Soil consolidation - From theory to practice

Consolidation - De la theorie a la pratique

Carl B. Crawford - Department of Civil Engineering, University of British Columbia, Vancouver, B.C., Canada

ABSTRACT: The theory o f consolidation seldom, if ever, applies in full scale loading of the natural ground. Consequently, engineers
use more judgement than theory in the prediction o f consolidation settlements. Judgement is usually based on the observation o f full
scale performance o f various kinds o f structures resting on the ground. This paper reviews several case studies to illustrate the
importance o f secondary-type consolidation in the ultimate settlement o f structures on some inorganic soils.

RESUME: La theorie de la consolidation etant pratiquement inapplicable au cas de chargements a echelle reelle du terrain naturel; les
ingenieurs ont recours au judgement plutot qu'a la theorie dans la prediction du tassement des sols. Le judgemnet est en general basee
sur 1'observation de la performance de differents types de structures reposant sur le sol. Cet article passe en revue plusieurs cas d'etude
afin d'illustrer l'importance de la consolidation secondaire sur le tassement ultime des structures sur certains sols inorganiques.

1. INTRODUCTION records o f performance with associated predictions made only

after the actual settlements are known.
The basic understanding o f soil consolidation began with The observations o f Terzaghi are as relevant today as they
Terzaghi's laboratory tests on soils and his development o f the were thirty years ago. This paper summarizes several case
theory o f consolidation more than seventy years ago. This studies that illustrate the difficulties o f making predictions based
fundamental work, first published in English by Terzaghi (1926) simply on theoretical analyses and the great value o f field
in a series o f articles in Engineering News-Record, was more observations to engineering practice.
than an explanation o f consolidation; it was the first statement o f Some o f the projects show primary consolidation occurring
the principle o f effective stresses which has dominated soil very rapidly followed by an equal amount o f secondary
mechanics research and practice ever since. consolidation and others where almost all o f the consolidation
Terzaghi never expected the theory o f consolidation to occurs at a constant effective stress (commonly considered to
provide the basis o f engineering design because the assumptions represent secondary consolidation). The paper is not an attack
in the theoretical analysis are seldom if ever satisfied in the field. on theoretical analyses, which are usually essential for
In an interview just weeks before his death in 1963, Terzaghi preliminary design, and always essential for understanding. On
was asked (Engineering News Record, 1963) to explain the the contrary it is an attempt to show the value o f good
unprecedented influence o f his 1925 articles on engineering in observations o f performance and the need to analyse, publish
America. He replied with two reasons: 1) "the papers contained and be guided by case studies that are relevant to all stages of
a wealth o f new and badly needed information concerning the design.
engineering properties o f the subsoil o f our structures" and 2)
"they aroused the unjustified hope that sooner or later all the
problems o f earthwork engineering could be solved by rigorous 2. SECONDARY-TYPE CONSOLIDATION IN THE
methods like those employed in the design o f superstructures". FRASER RIVER DELTA
He cautioned that rigorous solutions are not possible because "if
a body o f soil has an erratic pattern o f stratification its significant The Fraser River encloses a triangle delta between north and
engineering properties can be evaluated only in a very general south branches before emptying into the Pacific Ocean just south
way". He went on to state that “soil mechanics has explained o f the city o f Vancouver, Canada. The cases reported in this
the performance o f soils under load, shear and seepage paper are located within this triangular part o f the delta where
pressures. Therefore, it has become possible to locate in advance the soils consist o f interlayered sands, silts and clays generally up
o f construction, by boring and testing, the potential seats of to 300 m thick .
trouble in the subsoil o f our structures and to cope with them by More than forty years ago when the construction o f major
adequate design". He added that "the instances in which my facilities began to encroach on the soft, compressible alluvial
professional work called for advanced analytical procedures are soils in the delta engineers began to employ the technique o f
very rare". With respect to research carried out to fill gaps in preloading the ground in order to consolidate the subsoil before
our knowledge o f the engineering properties o f soils in general constructing large buildings or other heavy structures. It was
he pointed out that "the value o f these activities could be soon apparent that the complex interlayered subsoils were ideally
considerably increased if the study o f the long-term performance suited to this technique; they allowed rapid lateral drainage o f
of existing structures, such as earth dams, settling foundations, the porewater and sufficient consolidation could usually be
or moving slopes were more often included." In the decades achieved within acceptable time periods ranging from a few
since 1963 an enormous literature on consolidation has been months to a year. This method o f construction is now common
published but there are relatively few documented case histories practice in the delta area.
comparing predictions with performance and o f these most are

