CEE 340 · Worked Example 8 · Slope Stability

Modified Bishop's Method, Start to Finish

One problem, taken completely apart: a 12.7 m two-layer slope with seepage, and a slip circle that cuts through both soils. We'll build the geometry, cut the mass into slices, do one slice entirely by hand, then let the equation chase its own tail until it converges.

Long-term (drained) Two layers Seepage Live simulator
01

The Problem

Compute the long-term factor of safety for the given failure surface using the Modified Bishop's Method.

Geometry
Slope height12.7 m
Slope angle1.5 H : 1 V
Circle centre12.0 m horiz, 10.0 m above crest
Upper layer8.6 m thick
Soil
Upper c' / φ' / γ30 kPa / 28° / 17.9 kN/m³
Lower c' / φ' / γ10 kPa / 32° / 18.2 kN/m³
Waterphreatic surface, seepage to toe
γw9.81 kN/m³
🍩
Before any equations, picture what is being asked. A curved wedge of hillside is poised to rotate outward about a point in the sky. Gravity is trying to spin it; friction and cohesion along the curved base are trying to stop it. The factor of safety is simply how much spare grip there is — resisting moment divided by driving moment. Everything below is bookkeeping to get those two numbers.
The one thing that makes this problem harder than a textbook single-layer slope: the slip circle passes through two different soils and beneath a water table. So c', φ' and u all change from slice to slice. That is exactly why the method of slices exists.
02

What "Long Term" Changes

That phrase is not decoration — it decides which strength parameters you are allowed to use.

Short term (undrained)

Right after excavation, water has not had time to move. Use total stresses with φ = 0 and cu. Pore pressures are unknown, so they never appear in the equation.

Long term (drained) — our case

Years later, pore pressures have reached steady seepage. Use effective stress with c' and φ', and you must subtract u explicitly. The problem gives c' and φ' for both layers, which is the giveaway.

Core Idea

"Long term" means drained, effective-stress analysis. For a cut slope this is usually the critical case: excavation unloads the ground, pore pressures start low and rise over time toward equilibrium, so the slope gets weaker as the years pass. A cut that stands fine on the day it is dug can fail a decade later.

03

Step 1 — Pin Down the Circle

Most students lose this problem here, before any soil mechanics happens. Put down coordinates and everything becomes arithmetic.

1
Origin at the toe. Let x run into the slope (uphill) and y run up. Toe = (0, 0).
2
Locate the crest. The slope is 1.5 H : 1 V and 12.7 m tall, so it runs back \( 1.5 \times 12.7 = 19.05\text{m} \). Crest = (19.05, 12.7).
3
Locate the centre. It is given as 12.0 m horizontally and 10.0 m above the crest, so \( O = (19.05-12.0,\; 12.7+10.0) = (7.05,\; 22.70) \). Note x = 7.05 m sits directly above the slope face — exactly where you would expect a slip-circle centre.
4
Get the radius from the exit point. The surface daylights at the toe, so \( R = \sqrt{7.05^2 + 22.70^2} = 23.77\text{m} \).
5
Find where it re-enters the ground. Setting y = 12.7 on the circle gives \( x = 7.05 + \sqrt{23.77^2-10^2} = 28.61\text{m} \) — that is 9.6 m behind the crest, out on the flat top.
6
Check the depth. The lowest point of the arc is \( y = 22.70-23.77 = -1.07\text{m} \), i.e. about a metre below the toe. The layer boundary is at \( 12.7-8.6 = 4.10\text{m} \), and the arc crosses it at x = 21.85 m.
Step 6 is the moment the problem reveals itself. From the toe out to x = 21.85 m the sliding base sits in the lower layer (c' = 10, φ' = 32°); only the last stretch runs through the upper layer (c' = 30, φ' = 28°). If both soils weren't involved, the problem would not have bothered giving you two sets of properties. Always run this check — it tells you whether you have read the geometry correctly.

