Bowles On Strip Loads
Bowles On Strip Loads
Bowles On Strip Loads
(a) Case 1
(b) Case 2
Parallel
(C) Case 3
Figure 11-15
Procedures for location of point of application of Pa for (a) irregular backfill; (b) concentrated or
line load inside failure zone; (c) concentrated or line load outside failure zone (but inside zone ABC).
The author suggests the best solution for total wall force and point of application when
there are backfill loads of any type and location is to use one of the following:
1. If the backfill load is inside the Rankine zone, convert it to an equivalent surcharge over
the Rankine zone, then obtain the wall pressure and resultant using either the Coulomb or
Rankine equations.
2. If the backfill load is either inside or outside the Rankine zone, use the Coulomb or Rankine equations for the soil wedge with no backfill load. Next use the Theory of Elasticity
equations given in Sec. 11-13 to find the wall forces from the backfill loads. Then to find
total force and point of application use X P R X Pi and Ry = X Piyi-
Figure 11-16
Identification of terms used in the Boussinesq equation [Eq. (11-20)] for lateral pressure.
ar =
P \3rh
1 -2/x 1
2^ [ ^ - - R(RT7)\
(U 2 a)
where the several terms including 0, z, r, and R are identified on Fig. 11-16. This form of the
equation is particularly suitable for programming on a small calculator, since the point load P
is usually fixed with given x, y coordinates and we want to vary z to obtain the wall pressure
profile.
The computer programming of this equation allows one to solve any of the given backfill
surcharge loads of Fig. 11-17 defined as follows:
1. Point load. Use the equation in the given form.
2. Line load. Treat as either one load or a series of concentrated loads along a line of unit
width acting on unit areas.
3. Strip load. Treat as a series of parallel line loads acting on strips of some unit width.
4. Loaded area. Treat as a series of parallel line loads acting on strips of finite length.
We can easily analyze a constant uniformly loaded area (say, the interior part of an embankment) or one with a linear varying load (say, the embankment side slopes). In either of
these cases the loaded area is divided into strips with some load intensity q and some small
"unit" width B, on the order of 0.25 to 0.5 m. These strips are then subdivided into "unit"
areas of some length L also on the order of 0.25 to 0.50 m. These "unit" areas are treated as
a series of point loads of Q qBL acting at the center of each of the unit areas. The several
"unit" area contributions making up the total loaded area are then summed to obtain the total
lateral pressure acting at some point at the depth of interest (either in the soil or on a wall).
This is the procedure used in program SMBLPl (B-8) on your program diskette.
The general validity of using a form of Eq. (11-20) for surcharges was established in several publications, including the work of Spangler (1936), Spangler and Mickle (1956), Rehnman and Broms (1972), and others.
BxL.
Figure 11-17 Surcharge loads that can be used with the computer program SMBLPl (B-8) on your program
diskette. NSQW, NSQL = number of unit elements away from wall and parallel to wall, respectively, as used in
the computer program.
The early work of Spangler and Spangler and Mickle introduced an error into the general
application of the equations; however, that can be avoided by direct use of Eq. (11-20) and
an appropriate value for Poisson's ratio /JL.
When the work of Spangler was first published, he used /JL = 0.5 [and later in Spangler and Mickle (1956)], which substantially simplifies Eq. (11-20)but may not be correct.
Spangler's work consisted of trying to measure the lateral pressure against a 1.829 m (2.134
m total height) high X 4.572 m wide retaining wall with a constant stem thickness of 0.150
m. He used metal ribbons (since earth-pressure cells were not readily available in the early
1930s) and simply dumped a granular (WL = 17.5, Wp = 13.2%) backfill behind the wall
with no compaction at all to produce an extremely loose state. After a time, he had a truck
backed onto the loose backfill so that the rear wheels could simulate two concentrated loads.
To simulate a line load he laid a railroad crosstie parallel to the wall, onto which the rear
wheels (a single axle with dual tires) of a loaded truck were backed. Since the wall was only
4.572 m long and a railroad crosstie is about 3 m long, a strip model was not very likely to
have been produced.
From these efforts Spangler (both references) found that the measured lateral pressure was
about twice that predicted by Eq. (11-20) with /JL = 0.5. From the reported results, Mindlin
(1936a), in discussing Spangler's (1936) work suggested that the factor of 2 could be explained by a rigid wall producing the effect of a mirror load placed symmetrically in front of
the wall (Fig. 11-ISa). The author began looking at this problem more closely and decided
that the mirror load is not an explanation. As shown in Fig. 11-18&, a mirror load on a rigid
Rigid wall
(a) Mirrored loads on rigid wall.
