the modulus it asks for.

The sheet piles are designed from the same pieces. For each piece the largest bending moment is found — where the shear force changes sign, inside a span or at a strut where a cantilever meets it — and the largest of those is the design moment, per unit length of the wall. The required section modulus follows from the allowable flexural stress of the sheet-pile steel, and a section with at least that modulus is chosen from a table of sheet piles:

Mmax=maxpieces|M|,S=Mmaxσall(12)
ABCMmax​ = 27.03 kN·m/mM (per metre of wall)moment diagram of every piece, from the statics of the worked example
Figure 7. The moment along the sheeting of the worked example at the end of this page, piece by piece; the largest value anywhere on the wall is the design moment.
The module builds the moment diagram of every piece and takes the extreme anywhere along the wall. The sheeting is then given by its steel grade and its section modulus per metre of wall — the elastic or the plastic value a sheet-pile table prints — and it is verified to EN 1993-1-1 in the sections that follow, where the allowable stress of the source gives way to the resistance W fy / γM0 against the design moment γF M. The modulus that moment asks for is printed alongside, as the number to look the table up with.

9WalesA simple span s between struts: R s²/8.

Wales spliced properly may be treated as continuous along the wall. Conservatively they are taken as pinned at the struts: a simple span s loaded by the reaction of its level per unit length of the wall, so that at each level

Mmax=As28,(B1+B2)s28,(C1+C2)s28,Ds28(13)
R — the reaction of this level, per metre of wallstrutstrutM = R s² / 8 at mid-spans
Figure 8. A wale taken as pinned at two struts: a simple span s under the reaction of its level; its moment is R s²/8 at mid-span.

and the required section modulus follows as S = Mmax/σall. The level with the largest reaction governs. Wales are sometimes fastened to the sheet piles at points chosen to give them the lateral support they need.

The module takes the simple span, as the source does, with the shear R s/2 that goes with it. The wale is given either by its section modulus or as a section from the catalogue. Its compression flange bears against the sheeting along its whole length, so lateral-torsional buckling is not checked.

10The members as they are givenA steel grade and a section, by its numbers or from the catalogue.

The source sizes the members with an allowable stress and a table of sections. The module goes one step further and checks each member as it will be built. Every member — the sheeting, the wale and the strut — carries its own steel grade, and the grade hands over the yield strength and nothing else:

Table 1. The steel grades offered: fy of EN 1993-1-1 (hot-rolled steel, EN 10025-2), and ε
Gradefy (MPa), t ≤ 40 mmfy (MPa), 40 < t ≤ 80 mmε = √(235/fy)
S2352352151.00
S2752752550.92
S3553553350.81
The grade fills fy for elements up to 40 mm thick. “Typed” hands the box to the engineer, for a thicker element or a steel that is not in the list — a sheet-pile grade such as S270GP or S355GP of EN 10248, or S420 and S460.

The section is given the way a profile table gives it. The sheeting is given by its section modulus per metre of wall, since sheet piles are chosen from the maker’s tables. The wale is given by its section modulus about the axis it bends about, or as an IPE, HEA, HEB, HEM or UPN from the catalogue, or as a steel box. The strut is given by its area and its second moment of area about the weak axis — the two numbers its checks need — or as a round, square or rectangular tube or a rolled section from the catalogue.

A table prints two section moduli, and a modulus given by its number must say which of them it is:

sectionfy​fy​first yieldM = Wel​ fy​fy​fy​fully plasticM = Wpl​ fy​Wpl​ / Wel​ ≈ 1.10–1.25 for a rolled beam, 1.5 for a solid rectangle
Figure 9. The stress across a section at its elastic moment, when the outer fibre first reaches fy, and at its plastic moment, when the whole section has yielded.
Mel=Welfy,Mpl=Wplfy(14)

Whichever way a member is given, the calculation also prints what the load asks of it — the number to look a table up with:

Wreq=MEdγM0fy,Areq=NEdγM0fy(15)
A modulus does not say how thick the walls of the section are, and an area and an inertia do not either. For a member given by its numbers the class cannot be computed, and the shear area is unknown; the report states what was taken for each — class 1 or 2 for a plastic modulus, the class-3 resistance for an elastic one, class 3 or better for a strut given by A and I — and that the shear resistance was not verified.