73
Fig. 1. Settlement observations in the Fraser River Delta

2.1 Case I: Iona Island. o f secondary-type consolidation (i.e. a straight line on a semi-log
plot o f settlement with time).
An early example o f preloading o f the deltaic soils was begun in In all o f these cases the rate o f secondary-type consolidation
1959 for a sewage treatment plant on Iona Island, in the north settlement depended on the magnitude o f the loads and on the
arm o f the Fraser River. This was an extremely difficult site, magnitude and duration o f previous loads. The term "secondary-
much more difficult than earlier projects, because o f congestion, type consolidation" is used because there are insufficient
a wide variation in loads and a greater sensitivity o f the observations o f pore pressures and settlements at depth to know
structures to settlements. It required complex scheduling o f whether or not some primary consolidation is occurring in deep
preloads and surcharging based on careful interpretation o f soil strata while the surface observations indicate secondary
laboratory tests and earlier experiences. Fortunately the design, consolidation. There is enough variation from site to site to
construction and 27 years o f good settlement observations at the conclude that results are site specific but the nature o f the curves
plant have been described by Ripley (1995). The settlement appears to be sufficiently consistent and long term to aid in
observations on the structures having the maximum and settlement predictions for foundations on the delta.
minimum long term settlement are reproduced in Fig. 1(a)
(Crawford and Morrison, 1996).
3. CASE 4: BAY FARM ISLAND, CALIFORNIA
2.2 Case 2: Lulu Island, Primary Treatment Plant
In his Karl Terzaghi Lecture Duncan (1993) pointed out that
Another waste water treatment plant was built in the early conventional consolidation theory has found wide application in
1970's on a preloaded site on Lulu Island along the south arm o f geotechnical engineering practice and is commonly used by most
the Fraser River. At this Primary Treatment Plant a preload of practicing geotechnical engineers. He explained that
100 kPa was applied in twenty days over an area about 150 conventional theory is likely to continue to provide the most
metres square, causing a settlement o f 0.5 m. At the end o f the effective tool for teaching and learning the fundamental physics
loading period a secondary-type consolidation (0.42 m per log o f consolidation; he examined the reasons why the conventional
cycle o f time) began. After a total o f 400 days the application o f theory has serious limitations in practice and he supported his
a surcharge to 150 kPa caused a further 0.5 m o f primary views with tw o case histories.
consolidation settlement during the following 200 days. When an His first case is the Bay Farm Island landfill project on the
excavation in the fill was made for the plant there was a slight east side o f San Francisco Bay where an area o f 2.6 square km
rebound and during and after construction the secondary-type was loaded hydraulically with sand to a depth o f 2.5 m to 6 m.
consolidation continued at 0.25 m per log cycle for a further 20 Surface settlement plates were monitored at the site from 1967
years as shown in Fig. 1(b). to 1979, and the maximum and minimum observed settlements at
10 locations are shown in Fig. 2.
2.3 Case 3: Lulu Island, Secondary Treatment Plant In 1979, after the fill had been in place for 12 years, the
developer retained two independent consulting firms (both
In 1993 a site adjacent to Case 2, approximately 220 metres having experience at the site), to estimate the maximum amount
square, was preloaded for the construction o f a Secondary o f differential settlement that might occur in a 23 m square
Treatment Plant. A preload o f 225 kPa was applied in a month building supported on shallow foundations. One firm estimated a
causing a settlement o f about 1 m during loading. A further 1 m maximum differential settlement o f 300 mm for such a building,
occurred during the following 50 days and then a secondary-type while the second firm, using the same information, estimated a
consolidation began at 0.75 m/log cycle as shown in Fig. 1(c). maximum differential settlement o f only 30 mm.
There is no doubt that this was a difficult assignment as
2.4 Fraser River Delta Cases Compared illustrated by the 10-fold spread in the estimates o f differential
settlements. As it turned out, in 1992 the maximum differential
Details o f Cases 2 and 3 are given by Crawford and Morrison settlement measured in any o f the buildings was 60 mm. The
(1996) and the results compared with Case 1 on Iona Island and probable reason for the variation between the estimates of
three additional widely separated cases in the delta area. In all minimum and maximum settlements are discussed in detail by
six cases most o f the consolidation settlement occurred during or Duncan (1993). The last half o f the total settlement during the
shortly after the loading period followed by a continuing period 12-year period follows approximately a straight line on a semi-

