Missouri University of Science and Technology

Scholars' Mine
International Conferences on Recent Advances in 2001 - Fourth International Conference on Recent
Geotechnical Earthquake Engineering and Soil Advances in Geotechnical Earthquake Engineering
Dynamics and Soil Dynamics

Mar 26th - Mar 31st

Performance of Soil Improvement Techniques in

Earthquakes
Elizabeth A. Hausler
University of California, Berkeley, CA

Nicholar Sitar
University of California, Berkeley, CA

Follow this and additional works at: https://scholarsmine.mst.edu/icrageesd

Part of the Geotechnical Engineering Commons

Recommended Citation
Hausler, Elizabeth A. and Sitar, Nicholar, "Performance of Soil Improvement Techniques in Earthquakes" (2001). International
Conferences on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics. 6.
https://scholarsmine.mst.edu/icrageesd/04icrageesd/session10/6

This Article - Conference proceedings is brought to you for free and open access by Scholars' Mine. It has been accepted for inclusion in International
Conferences on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics by an authorized administrator of Scholars' Mine. This
work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder.
For more information, please contact [email protected].
 PERFORMANCE OF SOIL IMPROVEMENT TECHNIQUES IN EARTHQUAKES
Elizabeth A. Hausler Nicholas Sitar
University of California, Berkeley University of California, Berkeley
Berkeley, California-USA-94720 Berkeley, California-USA-94720

ABSTRACT

In the last four decades, there has been a steady trend toward the use of ground improvement as a countermeasure against the hazard of
liquefaction. It is well understood that sites with ground improvement suffer less ground deformation and subsidencethan adjacent, unimproved
areas. However, the lack of quantitative performance data has inhibited the development of empirical relationships between site design
parameters such as remediation zone depth and lateral extent and site performance parameters such as ground and buildmg settlement for a
given level of earthquake shaking.

To date, we have compiled over 90 case histories on the performance of improved sitesfiom 14 earthquakes in Japan, Taiwan, Turkey, and the
United States. The collection of field case histories, as the fist step towards a greater understandmg of the performance of improved soil sites
during earthquakes, are summarized in this paper. The field case histories cover a wide range of improvement methods, fiom conventional
densification methods like sand compaction piles to less common lateral restraint-based methods such as sheet pile walls or deep soil mixing
gnds. The collected data indicate that improved sites generally performed well. About 10 percent of the surveyed sites required sigmficant
post-earthquake remediation, repair or demolition. Unacceptable performance designations resulted most often for excessive ground
deformations in the presence of a severe lateral spreadmg hazard or because of an insufficient remediation zone depth.

INTRODUCTION to be able to predict the ground deformationsfor a given set of site

conditions and earthquake motion. Furthermore, a thoroughreview
Liquefaction-inducedfoundationdisplacement during earthquakes of the field case hstories has brought into light a set of
continues to be a major cause of damage to all types of structures, circumstances in whch ground improvement may not effectively
including buildings, dikes, levees, and. seawalls. Despite eliminate ground deformation, such as the use of conventional
widespread implementation of ground improvement to mitigate denslficationmethods in the presence of a severe lateral spreading
liquefaction-induced ground deformation over the past four hazard, or an inadequate remdation zone depth or lateral extent.
decades, until recently the effectiveness of improvementmethodsto
limit ground strain remained largely unevaluated due to a lack of In h s paper, we present a broad overview of the field case
field performance data under strong shaking (Dobry, 1996). There histories available for improved sites that have been subjected to
is an absence of seismic response observations of actual structures strong ground shakmg. Th~swork is part of a larger research
supported on improved ground. project intended to establish relationshps between remehation
zone geometry and building and ground settlement using empirical
However, there is clear hstorical evidence fiom events as far back data kom the field case histories and a series of dynamic
as the 1964 Niigata earthquake and most recently the devastating geotechcal centrifuge tests that is currently underway.
1995 Hyogoken Nanbu (Kobe), Japan, and 1999 Kocaeli, Turkey,
earthquakes that improved sites suffer less ground deformation and Information has been compiled for over 90 sites from 14 dflerent
subsidence than adjacent, unimproved areas. More importantly, earthquakes in four countries (Table 1). Space limitations for this
these most recent large earthquakes have sigmkantly helped in publication preclude the complete description of all of the field
extending the level of information available relative to the case hstories collected, their sources, and a 1 1 1 acknowledgement
effectiveness of ground improvement wtchell et al., 1998). of the contributors to the dataset. Complete case hstories with
W l e the case histories clearly indicate that ground improvement references and acknowledgments are available on the project
leads to a si@icant reduction, if not elimination, of large ground website, www.ce.berkeley.edu/-hausler/home.html.
displacements during seismic loadmg, the data is not yet sufXcient