11Verifying the steel to EN 1993-1-1Design actions, classes, bending and shear, buckling.

EN 1993-1-1 — one of the two design codes of the module — verifies each member against a resistance built from the yield strength and the partial factors. The loads that come off the envelope are working values — Peck’s envelope is an upper bound of measured loads — and they are raised once to design values:

Ed=γFEk(16)the three factors are on the Analysis settings stage, with the values of the tables below
Table 2. Partial factors on actions — EN 1997-1 (the same values as EN 1990, STR/GEO)
ActionFactorSet A1Set A2
PermanentunfavourableγG1.351.0
favourableγG1.01.0
VariableunfavourableγQ1.51.3
favourableγQ00
The earth pressure on the bracing is a permanent unfavourable action. Peck’s envelope is built from the soil strength as it is, so the factor goes on the action — set A1, as in Design Approaches 1 (combination 1) and 2: γF = 1.35. A surcharge is a variable action, 1.5 in the same set; the module puts one factor on all the bracing loads.
Table 3. Partial factors on resistances — EN 1993-1-1 and EN 1993-5, recommended values
Factorresistance ofEN 1993-1-1EN 1993-5 (piling)
γM0a cross-section, any class1.001.00
γM1a member to instability (buckling)1.001.10
γM2a cross-section in tension, to fracture1.251.25
A National Annex may set other values. γM2 is not used: no member of a braced cut is checked in tension.

The class of the cross-section

Whether a section can reach its plastic moment, or buckles locally first, is set by the slenderness of its walls — c/t against limits in ε:

ε=235fy(fy in MPa)(17)
Table 4. Limits of c/t for the parts the module classifies
Partclass 1class 2class 3
internal part, in compression33ε38ε42ε
internal part, in bending72ε83ε124ε
outstand flange, in compression9ε10ε14ε
tube, d/t50ε²70ε²90ε²
A rolled section is classed on the web between its fillets and on the outstand of its flange; a square or rectangular tube on the flat width of each wall. A class-4 section buckles locally before it yields: it is reported and not designed.

Bending: the sheeting and the wale

Mc,Rd=WfyγM0,W=Wpl for class 1 or 2,W=Wel for class 3,MEdMc,Rd≤1(18)

For the sheeting EN 1993-5 applies the same resistance to a unit width of wall, and reduces the modulus of a wall of single or double U-piles by a factor βB where the interlocks on the neutral axis may slip; its value is set by the National Annex, and it is 1.0 for Z-piles.

Shear in the wale

Vpl,Rd=Avfy3γM0,VEdVpl,Rd≤1(19)

Av is the shear area: for a rolled I-section A − 2btf + (tw + 2r)tf, never less than the web itself; for a rectangular tube Ah/(b + h). Above half the plastic shear the web cannot carry its full share of the moment, and the plastic modulus is reduced:

ρ=(2VEdVpl,Rd−1)2,MV,Rd=(Wpl−ρAw24tw)fyγM0≤Mc,Rd(20)

The reduction is written on the plastic modulus, so it applies to a class 1 or 2 section; past Vpl,Rd the section has already failed in shear and no reduction is made. The compression flange of a wale bears against the sheeting along its whole length, so lateral-torsional buckling is not the governing mode.