74
 ' 1-------------

AVERAGE Cf \\ 2 YEARS
LABORATORY TESTS
D IL U V IA L LA Y E R
o« \\ -

r
SET T LEM EN T

\
(7.3 m /cycle)
AIR PO R T
OPEN ING £
U \1
20 ‘
C A LC U L A T E D
J .
SET T LEM EN T
TO TAL
SE T T LE M E N T
(7.9 m/cycle)
VARIATION 1962 - 19M \
TIME - YE A R S
10 ------------- 1____ ...1. . . 1..............
0 20 40 60 eo 100
Fig. 2 Case 4: Settlem ent a t Bay Farm Island VERT IC A L EFFECTIV E ST R E S S - kP«
1 10
TIME - YEARS Fig. 4 Case 6: Settlem ent a t Gloucester
Fig. 3 Case S: Settlement at Kansai Airport

log plot with time, but unlike Cases 1, 2 and 3 the rate of sand layers on pore water drainage (Endo et al., 1991).
settlement is very high, ranging from 1.3 to 1.5 m per log cycle, Consequently, predictions are continually under review based on
and the records are too limited to draw conclusions with respect settlement observations at the site; observations that may be
to primary and secondary consolidation phases. Clearly the rate expected to guide engineers in that region for many decades to
will have to decrease and future observations at the site would come.
be extremely valuable to the understanding o f the actual
consolidation characteristics in this case.
5. CASES OF CLAY DESTRUCTURATION