Paper No. 10.15 1

Table 1. Field Case Kstories by Event
Earthquake Information
Year Earthquake No. Sites Magtu‘tude 0 Name and date of the earthquake
1999 921 Ji-Ji, Taiwan TBD 7.6 M W 0 Recorded ground motion at or near the site
1999 Kocaeli, Turkey 5 7.4 MW 0 Peak ground acceleration and bracketed duration of the
1997 KagoshimakenHoku, Japan 1 6.3 JMA
1995 Hyogoken Nanbu (Kobe), Japan 49 6.9 MW
earthquake measured at or near the site
1994 Sanriku Haruka Oh, Japan 3 7.5 JMA 0 Distance from site to the zone of energy release
1994 Hokkaido Toho Oki, Japan 4 8.1 JMA 0 Orientation of the site relative to the predominant direction
1994 Northridge, Califomia 5 6.7 M W of motion
1993 HokkaidoNansei Oki, Japan 4 7.8 JMA
1993 Kushiro Oki, Japan 5 7.8 JMA Performance Information
1989 Loma Prieta, California 12 6.9 MW Visual observations throughout and in the vicinity of the
1983 Nihonkai Chubu, Japan 1 7.7 JMA
site, includmg presence or absence of sand boils and
1978 Myagiken Oki, Japan 1 7.4 JMA
1968 Tokachi Oki, Japan 1 6.8 GR evidence of lateral spreading and ground craclung
1964 Niigata, Japan 3 7.3 GR Quantitativeinformation on performance of improved
ground and adjacent unimproved ground, such as amount of
DATA COLLECTION settlement of the ground, and settlement, tilt, and lateral
displacement of the structure
Categoriesof dormation in the database include the following: Details of damage to the superstructure and foundation
Functional state of the facility both immediatelyfollowing
General Site Information the earthquake and long-term
0 Site location and seismic setting Post-earthquake soil conditions, including subsurface
Site topography and proximity to other structures, slopes, or profiles throughout the site with SPT, CPT, or V,
port facilities measurements
Analyhcal studies and laboratory testing of post-earthquake
Soil Conditions soil properties
0 Initial @re-improvement)soil condtions, includmg Determination of acceptable or unacceptable performance,
subsurfaceprofiles throughout the site with SPT, CPT or V, “acceptable” is subdivided into no damage, tolerable
measurements, depth to groundwater, and grain size building settlement, differential settlement, or tilt, and minor
distributionof potentially liquefiable soil structural damage; “unacceptable” is dvided into si&cant
0 Design accelerationand results of liquefaction triggering ground displacement,sigTllficant structural damage,
analyses si&icant repair, remediation, or demolition.
0 Location and extent of nonliquefiable and liquefiable
material for the level of shaking experienced STRUCTURES AND FOUNDATIONS
Estimates of expected settlement and lateral spreading
As shown in Figure 2, the most abundant and complete subset of
Structure Information case lustones in this collectionare those for buildings and tanks on
0 Superstructuredetails, foundation type including pile head shallow foundations. Other collections exist for quay walls and
connection detail for pile foundations and depth of bridges (Dickenson and Yang (1998) and Mtchell -and Cooke
embedment for shallow foundations (1999, respectively).
0 Buildmg plan, column loads and foundation bearing
pressures, and presence of eccentricitiesor unique features
table Performance
Ground Improvement Information
Reason for improvement, such as increase in bearing Bn*, SMm Fdns
capacity or liquefaction mitigation Lak. ROE&, Rtmway~
Retrofit or new construction
Level of improvement required, in terms of density or
required SPT, CPT or V,values
Type of improvement considered and selected
Field tests, analytical studies, and laboratory testing Tanks, Wow Fdm
Depth and lateral extent of improvement, including as-built
drawings
Construction methods, problems and quality control
information I 0 5 10 15 20 25 30 35
Improved soil condtions, includmg subsurfaceprofiles
throughout the site with SPT, CPT or V, measurements Fig. I . Number of Field Case Histones by Faciliv Type