The strut: squashing and flexural buckling

Nc,Rd=AfyγM0,Ncr=π2EILcr2,λ‾=AfyNcr(21)
Φ=0.5[1+α(λ‾−0.2)+λ‾2],χ=1Φ+Φ2−λ‾2≤1,Nb,Rd=χAfyγM1(22)

A strut buckles about its weak axis, so I is the smaller second moment of area and Lcr = K Lb the effective length — the clear width of the cut unless a king post or a cross brace holds the strut in between. The imperfection factor α comes from the buckling curve of the section:

Table 5. The buckling curve about the weak axis, as the module takes it
SectionCurveα
hot-finished tube (EN 10210), up to S420a0.21
hot-finished tube, S460a00.13
cold-formed tube (EN 10219)c0.49
rolled I or H, h/b > 1.2, tf ≤ 40 mmb0.34
rolled I or H, h/b > 1.2, 40 < tf ≤ 100 mmc0.49
rolled I or H, h/b ≤ 1.2, tf ≤ 100 mmc0.49
rolled I or H, tf > 100 mmd0.76
channel (UPN)c0.49
For a strut given by its area and inertia the curve is chosen with them; curve c is the default and the safe side for every strut section in common use.
00.511.520.00.20.40.60.81.0relative slenderness λ̄χ1/λ̄² (Euler)curve a0 α = 0.13curve a α = 0.21curve b α = 0.34curve c α = 0.49curve d α = 0.76the strut of the cutthe module opens with:λ̄ = 0.934, χ = 0.579
Figure 10. The five buckling curves: χ falls with the slenderness, faster on the curves with the larger imperfection factor. Below λ̄ = 0.2 buckling may be ignored and the cross-section governs. Marked: the strut of the cut the module opens with.

For the cut the module opens with, the largest strut load is P = 271.26 kN, so NEd = 1.35 × 271.26 = 366.20 kN. With the strut the Sections stage starts from — A = 65.7 cm², I = 3598 cm⁴, fy = 275 MPa, curve c — over Lcr = 6.00 m the chain runs Ncr = 2071.5 kN, λ̄ = 0.9339, Φ = 1.1159, χ = 0.5791, Nc,Rd = 1806.8 kN and Nb,Rd = 1046.4 kN: buckling governs, and NEd/NRd = 0.350.

The strut is verified on the smaller of Nc,Rd and Nb,Rd, and the calculation prints both, with which one governs. The design load is the largest strut load of the cut times γF.

12Verifying the steel to AISC 360-22LRFD or ASD; classes, flexure and shear, the column curve.

AISC 360-22, the second design code of the module, verifies each member against a nominal strength Rn from its provisions for compression, flexure and shear. The method is chosen on the Analysis settings stage: by LRFD the load is factored and the strength reduced by φ; by ASD the working load is compared with the strength divided by the safety factor Ω.

Ru≤φRn(LRFD),Ra≤RnΩ(ASD)(23)

The loads that come off the envelope are taken as the lateral earth pressure H of ASCE 7-22. It enters the LRFD combinations with a load factor of 1.6 where it adds to the load effect — as it always does on the bracing of a cut — and the ASD combinations with 1.0:

Pu=γHP,Mu=γHM,γH=1.6Pa=P,Ma=M(24)γH is on the Analysis settings stage for LRFD; ASD has no factor to choose
Table 6. Resistance and safety factors — AISC 360-22
Limit statesectionφ (LRFD)Ω (ASD)
compressionevery section0.901.67
flexureevery section0.901.67
shearrolled I-shape, h/tw ≤ 2.24√(E/Fy)1.001.50
shearchannels, hollow and box sections0.901.67
The factors are fixed by the specification, so the stage asks for none of them. Steel is taken with E = 200 GPa (29,000 ksi).

The elements of the section

Before a strength is computed, the flange and the web are classed by the slenderness of their walls, b/t or h/tw, against limits in √(E/Fy) — Table B4.1a in compression (nonslender or slender), Table B4.1b in flexure (compact, noncompact or slender):

Table 7. Width-to-thickness limits for the parts the module classifies
Partratiocompression λrflexure λpflexure λr
flange of a rolled Ibf/2tf0.56√(E/Fy)0.38√(E/Fy)1.0√(E/Fy)
flange of a channelbf/tf0.56√(E/Fy)0.38√(E/Fy)1.0√(E/Fy)
web of an I or a channelh/tw1.49√(E/Fy)3.76√(E/Fy)5.70√(E/Fy)
wall of a rectangular hollow sectionb/t1.40√(E/Fy)1.12√(E/Fy) flange, 2.42√(E/Fy) web1.40√(E/Fy) flange, 5.70√(E/Fy) web
round tubeD/t0.11 E/Fy0.07 E/Fy0.31 E/Fy
h = d − 2(tf + r) for a rolled section; the flat width of a hollow section is taken as the outside dimension less 3t. A strut with a slender element needs its effective area, and a wale with a slender element its effective section: both are reported and not designed.