4. CASE 5: KANSAI INTERNATIONAL AIRPORT, JAPAN 5. 1 Case 6: New Liskeard, Ontario, Canada

The second case discussed by Duncan (1993) is the airport built The planned stage construction o f a highway embankment over
on a man-made island in Osaka Bay, Japan. This project began in varved clay at New Liskeard had to be abandoned because pore
1987 with the dumping o f 33 m o f fill into water 18 m deep over pressures under the embankment did not decrease and shear
a huge area 4.3 km by 1.3 km. The underlying natural soils strengths did not increase after more than two years (Stermac et
consist o f 20 m o f soft, alluvial clay lying over more than 400 m al., 1967). In a discussion o f the paper by Crawford and Eden
of somewhat over-consolidated diluvial clay o f Pleistocene age (1967) it was noted that the imposed total stress increase in the
with layers and lenses o f sand. In 1986, the 50-year settlement consolidating layer was approximately equal to the measured
estimates were 6.5 m in the alluvial clay and 1.5 m in the diluvial preconsolidation stress o f the natural soil. On the basis o f this
clay, for a total o f 8 m. observation and simultaneous work on the Leda clay at Ottawa
Sand drains were used through the 20 m alluvial layer and as it was concluded that "owing to structural breakdown in the
expected consolidation o f this layer occurred very quickly, natural soil, the pore pressures were increasing as quickly as
almost coincidentally with the applied load, but a little less than they could be relieved by drainage"
the predicted value. The diluvial layer, on the other hand,
consolidated much more and at significantly faster rates than 5.2 Case 7: Vasby, Sweden
predicted. At the end o f 1990 the measured settlement o f the
alluvial layer was about 6 m and o f the diluvial layer more than A further dramatic example o f pore pressure generation by
2 m for a total settlement o f about 8 m after 3 years, an amount structural collapse was given by Chang et al. (1973) for a test fill
equal to the 50-year prediction. Consolidation o f the diluvial in Sweden. It was shown that 22 years after loading, the pore
deposit was continuing at a rate o f about one-half metre a year, pressures at the middle o f the consolidating layer were still
so the prediction was re-assessed on the basis o f the actual approximately equal to the total stress increase at that level and
settlement records. it was concluded that "the dissipation o f pore pressures due to
In a recent publication Shiraishi and Suzuki (1995) explained consolidation was compensated by pore pressure regeneration.
that with the re-analysis, based on modified drainage conditions, This portion, called secondary consolidation, has generally been
the settlement o f the Pleistocene layers was calculated to be 4 m considered to begin after the completion o f the primary phase,
at the time o f airport opening in 1994 and a further 1.5 m during and very low excess pore pressures have been thought to exist
the next 50 years. The observed and re-calculated settlements during the secondary phase".
are reproduced in Fig. 3. Since 1989, the consolidation o f the
Pleistocene soils has followed a secondary-type curve although 5.3 Case 8: Gloucester, Ontario, Canada
the re-calculated curve indicates that the rate should soon
decrease. Only future observations can confirm this prediction. Long term settlement observations, beginning in 1955, were
The profession is indebted to Japanese colleagues for made on a building founded on a gravel fill over the sensitive
publishing information on the performance o f the Kansai airport Leda clay near Ottawa, (total applied stress = 35 kPa). Most of
site from the beginning o f the project. This case has illustrated the consolidation o f the clay occurred in a 3 .6 m layer below a
the great difficulty o f predicting consolidation settlements in 2 m thick crust. During the first year the increase in pore
very deep soil strata. Estimates for the upper alluvial soils were pressures limited the increase in effective stress at mid-layer to
quite adequate but for the lower diluvial layers both the the laboratory measured preconsolidation stress, but the vertical
predicted magnitude and rate o f consolidation were much too compression was about 60 mm. After 20 years, there was still
low. In their own view this was caused primarily by uncertainties little change in the effective stress at mid-layer but the
in determining the pre-consolidation stresses from standard compression o f the layer had increased to 250 mm, indicative o f
consolidation tests, and in assessing the influence o f thin diluvial secondary consolidation. Similar observations were made under

