7058 S The influence of geological features
of clay deposits on the design and
performance of sand drains
P. W. ROWE, DSc.-MICE.
The following is a summary of the Paper to be discussed at an ordinary meeting
at 5.30 pm. on Tuesday, 7 May, 1968. The full Paper is available as a supple-
ment to the Proceedings (members free; non-members, 1Os.).
Sand drains have been usedto accelerate the consolidation of soft clay founda-
tions of embankmentsfor over 30 years. During this period manyobservations
of fieldrates of settlement and an increasing number of pore-pressure measure-
ments have led to inconclusive opinions as to the eficiency of sand drain
installations.
Following a review of the literature, the results of recent field measurements
of rates of settlement and pore pressure dissipation of clay foundation for
earth embankments with and without sand drain installations are compared
with predictions based on laboratory and in situ tests.
It is shownthat the real drainage behaviour of a deposit as a whole depends
on the geological details of its formation. Quite small layers, Geins of silt
along fissures, or organicinclusions can transform the permeability of the mass
compared with that of small samples. The conclusions are as follows.
The identification tests, liquid limit, plastic limit and clay fraction, give no
indication of drainage properties of undisturbed strata, because the tests are
necessarily conducted on remouldedsamples after destruction of all the
essential geological features.
The 3 in. dia. oedometer test (and similarlythe 13 in. dia. triaxial consolida-
tion test) can givecompletelymisleading laboratory measurements of the
coefficient of consolidation for other than uniform clays, which are rarely to
be found.
On present information, consolidation samples 10 in. in dia. by 5 in. thick
appear to be large enough to represent the natural structure of the more
common non-uniform clay deposits, and with attention to the appropriate
drainage direction give fair agreement with in situ measurements and field
pore-pressure dissipation records at the same stress level.
In situ permeability measurements, coupled with laboratory measurements
of compressibility which are not very sensitive to sample size, give coefficients
of consolidation in fair agreement with field performance and large samples
at the same stresslevel.Since for design purposes in situ tests have to be
conducted at relatively low effective stress levels compared with those acting
later under an embankment, the measurement should becoupledwith at
Professor, Dept of Civil Engineering, University of Manchester.
465
Downloaded by [ York University] on [18/09/16]. Copyright © ICE Publishing, all rights reserved.
ROWE
least a few laboratory tests on large samples in order to ascertain the influence
of increase in effective stress. Soft organic clays containing rootlets, existing
under low stress levels, may beparticularly sensitive to later changes in effective
stress.
Variations in in situ and large sample test values will occur in depth with
variation in major stratification and all testing should be preceded by visual
examination of continuous cores which may be 2 in. in diameter. The drain-
age of the mass may be dominated by specific permeable layers, be they only
1% in. thick at 10 ft intervals.
Multi-fissuredfirmshale,derived from a coarsely layered clay,clays
coarsely layered with sand and silt, silts, and boulder clays having connected
silt, sand or gravel pockets or seams may be expected to show coefficients of
consolidation of the mass in excess of 1000 sq. ft/year and do not benefit from
sand drain installation. Such deposits, including those having coefficients
down to 500 sq. ft/year, respond to dewatering by vacuum well points.
Clays finely layered with thin silt sand seams and multi-fissured mudstone
having coefficientsof consolidation of 500 sq. ft/year do notrequire sand drains
for embankments of the order of 100 ft high constructed over a twoyear
period.
The above permeable type clays, which have often yielded low undrained
strengths in the past due to faulty sampling technique, and have therefore the
more often been considered to require assistance in consolidation, are the most
capable of consolidation by natural means.
Sand drains are effective in deposits with coefficients ofconsolidation of the
order of 100 sq. ft/year or less, as may apply to laminated lacustrine clay
deposits with only occasional silt dustings, organic, non-layered clays under
medium to higher effective stress, and uniform clays.
The drainage rates of boulder clays which may be sensibly uniform or may
contain permeable inclusions are likely to be particularly difficult to assess
on the basis of geological description owing to their erratic nature.
Where sand drains are tobe used,the driven mandrel method of construction
induces excess pore pressures during driving, increases settlement, and r e
moulds the clay around the drain. Its use may therefore be suited more to the
case of high banks where the induced pore pressures during driving are small in
comparison to those induced by the bank itself and to nearly uniform deposits
whose permeability is not greatly affected by remoulding, should drains be
decided upon, and especially if the driven mandrel offers marked economic
advantage. Construction should be preceded wherever possible by field trials
over areas with variable drain spacing and areas without drains.
466
Downloaded by [ York University] on [18/09/16]. Copyright © ICE Publishing, all rights reserved.