PaperNo. 10.15 2
SOIL TYPE hstories presented here is to restraint the lateral flow of soil.

As illustrated in Fig. 2, the majority of field case hstories involve The average increase in SPT resistance in terms of N1,60is listed in
remediation of reclaimed ground. Table 2 by improvement method. Generally, improvement
methods incorporatingvibration and compaction(sand compaction
piles, stone columns) show a greater increase in SPT N-value than
Coarse dune do those that rely on compaction only (deep dynamic compaction,
stratifid sand Industrial preloading). Without additional consideration of soil type,
sandand 3% waste constructionprocedures, and treatment ratio, &IS statistic must be
clay deposits7 < 1% taken at face value.

Table 2. Improvement Methods Used in Field Case Histories

Performance Average
Method (Acceptable1 Increase in
Unacceptable) N1,60
DensiJicationthrough vibration and compaction
Sand compaction piles 2615 11
Deep dynamic compaction 1510 5
Vibrorodlvibroflotation 11 / 6 . 13
Stone columns 71 1 8
Preloading 510 5
Compaction grouting 111 n/a
Timber displacement piles 1I O da
Dissipation of excesspore water pressure
Gravel drains 510 7
Sand drains 510 9
Fig. 2. Field Case Histoly Breakdown by Soil Type (inPercent).
Wick or paper drains 210 nla
Restraining efect through inclusions
COMMON IMPROVEMENTMETHODS Deep soil mixing 41 1 n/a
Diaphragm walls 01 1 da
The improvementmethods used in the field case historiescollected shflening through chemical or cement addition
fall into four main functional categories (Table 2). With the Jet grouting 510 da
exception of displacement piles, the use of piles as a foundation Chemical grouting 1IO da
alternative without additional modification of the soil properties
was not considered an improvement method. The improvement LEVEL OF SHAKING
methods listed here do not cover the entire spectrumofremediation
technologies available to limit ground deformation during The mformation in the database can be broken down by peak
earthquakes and they are heavily dnven by the Japanese practice ground acceleration measured at or near the site (Figure 3). The
during the 1970’s and 1980’s. Several excellent publications field case hstories are most plentiful for levels of shakmg below
describe techcal specrfications, h t a t i o n s , applicability and 0.5g. Interestingly,unacceptableperformance was documented at
design procedures of remdation technologies, such as Mtchell sites with peak ground acceleration as low as 0.13g.
(1981), Schaeffer (1997), PHRI (1997), and JGS (1 998).
GnatuthwO.8 W Acceptable Performance
The improvementmethods included in the field case h r y dataset
07-04 nce
are categorized according to their primary means of limiting
ground deformation, even if the effects of earthquake strong 0 6 07g
shaking may be mitigated in more than one way. For instance, the
primary intent of sand or wick drains is to accelerate the os-06g

consolidation of soft clay through the dissipation of excess pore

04-0%
water pressure. The same function for the saturated, loose
cohesionlessmaterial through whch the drains may pass may be 0 3 0%
served during and immediately following an earthquake,although
02-03g
the effectivenessof sand or wick drains for this purpose has not
been irrefutably proved, and the act of drainageitselfmay in effect 01-02g
increase the amount of settlement of the improved ground. The
true reduction in ground deformation provided by h s type of LfSl

improvement method may result fiom localized densifcation due

to vibration during installation. Similarly,while deep soil mixing 0 5 10 15 20 25
results in local stiffening of soil, its main function in the case Fig. 3. Field Case Histories by Peak Ground Acceleration.

