F
The design of vibro replacement
by Heinz J Priebe, Keller Grundbau GmbH,
Kaiserleistr. 44, 63067 Offenbach.
Vibro replacement is an accepted method for
subsoil improvement at which large-sized
columns of coarse backfill material are
installed in the soil by means of special depth
vibrators. The performance of this composite
jystem consisting of stone columns and soil,
O is not determinable by simple investigation
methods like soundings, and therefore such
methods are not suitable for design purposes.
However, theoretically, the efficiency of vibro
replacement can be reliably evaluated. The
method elaborated on a theoretical basis and
described in this contribution, is easy to
survey and adaptable to different conditions
due to the separate consideration of
significant parameters. Practically, it
comprises design criteria for all frequently
1
occurring applications.
5 surrounding soil.
's,.
Introduction
Vibro replacement is part of the deep vibratory compaction
*echniques whereby loose or soft soil is improved for building
I
~ u r p o s e sby means of special depth vibrators. These. techniques
as well as the equipment required is comprehensively described
elsewhere'.
Contrary to vibro compaction which densifies noncohesive
soil by the aid of vibrations and improves it thereby directly,
vibro replacement improves non compactible cohesive soil by
the installation of load bearing columns of well compacted,
coarse grained backfill material.
The question to what extent the densiry of compactible soil
will be improved by vibro compaction, depends not only on the
parameters of the soil being difficult to determinel but also on
the procedure adopted and the equipment provided. However,
the difficulry of a reliable prognosis is balanced by the fact that
the improvement achieved can be determined easily by
soundings.
With vibro replacement the conditions are more or less
revers. Considerable efforts only like large-scale load tests can
prove the benefit of stone columns. However, a reliable conclu-
sion can be drawn about the degree of improvement which
results from the existence of the stone columns only without
any densification of the soil between. This is possible because
the essential parameters attributable to the geometry of the
layout and the backfill material can be determined fairly well.
In such a prognosis the properties of the soil, the equipment
and the procedure play an indirect role only and that is mainly
in the estimation of the column diameter.
Basically, the design method described was developed some
20 years ago and published already'. However, in the meantime
it came to several adaptions, extensions and supplements which
justify a new and comprehensive description of the method.
Nevertheless, the derivation of the formulae is renounced with
reference to literature.
It must be emphasised that the design method refers to the
improving effect of stone columns in a soil which is otherwise
unaltered in comparison to the initial state. In a first step a
factor is established by which stone columns improve the
performance of the subsoil in comparison to the state without
columns. According to this improvement factor the
deformation modulus of the composite system is increased
respectively settlements are reduced. All further design steps
refer to this basic value.
In many. practical
. cases the reinforcing effect of stone
columns installed by vibro replacement is superposed with the
densifying effect of vibro compaction, ie the installation of
stone columns densifies the soil between. In these cases, the
densificadon of the soil has to be evaluated and only then - on
the basis of soil data adapted correspondingly - the design of
vibro replacement follows.
Determination of the basic improvement
factor
-
The fairly complex system of vibro replacement allows a more
or less accurate evaluation only for the well defined case of an
unlimited load area on an unlimited column grid. In this case a
unit cell with the area A is considered consisting of a single
column with the cross section Ac and the attributable
Furthermore the following idealized conditions are assumed:
*The column is based on a rigid layer
*The column material is uncompressible
*The bulk density of column and soil is neglected
Hence, the column can not fail in end bearing and any
settlement of the load area results in a bulging.of the column
which remains constant all over its length.
The improvement of a soil achieved at these conditions by
the existence of stone columns is evaluated on the assumption
that the column material shears from the beginning whilst the
surrounding soil reacts elastically. Furthermore, the soil is
assumed to be displaced already during the column installation
to such an extent that its initial resistance corresponds to the
liquid state: ie the coefficient of earth pressure amounts to
K= 1. The result of the evaluation is expressed as baslc
lmprovemrnt factor n,,.
A poisson's ratio of ps=113 which is adequate for the state of
final settlement in most cases, leads to a simple expression.
.
GROUND EN(3INEERING DECEMBER. 19'45
*.
TOP: Figure 1. Design c h a r t
for vibro replacement.
BOTTOM: Figure 2.
Consideration of column
compressibility. 5--- -

T h e relation between the

improvement factor no, the
reciprocal area ratio NA, and the
friction angle of the backfill
material cp, which enters the
derivation, is illustrated in the well
known diagram of Figure 1.

Consideration of the 1 -,
1 2 3 4 5 6 7 8 9
l

10
column compressibility
The compacted backfill material Area Ratio NAc
of the columns is still
compressible. Therefore, any load
.auses settlements which are not
(7 ~onnectedwith bulging of the
columns. Accordingly, in the case . .- - - ..- ..... . . .. .

of soil replacement where the area

ratio amounts to NAc= 1, the
actual improvement factor does
not achieve an infinite value as
determined theoretically for non
compressible material, but it
coincides at best with the ratio of
the constrained moduli of column
material and soil. In this case for
compacted backfill material as well
as for soil, a constrained modulus
is meant as found by large scale
oedometer tests. Unfortunately, in
many cases soundings are carried
out within the columns and wrong
conclusions about the modulus are 1 2 3 4 6 8 10 20 30 40 60 80 loo
drawn from the results which are
sometimes only Constrained Modulus Ratio DdDs
. very. moderate.

I It is relatively easy to determine

at which area ratio of column cross iection and grid size
:AcIA), the basic improvement factor no corresponds to the
ratio of the constrained moduli of columns and soil DcIDs. For
example, at ps=113 the lower positive result of the following
expression (with no = DcIDs delivers the area ratio (Ac/A),
concerned.
As an approximation, the compressibility of the column
material can be considered in using a reduced improvement
factor no which results from the formula developed for the basic
improvement factor no wheri the given reciprocal area ratio
MAc is increased by an additional amount of A(A/A,.-J.
In using the diagram in Figure 1, this procedure corresponds to
such a shifting of the origin of the coordinates on the abscissa
which denotes d e area ratio A/A,that the improvement factor
32 n, to be drawn from the diagram, begins with the ratio of the
constrained moduli and not with just an infinite value. The
additional amount on the area ratio A ( N A 3 depending on the
ratio of the constrained moduli Dc/Ds can be readily taken
from the diagram in Figure 2.
Consideration of the overburden
T h e neglect of the bulk densities of columns and soil means
that the initial pressure difference between the columns and the
soil which creates bulging, depends solely on the distribution of
the foundation load p on columns and soil, and that it is
constant all over the column length. As a matter of fact, to the
external loads the weights of the columns W, and of the soil
Ws which possibly exceed the external loads considerably, has
to be added. Under consideration of these additional loads the
initial pressure difference decreases asymptotically and the
bulging is reduced correspondingly. In other words, with
increasing overburden the columns are better supported
laterally and, therefore, can provide more bearing capacity.
Since the pressure difference is a linear parameter in the
derivations of the improvement factor, the ratio of the initial
pressure difference and the one depending on depth -
expressed as depth factor fd - delivers a value by which the
improvement factor n, increases to the final improvement
factor n,=fd x n , on account of the overburden pressure. For
example, at a depth where the pressure difference amounts to
50% only of the initial value, the depth factor comes to f, = 2 .

GROUND ENGINEERING DECEMBER - 1995