Flexure: the sheeting and the wale

The wale is held laterally along its whole length by the sheeting its compression flange bears against, so Lb = 0 and lateral-torsional buckling is not a limit state. A compact section reaches its plastic moment:

Mn=Mp=FyZx(compact),φbMn≥MuorMnΩb≥Ma(25)

A noncompact flange of an I-section falls linearly from Mp to 0.7FySx as λ goes from λp to λr:

Mn=Mp−(Mp−0.7FySx)λ−λpfλrf−λpf(noncompact)(26)

A rectangular hollow section with a noncompact flange or web and a round tube follow provisions of their own. The sheeting and a wale given by their modulus are taken at the strength the modulus claims — Fy Z for a compact section, Fy S at first yield.

Shear in the wale

Vn=0.6FyAwCv,Aw=dtw(I-shape),Aw=2ht(hollow section)(27)

For a rolled I-shape with h/tw ≤ 2.24√(E/Fy) — every rolled section of the catalogue — Cv1 = 1.0 and φv = 1.00. Unlike EN 1993-1-1, AISC 360-22 does not reduce the moment strength of a beam for the shear it carries.

The strut: flexural buckling

Fe=π2E(Lc/r)2,Fcr=0.658Fy/FeFyifLcr≤4.71EFy,elseFcr=0.877Fe(28)
Pn=FcrAg,φcPn≥PuorPnΩc≥Pa(29)

One column curve serves every section — there are no buckling curves to choose, as there are in EN 1993-1-1. r is the smaller radius of gyration and Lc = K Lb the effective length, the clear width of the cut unless a king post or a cross brace holds the strut in between. The specification recommends Lc/r ≤ 200.

0501001502002500.00.20.40.60.81.0slenderness Lc​/rFcr​/Fy​Fe​/Fy​ (Euler)4.71√(E/Fy​) = 127inelastic:Fcr​ = 0.658Fy​/Fe​​ Fy​elastic:Fcr​ = 0.877 Fe​Fy​ = 275 MPa, E = 200 GPathe strut of the cutthe module opens with:Lc​/r = 81.1, Fcr​/Fy​ = 0.682
Figure 11. The column curve of AISC 360-22: inelastic buckling up to Lc/r = 4.71√(E/Fy), elastic beyond it at 0.877 of the Euler stress. Marked: the strut of the cut the module opens with.

For the cut the module opens with, the largest strut load is P = 271.26 kN. By LRFD Pu = 1.6 × 271.26 = 434.02 kN. With the strut the Sections stage starts from — Ag = 65.7 cm², r = 7.40 cm, Fy = 275 MPa — over Lc = 6.00 m the chain runs Lc/r = 81.1, Fe = 300.3 MPa, Fcr = 187.4 MPa, Pn = 1231.5 kN and φcPn = 1108.3 kN: Pu/φcPn = 0.392. By ASD the same strut has Pn/Ωc = 737.4 kN against Pa = 271.26 kN, a ratio of 0.368.

LRFD and ASD share the nominal strengths — the same classification, the same column curve, the same flexural limit states; only the margin is placed differently. With γH/φ = 1.6/0.90 = 1.78 against Ω = 1.67, LRFD asks a little more of the bracing than ASD does.

13The envelopes against real cutsBoston and Chicago, from the case studies.

The procedure is straightforward once an envelope is chosen, and choosing it is the hard part. Two instrumented cuts show how much judgment that takes.