Paper No. 10.15 3

REMEDIATIONZONE GEOMETRY improvement depth equals 1, especiallyat levels of shakinghigher
than 0.3 g. Second, there is a dearth of field case h s t o q
The required depth of treatment is typically governed by the extent dormation for sites with a normalized improvement depth of less
of material that is potentially liquefiable and the degree of than a half It is not common to remediate such a shallow zone.
settlement that is likely. In U.S. practice, remdation zone depth is W d , the data from the Liu and Dobry (1997) centrifuge tests
typically determined using SPT, CPT, or V, measurements and appears to follow the same trend, but with larger normalized
simplified liquefaction triggeringprocedures developedinitially by building settlement. Finally, it should be noted that the
Seed and Idriss (197 l), updated by Youd and Idtlss (1997) and exceptionally large normalized settlements for all but one of the
revised most recently by Cetin (2000). These common liquefaction outlying data points can be explained by the presence of a
triggering analyses produce a hckness of potentially liquefiable sigmficant lateral spreadmghazard, as discussed more filly in the
material based on a deterministicearthquake hazard assumption, following section.
and from thls the depth and degree of improvement required can
found. It is common practice to remdate throughout the full
liquefiable thickness. However, of the 53 sites with s a c i e n t
information to evaluate the ratio of the hckness of the improved
layer to the thickness of the liquefiablelayer, only 24, or 45%, were D O PGA less than 0.39
improved throughout the full liquefiablehckness.
g 0.04
A Liu and Dobry (1997), 0.29
"
c
I-

The required lateral &stanceor width of soil improvementoutside

the perimeter of the structure is limited to the area that controlsthe
stability of the structure, even if liquefaction occurs over a wide
area (PHRI, 1997). However, the lateral distance that is
necessary to protect the treated zone beneath the structure fiom
si@icant post-earthquakestrength loss and settlement is a source
of great uncertainly Wtchell et al., 1995). Mitchell reports that
commonpractice involves extending the treatment a distance equal
to the depth of the layer being denslfied, a recommendationbased
on extensivefield experience and centrhge model tests performed 0 0.2 0.4 0.6 0.8 1
by Iai and others (1988). Improved ThlcknessI Llqueflable Thlckness

Only 5 of the 25 case histories with data sufticient to evaluate the Fig. 4. Normalized Improvement Depth vs. Normalized Building
ratio between the lateral extent of improvement and the depth of Settlement.
improvement actually extended the improvement laterally equalto
the improved depth. It is ofien not possible to extyd improvement SITES WITH UNACCEPTABLEPERFORMANCE
$

to a distance equal to the depth because of the presence of other

structures and efforts to limit costs. Of the 20 field case histories An unacceptableperformance designation results from sigruficant
with lateral improvementratios less than one, 18experienced some ground deformation, sigmficant structural damage, or necessary
degree of settlement or tilt of the structure. However, at all 18 extensive repair, remdation, or demolition. In the majority of
sites, more severe evidence of the consequences of liquefaction cases, an unacceptable performance determination can be
was documented in adjacent, unimproved areas. explained by excessive ground displacements due to laterally
spreading soils or inadequate remdation zone depth or lateral
NORMALIZED 'BUILDING SETTLEMENT VERSUS extent. Some examples from the 1995 Hyogoken Nanbu (Kobe)
NORMALIZED REMEDIATIONZONE GEOMETRY earthquake follow.