The Set-Up, to Scale

Toe at the origin, crest 19.05 m back and 12.7 m up, centre O floating 22.7 m above the toe line. The arc leaves the toe, dips below it, crosses into the upper layer at x = 21.85 m, and re-emerges 9.6 m behind the crest.

04

Step 2 — Cut It Into Slices

🥪
Why slice at all? Because nothing about this wedge is uniform. Its height changes, the soil under it changes, the water pressure changes, and the base tilts differently everywhere. There is no single "average" that works. So we cut it into strips narrow enough that inside one strip everything is near enough constant — one height, one soil, one pore pressure, one base angle — solve each strip, and add them up. It is a definite integral performed with a bread knife.
Per slice you need

Six numbers

Mid-point x, height h, weight W, base angle α, base length ℓ, and pore pressure u. Plus which soil the base sits in.

Weight, carefully

Split at the boundary

A slice straddling the layer boundary is part 17.9 and part 18.2 kN/m³. Compute each piece separately — \( W = b(\gamma_1 h_1 + \gamma_2 h_2) \).

Strength, differently

Base only

c' and φ' come from whichever soil the base passes through — not the soil above it. Shearing happens on the base.

Core Idea

Weight is about everything stacked above the base; strength is about what the base is made of. Mixing these up is the single most common slicing error. A slice can be 90% stiff clay by volume and still shear through weak sand because that is where its base lies.

Sign convention that trips everyone

\( \sin\alpha = (x_m - x_O)/R \). Slices uphill of the centre get positive α and drive the failure; slices near the toe, downhill of the centre, get negative α and resist it. In the table below slice 1 has α = −13.7° and contributes −16.6 kN/m — it is holding the slope back, and the arithmetic says so on its own.

05

Step 3 — One Slice, in Full

Ten slices done ten times is not ten times the learning. Here is slice 7 from end to end; every other slice is the same six lines with different numbers.

Slice 7 of 10

b = 2.861 m, centred at x = 18.60 m

1
Ground surface. x = 18.60 m is still on the slope face (face runs to 19.05), so \( y_g = 18.60/1.5 = 12.40\text{m} \).
2
Base of the slice. \( y_b = 22.70-\sqrt{23.77^2-(18.60-7.05)^2} = 22.70 - 20.77 = 1.93\text{m} \). Slice height \( h = 12.40-1.93 = 10.47\text{m} \) — the deepest slice in the problem.
3
Weight, split at y = 4.10 m. Upper layer thickness \( 12.40-4.10 = 8.30\text{m} \); lower layer \( 4.10-1.93 = 2.17\text{m} \). \( W = 2.861\left[17.9(8.30) + 18.2(2.17)\right] = 2.861(148.6+39.5) = 538\text{kN/m} \)
4
Base angle. \( \sin\alpha = (18.60-7.05)/23.77 = 0.486 \Rightarrow \alpha = 29.07^\circ \). Base length \( \ell = b/\cos\alpha = 2.861/0.874 = 3.27\text{m} \).
5
Pore pressure. The phreatic surface here sits about 1.9 m above the base, so \( u \approx 9.81(1.9) = 18.6\text{kPa} \).
6
Which soil? The base is at y = 1.93 m, below the boundary at 4.10 m → lower layer: c' = 10 kPa, φ' = 32°. Note the slice is mostly upper-layer soil by volume, but it shears through the lower one.
7
Its share of the driving force. \( W\sin\alpha = 538 \times \sin 29.07^\circ = 262\text{kN/m} \) — the largest single contribution in the problem.
Slice 7: W = 538 kN/m, α = 29.1°, ℓ = 3.27 m, u = 18.6 kPa, c' = 10 kPa, φ' = 32°. Repeat nine more times and Step 4 is done.
06

Step 4 — The Driving Force

The denominator is the easy half: it needs no iteration and no strength parameters at all. Just add up W sinα.

Result
\[ \sum W\sin\alpha = 1199\ \text{kN/m} \]

Multiply by R and you have the driving moment about O. Since R divides out of both sides of the factor-of-safety equation, we never actually need it.