Boston: the subway extension

Three test sections of a subway extension in Boston were instrumented. At test section B the cut was 58 ft (17.7 m) deep, through some 26 ft of gravel, sand, silt and clay fill, an organic silt, and a layer of sand and gravel with some clay over rock, braced at five levels with the struts 12 ft (3.66 m) apart. The envelopes exist for sand and for clay only, so the profile was bracketed with both. As a sand, with γ = 114 lb/ft³ and Ka = 0.26, σa = 0.65γHKa = 1117 lb/ft² (53.5 kPa); as a soft to medium clay with c = 890 lb/ft², σa = γH[1 − 4c/γH] = 3052 lb/ft² (146.1 kPa).

0200400600kipS-1S-2S-3S-4S-5computed, envelope for sandcomputed, envelope for claymeasured
Figure 12. The strut loads of test section B: computed from the envelope for sand and from the envelope for clay, and measured on site.

The measured loads differed widely from both predictions, mainly because the soil parameters were uncertain, and the loads the struts were actually designed for were well above those measured.

Chicago: the National Plaza

The south half of the National Plaza in Chicago needed a cut 70 ft (21.3 m) deep, braced at six levels, through 13 ft of sand fill over clays of increasing strength. To compare Peck’s envelope with the wale loads measured, the profile is converted to one equivalent clay: the clays first by thickness, then that clay together with the sand. The module’s own conversion gives cav = 729 lb/ft² (34.91 kPa) and γav = 125.8 lb/ft³ (19.76 kN/m³); γavH/cav = 12.08 is above 4, so the soft-to-medium envelope applies with σa = 5889 lb/ft² (282.0 kPa).

Peck’s envelope turned out to give about 1.8 times the pressure actually observed. That is not surprising: an envelope drawn around many cuts in many places is conservative for any one of them, and experience with similar soils on a site can reduce the over-design substantially.

The module takes the envelope as the source gives it and does not scale it to local experience; the envelope, the stiff-clay coefficient and the conversion coefficients are all shown in the calculation, so the engineer can judge the result against what is known of the site.

14The base of a cut in clay: bottom heaveTerzaghi’s column, revised.

A braced cut in clay can fail without the bracing failing at all: the weight of the soil beside the cut, and any surcharge, squeezes the clay below the base inward and up. The column of soil of width B′ beside the cut as a load on the clay at the level of the base, with an ultimate bearing capacity cNc (Nc = 5.7 for a perfectly rough base), relieved by the shear cH along its side:

FS=cNcγH+q−cHB′(30)
FS=cNc(1+0.2B′L)γH+q−cHB′,B′=T or B2, the smaller(31)for a cut of limited length L

B is the width of the cut, H its depth, T the thickness of clay below the base and q a surcharge beside the cut. The equation is revised in three ways: the shear along the side of the soil column is an added resistance rather than a relief of the load; the strip of width B″ at the base acts as a footing loaded upward; and, the base of the cut being free of restraint, Nc is 5.14, the value for a smooth footing:

FS=5.14c(1+0.2B′′L)+cHB′γH+q,B′=T(T≤B2),B′=B2(T>B2),B′′=2B′(32)
hard stratumqγH + qbase heavesB′B″ = √2 B′HTBB′ = T when the hard stratum is shallow, otherwise B/√2 — the width of soil that can be drawn in under the wall
Figure 13. Bottom heave: the overburden γH + q beside the cut pushes the clay below the base up into the excavation along a surface of width B′, limited by the hard stratum (T) or by the width of the cut (B/√2).

Checked against compiled cases of bottom heave, the revised equation gives factors of safety close to 1 for the cuts that failed completely and slightly above 1 for those that failed partly or nearly. A factor of safety of about 1.5 is recommended.