75
an earth embankment in the same vicinity at Kars, Ontario (Eden 1) The sampling and testing o f soils at relatively shallow depths
and Poorooshash, 1968). will usually provide reasonable estimates o f the amount o f
From the 20th year until the 33rd year the maximum com­ consolidation settlement under load but the predicted rate o f
pression o f the 3 .6 m consolidating layer increased from 250 mm settlement is often not reliable.
to 280 mm (i.e., from 6.9% to 7.8%), and was then occurring at 2) When the applied stress increase extends to great depths the
a rate o f about 1 mm per year. The discrepancy between the prediction o f consolidation settlements is in jeopardy.
stress-compression measured in situ and from incrementally 3) In addition to applied loads, Leroueil (1996) has identified
loaded laboratory tests is shown in Fig. 4 (Crawford and several other important factors that influence consolidation.
Bozozuk, 1990). For more than 30 years a secondary-type 4) The separation o f consolidation into primary and secondary
consolidation had occurred at an almost constant vertical phases is an arbitrary separation that often tends to confuse
effective stress equal to the measured preconsolidation stress. the consolidation process.
5) Predictability would be greatly enhanced if publications o f
6. DISCUSSION theoretical analyses were accompanied by observations o f
settlement and pore pressure at depth.
In a summary o f the highlights o f the 4th International 6) The search for near perfect, widely applicable analytical solu­
Conference on Soil Mechanics and Foundation Engineering, D.J. tions to consolidation problems has been and will probably
Maclean (1957) expressed disappointment that there was an continue to be in vain because o f sampling difficulties and the
inability to assess the rate o f pore pressure dissipation in situ.
"absence o f any significant advance in obtaining a better
understanding o f secondary consolidation". Thirty-six years later
Duncan (1993) stated that "there is still important progress to be REFERENCES
made to improve our ability to anticipate accurately the
magnitudes and rates o f consolidation settlements". The cases Chang, Y.C.E., B. Broms, R.B. Peck. 1973. Relationship
described above indicate the discrepancies from theory to between the settlement o f soft clays and excess pore pres­
practice that can occur and that probably do occur often. sures due to imposed loads. Proc., 8th ICSMFE. M oscow
It is often useful to separate consolidation into primary and 1:1, 93-96.
secondary consolidation phases but these studies emphasize the Crawford, C.B. and M. Bozozuk. 1990. Thirty years of
value o f the definition o f Terzaghi (1943) that “Every process secondary consolidation in sensitive marine clay. Can. Geot.
involving a decrease o f the water content o f a saturated soil J., 27: 315-319.
without replacement o f the water by air is called a process o f Crawford, C.B. and W.J. Eden. 1967. The performance o f an
consolidation”. It follows that consolidation is a continuous embankment on a deep deposit o f varved clay. Discussion.
process, interrupted temporarily by pore pressures caused by Can. Geot. J., 4 : 63-64.
low permeability o f the soil and in some cases by a collapsing Crawford, C.B. and K.I. Morrison. 1996. Case histories
soil structure. illustrate the importance o f secondary-type consolidation
The literature on soil consolidation is heavily weighted settlements in the Fraser River delta. Can. Geot. J. (in press)
toward laboratory tests and theoretical analyses without Duncan, J.M. 1993. Limitations o f conventional analysis of
corresponding full-scale performance studies. The probable consolidation settlement. ASCE J. o f Geot. Eng., 119: 1333-
reason is that laboratory and desk studies fit nicely into student 1359.
research projects that are valuable training exercises. Case Eden, W.J. and H.B. Poorooshasb. 1968. Settlement
histories, on the other hand, are time consuming and very observations at Kars bridge. Can. Geot. J. 5: 29-45.
expensive if field work is required. This creates a dilemma for Endo, H., K. Oikawa., A. Komatsu and M. Kobayashi. 1991.
research managers but the solution could be in the development Settlement o f diluvial clay layers caused by large scale man-
o f improved liaison between researchers, consulting firms and made island. Proc., Geo-Coast ‘91: 177-182, Yokohama.
government agencies. Practicing engineers always have many Engineering News-Record. 1963. Karl Terzaghi’s last writings
records o f foundation designs and performance in their files but on soils. Nov. 21: 39-40.
seldom have time to prepare them for publication. The Leroueil, S. 1996. Compressibility o f clays: fundamental and
opportunity for graduate students to work with consulting practical aspects. J. Geot. Eng., ASCE 122: 534-543.
engineers on case histories would be excellent experience for Maclean, D.J. 1957. Report on 4th ICSMFE, London. Nature,
them, would help seasoned engineers to improve their expertise, October 26: 839-840.
and would produce valuable publications. Ripley, C.F. 1995. Preloading thick compressible subsoils - a
Leroueil (1996) has brought much o f the consolidation case history. Can. Geot. J., 32: 465-480.
research into focus in a recent paper which explores the Shiraishi, N and S. Suzuki. 1995. Settlement management
influence on compressibility o f natural clays by "strain rate, system for the construction o f Kansai International Airport.
temperature, sampling disturbance, stress path and structuring Compression and Consolidation o f Clayey Soils. Balkema,
phenomena". His conclusions are based on a blending of Rotterdam : 647-651
laboratory and field research that show influence trends o f the Stermac, A.G., K.Y. Lo and A.K. Barsvary. 1967. The perform­
above factors and finally he noted that "using judgment and local ance o f an embankment on a deep deposit o f varved clay.
experience, the engineer will decide how to calculate the Can. Geot. J., 4 : 45-65.
settlements o f the considered structure” Terzaghi, K. 1926. Principles o f Soil Mechanics. Reprint o f eight
articles published Nov.-Dee. 1925 in Engineering News-
Record, Vol. 25. McGraw-Hill, New York.
CONCLUSIONS Terzaghi, K. 1943. Theoretical Soil Mechanics. John Wiley &
Sons Inc., New York, N Y.
These case studies describe some extreme departures from
calculated consolidation settlements and lead to the general
conclusion that the approach to consolidation calculations needs
revision, taking account o f Terzaghi’s original concepts. Some
specific conclusions are:

76