One of the most potentially usell evaluations of the dataset Mikage Hama LPG Storage Tank Site
involves a comparison between building settlement normalizedby
potentially liquefiable thickness and improvement depth In the case of the Mikage Hama LPG storage tank yard locatedjust
normalized by liquefiable hckness, as shown in Figure 4. The north of Rokko Island in Kobe (Ishhara, 1997), the use of
data presented in the figure includesbuildings andtankson shallow vibroflotation to a depth of about half of the liquefiable hckness
foundations supportedon unimproved ground or ground q r o v e d was not able to prevent si@icant settlement and tilt of several
using densification or compaction methods. For comparison, the tanks located near a quay wall. The subsurfaceat the site consisted
results of a centnfuge test series performed by Liu and Dobry of about 15 m of loose, reclaimed decomposed graniteunderlain by
(1997) are also plotted. The authors varied the depth of soft silty clay. The groundwater table was wittun 3 m ofthe ground
compaction below a circular footing while keeping the lateral surface at the time of dnlling.
extent of the improvementrelative to the footing diameterconstant
During the Kobe earthquake, the peak ground acceleration
Several observations can be made from th~splot. First, some recorded about 1 lan from the site was 0.33 g. The quay wall
deformation should be expected even when the normalized underwent about 2 m of horizontal movement. The 30 m dameter

Paper No. 10.15 4

tank located nearest the wall was founded on piles in unimproved ACKNOWLEDGMENTS
soil. The piles extended about 12 m below the liquefiable soil.
Although the tank itself did not settle and the tilt was minor, the ’ h swork was supported primarily by the EarthquakeEngmeering
surrounding ground subsided between 35 and 60 cm. hsociated Research Centers Program of the National Science Foundation
piping on shallow foundationsbecame dislodgedfi-omthe tankdue under Award Number EEC-9701568.
to ground subsidence and major leakage occurred. Of the two
large tanks supportedon shallow foundationson improved ground REFERENCES
the tank nearest the wall settled 62 cm and tilted 1/80 toward the
wall. The tank furthest fi-om the wall settled 44 cm with minimal Adalier, K., Elgamal, A-W., and G.R. Martin. [1998].
tilt. “Foundation Liquefaction Countermeasures for Earth
Embankments”,J. Geotech and Geoenv. Eng., 124(6),pp. 500 -
Hamakoshen Apartment Buildings 517.

The Hamakoshen apartmentbuildmgs are located in Nishinomiya Cetin, K.O. [2000]. “Reliability-Based Assessment of Seismic
City in Hyogo Prefecture, Japan (Tanaka et al., 1996). The circa- Soil Liquefaction Initiation Hazard, Dissertation Submitted in
1962 5-StOIy buildmgs are supported on shallow foundations on Partial Fulfillment of the Requirementsfor the Degree ofDoctorof
ground improved with vibrocompozer (sand compaction) piles. Phlosophy, University of Califomia at Berkeley.
Peak ground accelerationsbetween 0.27g and 0.47gwere recorded
w i b 3 km ofthe site. Of the 150 buildings at the site, five had Dickenson, S.E. and D-S. Yang. [19981. “Seismically-Induced
u n a d t a b l e performance in terms of tilt greater than 1/100. The Deformations of Caisson Retaining Walls in Improved Soils”,
maximum d&erential settlement measured was 20 cm. Mer the Geotechcal Earthquake Engineering and Soil Dynamics III,
earthquake, the soil under those five buildings was “ h a t e d Geotechnical Special Publication 75, ASCE, August, pp. 1071 -
using compaction grouting. 1082.

Unllke the majority of sites damaged during the Kobe earthquake, Dobry, R. [19961. “Geotechnical Engineering Performance
the Hamakoshien apartments were not subjected to a si&icant QuantScation of Foundations”, Issue Paper No. 5, Performance
lateral spreadinghazard. The subsurfaceconsists of about 10m of Based Seismic Design of Buildmgs: An Action Plan for Future
well-graded river and landfilled sand with SPT Nm values between Studies, Prepared for Federal Emergency Management Agency,
6 and 12 blows per foot in the unimproved zones. SPT Nso values March, pp. 54-62.
in an improved area under a building range between 12 and 24
blows per foot. Improvementwas extended to a depth of 4 to 7 m, Iai.S., Koizumi, K., Noda, S., and Tsuchda, H. 119881. “Large
and laterally to a distance equal to one-quarter to one-half of the Scale Model Tests and Analysis of Gravel Drains,” Proc., gfh
improved depth. Thus the remediation zone was quite limited. World Conference on Earthquake Engineering, Japan Assn. for
Earthquake Disaster Prevention, Tokyo, Vol. 111, pp. 26 1 - 266.
Evidence of liquefaction was found along roads, in unimproved
open areas between buildmgs, and in particular, adjacent to all but I~hlhara,K. [19971. “Geotechnical Aspects of Ground Damage
one of the buildings with unacceptable degrees of tilt. Dunng the Kobe-Awaji Earthquake,” Proc., Is‘ Int’l. ConJ on
Earthquake Geotechnical Engineering, Ishhara (ed.), Balkema,
SUMMARY pp. 1327 - 1331.