Figure 11-18
wall would simply cancel the lateral shear stresses in the wall and certainly not double the
horizontal pressure. A flexible wall could possibly double the lateral pressure but would have
to be extremely flexible (and have the loads applied sequentially). Referring to Fig. 11-18Z?,
we see that the horizontal pressure ah produced by P0 (applied first) acting on a rigid area of
dydz would develop shear resistance r such that rt(2dz+2dy) = a^ When mirror load P0 is
applied, a second shear stress r would develop on the element but in the opposite direction, so
the shear stress would cancel and we would simply have on each side of the wall a horizontal
stress ah (not 2ah). If the wall is flexible and little shear stress develops, the element would
(if the loads were applied sequentially) displace laterally toward the -a direction to produce
a resisting soil stress on the a side of a'h = ah r(2dz + 2dy). Since this would become
locked in when P'o is applied, the stress on the left would become ah + a'h, and since a'h is
transmitted through the wall, the right side would also have the existing cr^+ transmitted a'h
value from load P0. If the loads were applied simultaneously the stresses would simply be
ah on each side (and not 2ah).
Mindlin got around this complication by inserting a statement that the wall was rigid but
could not carry shear. There is no such wall type known to the author.
Because the Spangler work was done in the early 1930s, it is difficult to speculate on
the cause of the high stresses except to note that the wall had rather finite dimensions. The
surcharge load was caused by a truck backing onto the backfill. When it stopped at the desired
position it would have produced an inertial force that was amplified because the fill was not
well compacted. The backfill probably was of limited extent, so that it is also possible some
type of arching occurred that increased the lateral pressure.
More recently Rehnman and Broms (1972) showed (using modern earth-pressure cells)
that when the soil behind the wall was dense the lateral pressure from point loads was much
less than when the soil was loose. They also found that gravelly backfill produced larger
lateral pressures than finer-grained materials. This observation indicates that both soil state
and Poisson's ratio are significant parameters.
0.3
0.42
0.33
0.50
0.35
0.54
0.40
0.67
0.45
0.82
0.50
1.00
0.60
1.50
Table 11-7 illustrates the case of a small retaining wall with a concentrated load at varying
distance, using a range of Poisson's ratio. This wall also includes the trial wedge solution for
the several load positions and the Rankine lateral pressure computed for no surcharge. From
this table several conclusions can be drawn:
1. The trial wedge method gives the Coulomb (or Rankine if S = 0) wall force for a horizontal backfill and correctly locates the failure surface using angle p measured from the
horizontal. For concentrated surcharges on the backfill, much larger wall forces are obtained than by any other method.
2. Poisson's ratio/x = 1 gives a substantial increase in wall pressure versus LL = 0.3 to 0.5.
A plane strain LL' = 1.00 may be possible for soil in a very loose state.
TABLE 11-7
Comparison of trial wedge and Boussinesq wall forces computed using Eq. (11-20).
Also shown is the Coulomb active pressure force.
2.217 m
Load position
from wall, m
Trial wedge, kN
0.3
0.6
1.0
1.3
1.6
2.0
2.3
2.6
3.0 m
65.0
65.0
65.0
65.0
64.8
63.4
61.6
59.3
55.5
fi = 0.3
0.5
0.7
1.0
Outside
Inside
Boussinesq
8.8
13.1
18.0
25.3
8.3
11.5
14.8
19.9
5.3
7.5
9.7
13.0
4.1
5.8
7.5
10.0
Inside
3.3
4.6
6.0
8.1
2.5
3.6
4.6
6.2
2.1
3.0
3.9
5.2
1.8
2.5
3.3
4.5
1.4
2.1
2.7
3.7
Outside
3. Concentrated loads well outside the Rankine zone contribute to Pa in the trial wedge case,
leading one to the opinion that the trial wedge is not correct, but conservative.
4. Since Eq. (11-20) gives small lateral pressures when the load is very close to the wall, this
result may mean either that the surcharge load is being carried downward by vertical wall
friction rather than by lateral pressure or that Eq. (11-20) is not valid for a load close to
such a massive discontinuity in the elastic half-space.
5. Wall pressures computed by Eq. (11-20) are rather small once the wall-to-load distance is
greater than the Rankine zone.