In homogeneous clay, when the factor falls below 1.5 the sheet piles are driven deeper, usually not more than B/2 below the base. The buried length then carries a force per unit length, which the source takes from published design charts:

P=0.7(γHB−1.4cH−πcB)(d>0.47B),P=1.5d(γH−1.4cHB−πc)(d<0.47B)(33)
The module uses the revised equation, with the required factor of safety on the Base heave stage (1.5 by default) and L blank for a long cut. The embedment of the sheeting is drawn but not credited — the wall is taken to end at the base, the safe side — and the force on a buried length is not computed. In a layered profile the 5.14c term takes the clay the base sits in, and the side term the strength averaged over the height beside the cut. The check itself can be left out on the same stage; the report then states that the base has not been checked against bottom heave.

15What the method assumes, and what it leaves outThe limits of the envelopes; wall movement and settlement.

What the calculation rests on, in one place:

  • The envelopes were assembled from cuts deeper than about 6 m; a shallower cut lies outside the data they summarise, and the module notes it.
  • They assume the water table below the base. The clay envelopes are total-stress constructions, so a table above the base is noted, not added.
  • An envelope is not the real pressure on the wall at any one time. It gives safe strut loads; it does not give the deflection of the wall.
  • The sheeting is idealised as a chain of determinate beams, hinged at the chosen strut levels; its own stiffness and the embedment below the base are not used.
  • The wales are simple spans between the struts, the safe side for a wale that runs continuously past several of them.
  • The steel is verified to EN 1993-1-1; class-4 sections are not designed, and the lateral-torsional buckling of the wales is not checked.

Lateral yielding of the wall, and settlement

Some lateral movement of the sheeting is to be expected. How much depends above all on the time between digging a lift and installing its wales and struts. Sheet piles are often driven below the base to reduce the yielding in the last stages of the excavation: with deep clay below the cut this helps considerably while γH/c is below about 6, but little when it is about 8, when large yielding and even the collapse of the bracing can follow; where a hard layer lies below the clay, the piles should be embedded in it.

The yielding of the wall lets the ground around the cut settle — the ground loss. Published curves give the settlement to be expected in different soils, a general guide only. From about 150 case histories, mostly in soft clay, the largest settlement averages about 1.1 per cent of the depth of the cut in soft clay, 0.2 per cent in stiff clay and 0.3 per cent in non-cohesive soils, usually within half the depth of the cut from the wall; the largest horizontal deflection of the wall in soft clay averaged about 0.9 per cent of the depth.

The module does not compute the deflection of the wall or the settlement behind it. These figures are for orientation; where they matter, a soil–structure analysis of the stages of the excavation is the tool.

16Worked exampleEvery number computed by the module’s own code.

The cut in clay is the one this module opens with: 7 m deep in a clay with γ = 18 kN/m³ and c = 35 kPa, struts at 1, 3.5 and 6 m below the surface and 3 m apart in plan, and an allowable stress of 170 MPa.

The envelope

γHc=18×735=3.60≤4⇒σa=0.3γH=0.3×18×7=37.80kPathe stiff-clay envelope, ramping over the top and bottom 1.75 m
zc=2cγ=2×3518=3.89m>1.00mthe first strut is above the tension-crack depth

The strut loads

With the hinge at B the wall is two pieces, symmetric about B:

A=54.02,B1=45.21,B2=45.21,C=54.02kN/m
PA=162.06,PB=271.26,PC=162.06kN

The sheet piles

The shear force changes sign 1.196 m from B, where B1 has been used up by the pressure; the moment there is the largest on the wall:

Mmax=27.03kN·m/m,S=27.03170000=15.9×10−5m3/m

The wales

Mmax=(B1+B2)s28=90.42×328=101.72kN·m
S=Mmaxσall=0.598×10−3m3

To EN 1993-1-1 in S275, with γF = 1.35, the same moments ask for Wreq = 133 cm³/m for the sheeting and 499 cm³ for the wale at level B.

With the hinge convention Interior struts, γF = γM0 = 1 and fy typed as 170000 kPa — the allowable stress of this example — the module prints exactly these numbers: its Wreq = γF M γM0 / fy is the S of the equations above. It opens with γF = 1.35 and S275, with the same strut loads and moments.