Field case histories have been compiled for over 90 sites where Japanese Geotechnical Society (eds.). [1998]. Remedial
countermeasures against liquefaction have been employed and Measures Against Soil Liquefaction: From investigation and
tested with strong ground shaking. While there is a growing design to implementation. A.A. Balkema.
database of observed field performance, it is still limited by the fact
that large earthquakes are infrequent and, unfortunately,the quality Liu, L. and R. Dobry. [19971. “Seismic Response of Shallow
of the data from many of the sites is marginal. In many cases, the Foundation on Liquefylng Sand.” J. Geot. and Geoenv. Eng.,
performance assessment is based on statements such as “no 123(6), pp. 557 - 567.
damage was observed“ or “evidence of liquefaction was not
found. Only in a minority of case hstories are there Mitchell, J.K. [19811. “Soil Improvement - State-of-the-Art
measurements of pre- and post-improvement soil properties, post- Report”, Proc., IdhICWFE, Stockholm, Vol. 4, pp. 509 - 565.
earthquake settlement, and budding tilt.
Mitchell, J.K., Baxter, C.D.P., and T.C. Munson. [1995].
The research described herein is intended as a building block for “Performance of Improved Ground During Earthquakes”, Proc.,
sustained future efforts on the evaluation of the performance of Soil Improvements for Earthquake Hazard Mitigation,
improved ground techniques. More detailed studles of the field Geotechcal Special Publication No. 49, ASCE, pp. 1-36.
case histories coupled with carefully designed experimentaltesting
are needed and currently underway. Mtchell, J.K., Cooke, H.G., and J.A. Schaeffer. [1998]. “Design
Considerations in Ground Improvement for Seismic Risk

PaperNo. 10.15 5
Mtigation”, Geotechcal Earthquake Enweering and Soil
Dynamics HI, Geotechcal Special Publication No. 75, ASCE,
August, pp. 580-613.

Mitchell, .J.K. and H.G. Cooke. [1995]. “Assessment and

Mitigation of Liquefaction a s k for Existing Highway Bridge
Foundations”,Proc., National Seismic Conference on Bridges and
Highways, San Diego, CA, December, 20 pp.

Port and Harbour Research Institute. [1997]. Handbook on

Liquefaction Remediation of Reclaimed Land. A. A. Bakema.

Schaeffer, V.R. (ed). [1997]. Ground Improvement, Ground

Remforcement, Ground Treatment -Developments 1987 - 1997,
Geotechcal Special Publication No. 69, ASCE, 6 18 pp.

Seed, H.B. and I.M. Idriss. [1971]. “Simpllfied Procedure for

Evaluating Soil Liquefaction Potential,” J. Soil Mech. and Fdn.
Engrg. Divn., ASCE, GT7 (1 04),July.

Tanka, M., Akiyama, T., and N. Murata. [1998]. “Restoration

and Redorcement of Apartment Housing Foundation,” The
Foundation Engineering and Equipment, Special Issue on
Lessons Leamedfi-om the Hanshin Awaji Earthquake, pp. 108-
113 (in Japanese).

Youd, T.L. and I.M. Idriss. [19971. Proceeding of the NCEER

Workshop on Evaluation of Liquefaction Resistance of Soils,
Techcal Report NCEER-97-0022.

PaperNo. 10.15 6