Read the α column, not just the total. It climbs from −13.7° to +57.9°. The first two slices sit downhill of the centre and subtract nearly 40 kN/m — the toe is buttressing the slope. This is precisely why excavating at the toe of a marginal slope is so dangerous: you remove the slices that were helping you.
07

Step 5 — Why Bishop Chases Its Tail

Modified (Simplified) Bishop
\[ FS = \frac{\sum \dfrac{c'b + (W-ub)\tan\phi'}{m_\alpha}}{\sum W\sin\alpha} \]
\[ m_\alpha = \cos\alpha + \frac{\sin\alpha\,\tan\phi'}{FS} \]
Spot the problem

FS appears on both sides. It is in the answer and it is buried inside mα in every single term. There is no algebra that isolates it — the equation is implicit, so it must be solved by iteration.

🔄
It is a thermostat, not a formula. You guess a temperature, the system responds, and you adjust toward what it tells you. Guess FS, compute what FS that implies, feed that back in, repeat. Each pass overshoots less than the last, and after three or four rounds the number stops moving. That settled value is the answer — not any single pass.
Core Idea

mα is where Bishop earns its accuracy. The Ordinary Method simply ignores the forces between slices; Bishop keeps horizontal equilibrium, and mα is exactly the correction that bookkeeping introduces. That single term is the entire difference between the two methods — and it is worth roughly 7% here.

08

Step 6 — Iterate to Convergence

Start from the Ordinary Method answer — it is free, and always a little low, which makes it an excellent first guess.

Four passes and it is done. The first correction is large (+0.11), the second tiny (+0.015), the third negligible. Bishop's method converges fast because mα depends only weakly on FS. If your iteration ever diverges, it is almost always because mα has gone near zero or negative in a steep slice — a sign the trial circle is unreasonable, not that the arithmetic is wrong.
09

The Live Simulator

The whole solution, rebuilt on every keystroke. Click any slice in the drawing to see its numbers. Change the slice count and watch the answer converge; switch the water off and watch it jump.

10

The Answer & Three Sanity Checks

Answer
\[ FS_{\text{long term}} \approx 1.9 \]

Modified Bishop, converged, using 20+ slices. With the 10 slices tabulated above you get 1.887; refining to 20 gives 1.906. Quote it as 1.9 — the geometry is read off a drawing and the phreatic surface is approximate, so three decimals would be false precision.

Never report a factor of safety without testing it. Three checks, each taking seconds:

Check 1 — against OMS

1.79 vs 1.90

The Ordinary Method gives 1.79 — about 6–7% lower. Bishop should always come out higher, by roughly 5–20%. If yours came out lower, the iteration or a sign is wrong.

Check 2 — slice count

10 → 20 → 40

1.887, 1.906, 1.901. Changes in the second decimal only, so the answer has converged. If it were still marching in one direction you would need more slices.

Check 3 — turn the water off

1.90 → 2.14

Dry, the same slope gives 2.14. Water costs 11% of the factor of safety. It must move in that direction, and by a plausible amount.

And one thing this answer is not. FS = 1.9 belongs to this circle only. The question handed us the failure surface; a real analysis would search hundreds of circles and report the lowest. That is what the critical circle search in the slope stability chapter does. Never present a single-circle result as "the" factor of safety of a slope.
11

What to Carry Away

"Long term"drained → c', φ', subtract u
Geometry firstcoordinates before any soil mechanics
Depth checkwhich layers does the arc actually cut?
Weighteverything above the base, layer by layer
Strengthfrom the soil at the base only
α sign(xm − xO)/R; toe slices resist
DenominatorΣW sinα, no iteration needed
mαcosα + sinαtanφ'/FS — the implicit bit
First guessuse the OMS answer
Convergence3–4 passes is normal
Bishop vs OMS5–20% higher, always
One circleis not the slope's FS
The Method in One Sentence

Cut the sliding mass into strips, work out what gravity asks of each strip and what its base can supply, add both columns up, and — because the strength term secretly depends on the answer — guess, recompute, and repeat until the number stops moving.