Bottom heave

The base is checked for a cut 4 m wide and 15 m long, 6 m deep in a clay with γ = 17 kN/m³ and c = 40 kPa, over a hard stratum 2 m below the base. B/√2 = 2.83 m is more than T, so B′ = T:

B′=2.00m,B′′=2B′=2.83m
FS=5.14×40(1+0.2×2.8315)+40×62.0017×6=3.27above the 1.5 recommended

The same cut, verified to EN 1993-1-1

The module opens with this cut, with γF = 1.35, S275 for every member, and the members the Sections stage starts from: the sheeting with Wel = 600 cm³/m, the wale with Wel,y = 1000 cm³, and the strut with A = 65.7 cm² and I = 3598 cm⁴ on curve c. Pressing Analyze gives:

ηsp=MEdMc,Rd=1.35×27.03600×10−6×275000=0.221the sheeting; it asks for Wreq = 133 cm³/m
ηwale=1.35×101.721000×10−6×275000=0.499the governing wale; it asks for Wreq = 499 cm³
NEd=1.35×271.26=366.20kN,λ‾=0.9339,χ=0.5791
ηstrut=NEdNb,Rd=366.201046.36=0.350the largest strut, over Lcr = 6.00 m, the clear width of the cut

All three are below 1, so each member carries what is asked of it; the Summary of the report shows the same three verdicts, with the base check beside them.

17ReferencesSources.

  • Das, B.M. (2016). Principles of Foundation Engineering, 8th ed. Cengage Learning — Chapter 15, “Braced Cuts”; this page follows it.
  • Peck, R.B. (1943). Earth pressure measurements in open cuts, Chicago (Ill.) subway. Transactions, ASCE, 108, 1008–1058.
  • Peck, R.B. (1969). Deep excavation and tunneling in soft ground. Proceedings, 7th International Conference on Soil Mechanics and Foundation Engineering, Mexico City, State-of-the-Art Volume, 225–290.
  • Terzaghi, K. (1943). Theoretical Soil Mechanics. Wiley, New York.
  • Lambe, T.W. (1970). Braced excavations. Proceedings, Specialty Conference on Lateral Stresses in the Ground and Design of Earth-Retaining Structures, ASCE, 149–218.
  • Swatek, E.P., Asrow, S.P. and Seitz, A. (1972). Performance of bracing for deep Chicago excavation. Proceedings, Specialty Conference on Performance of Earth and Earth-Supported Structures, ASCE, Vol. 1, Part 2, 1303–1322.
  • Bjerrum, L. and Eide, O. (1956). Stability of strutted excavations in clay. Géotechnique, 6(1), 32–47.
  • Chang, M.F. (2000). Basal stability analysis of braced cuts in clay. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 126(3), 276–279.
  • U.S. Department of the Navy (1971). Design Manual — Soil Mechanics, Foundations, and Earth Structures, NAVFAC DM-7, Washington, D.C.
  • Moormann, C. (2004). Analysis of wall and ground movements due to deep excavations in soft soil based on a new worldwide database. Soils and Foundations, 44(1), 87–98.
  • EN 1990:2002, Eurocode — Basis of structural design. CEN.
  • EN 1993-1-1:2005, Eurocode 3: Design of steel structures — Part 1-1: General rules and rules for buildings. CEN.
  • EN 1993-5:2007, Eurocode 3: Design of steel structures — Part 5: Piling. CEN.
  • ANSI/AISC 360-22, Specification for Structural Steel Buildings. American Institute of Steel Construction, Chicago, 2022.
  • ASCE/SEI 7-22, Minimum Design Loads and Associated Criteria for Buildings and Other Structures. American Society of Civil Engineers, Reston, 2022.

From theory to your own design

Design your own braced cut in clay in minutes

  • The pressure envelope chosen from the stability number, and every strut load
  • Sheeting and wale sections, with the steel checked to EN 1993-1-1 or AISC 360-22
  • The factor of safety against bottom heave
  • A clean calculation report, saved as a PDF. Online: nothing to install, no dongle, on a phone, a tablet or a PC

More on braced cuts