Grouting 2017 GSP 289 279

Design and Construction of Column Supported Embankment Using Deep Soil

Mixing Columns and Reinforced Concrete Load Transfer Slab

King H. Chin, P.E., M.ASCE1; Heidi P. Cashman, P.E., M.ASCE2; and

Chris Barnes3
Downloaded from ascelibrary.org by Tufts University on 03/17/18. Copyright ASCE. For personal use only; all rights reserved.

1
GeoEngineers, Inc., 8410 154th Ave. NE, Redmond, WA 98052. E-mail:
[email protected]
2
GeoEngineers, Inc., 8410 154th Ave. NE, Redmond, WA 98052. E-mail:
[email protected]
3
City of Seattle Department of Transportation, P.O. Box 34996, Seattle, WA 98124. E-mail:
[email protected]

Abstract

This paper presents the design and construction of an embankment supported on a reinforced
concrete load transfer slab founded on deep-soil-mixing columns completed as part of the
seismic upgrade of the Airport Way South Viaduct project in Seattle, Washington. The project
included replacement of the timber trestle approach structures constructed in 1928 with
mechanically stabilized earth (MSE) approach embankments. Ground improvement was
incorporated in the design to mitigate compressible/liquefiable soil conditions within the new
MSE approaches. Both the limit equilibrium methods and 3D numerical modeling were used to
evaluate the performance of the new embankments and to ensure strain compatibility between
the soils and the DSM columns. By including the reinforced concrete load transfer slab in the
analysis, the amount of DSM columns and ground improvement in the original design were
significantly reduced. Settlement monitoring was completed throughout the construction and
confirmed that the specified performance criterion was met.

INTRODUCTION

The Airport Way South Viaduct was originally built in 1928 and consisted of a 517-m
long elevated structure (a 125-m North approach, a 251-m Main Span, and a 140-m South
approach) to provide grade separation between Airport Way South and the Union Pacific,
Northern Pacific and Seattle to Tacoma Interurban railroad tracks. With over 13,000 vehicles
traveling on Airport Way South daily, the structure experienced increased deterioration over
time. The project consists of replacing the north and south timber trestle approach structures with
additional bridge spans and mechanically stabilized earth (MSE) fill approaches. The project site
is shown in Figure 1.

© ASCE

Grouting 2017
 Grouting 2017 GSP 289 280
Downloaded from ascelibrary.org by Tufts University on 03/17/18. Copyright ASCE. For personal use only; all rights reserved.

Figure 1. Site Map aand Location

n.
The
T MSE fill approachees (North an nd South Appproach) weere originallyy designed to be
supported d on improvved ground with
w compacction groutinng techniquees (Shannon & Wilson, 22010)
with a taargeted 0.6-mm-diameter grout colum mns and an aarea replacem ment ratio oof 10 percennt. As
constructtion of the compaction grouting
g beggan the resultt appeared ppositive, but in the proceess of
adding adjacent
a grrout column ns, more ground
g movvement wass observed. This beecame
increasin
ngly concern ning as the coolumns weree moving clooser to an exxisting sensittive gas mainn that
was with hin 1.4 m at a its closestt point and paralleled the North A Approach onn one side; with
tolerancee of 25 mm maximum that t was quuickly exceedded. The pproduction w was halted inn this
location and the conttractor was directed
d to move
m to a neew area wheere experimeentation of sllower
injection rates was performed. Experimenta
E tion with low
wer injectionn rates yieldded questions that
the targeted 0.6 m diiameter colu umn was nott being achieeved in the ttop 1.5 to 4.6 m due to early
return off the compacction grout. The contraactor attemptted an injecttion rate of 0.03 cubic m meter
per minuute or less ad djacent to th
he gas main,, again, the movement nnearly doublled the allow wable
maximum m. Based on the con nstruction observations
o of the commpaction grrouting, thee city
concludeed that compaction groutting was not constructiblle due to thee presence off the sensitivve gas
main at the
t north ap pproach and was of som me concern ffor the southh approach. Due to thee cost
increase associated with delayss and impaccts to the c ompletion ddate of the project, thee city
decided to redesign n the groun nd improvem ment prograam using ddeep soil m mixing technnique,
presentedd below.

SUBSUR
RFACE SOIIL CONDIT
TIONS

Subsurface so oils encounttered at the site generallly consistedd of fill, alluuvium, estuaarine,
beach annd colluvium
m deposits ovverlying the Blakely
B form
mation (bedrrock). The suubsurface sooils at
the Northh Approach generally
g co
onsist of 6 to 8 meters off loose to meedium dense sand overlyying 2
to 6 meteers of mediu
um stiff to sttiff clayey siilt. Bedrock was encounntered at deppths ranging from
6 to 13 meters.
m The bedrock
b wass found to beb dipping frrom east to w west and froom north to south
within th
he North Ap pproach emb bankment foo otprint. The subsurface soils at the South Apprroach
generally
y consist of 7 to 12 metters of loosee to medium m dense sandd overlying 6 to 14 meteers of

© ASCE

Grouting 2017
 Grouting 2017 GSP 289 281

medium stiff to stiff clayey silt. Bedrock was encountered at depths ranging from 16 to 24
meters. The bedrock was found to be dipping from north to south within the South Approach
embankment footprint. Groundwater was estimated at depths ranging from 4 to 9 feet below the
ground surface.
Based on the laboratory testing and in-situ test results, the clayey silt is determined to be
highly susceptible to consolidation settlement. In addition, the loose to medium sand below the
groundwater table is prone to soil liquefaction during the design earthquake event. Based on our
settlement analysis, we estimated that the total static settlement under the proposed embankment
Downloaded from ascelibrary.org by Tufts University on 03/17/18. Copyright ASCE. For personal use only; all rights reserved.

loads range from 50 to 100 mm. Our liquefaction analyses indicated that liquefaction induced
settlement of 100 to 300 mm can be expected after a design earthquake event, which is defined
as a M6.8 earthquake with peak ground acceleration of 0.47g, which is associated with a 7
percent probability of exceedance in 75 years.

DEEP SOIL MIXING DESIGN

The DSM ground improvement system utilizes a reinforced concrete slab as the load
transfer platform instead of a gravel layer to distribute the load from the MSE walls to the DSM
columns in a uniform manner and to more effectively engage the load resisting capacity of the
DSM columns. This results in a more optimized design that requires less DSM columns and
reduces the impact to the construction schedule. Another important feature of the recommended
DSM ground improvement system is the use of a load transfer gravel layer between the
reinforced concrete slab and the DSM columns. One of the key functions of this load transfer
gravel layer is to isolate the DSM columns from the reinforced concrete slab to minimize the
transfer of shear forces from the MSE wall to the DSM columns, especially under seismic
conditions. The high friction resistance of the gravel layer will also prevent the MSE wall from
sliding under seismic conditions. The DSM design process is an iteration process between
simplified and semi-empirical analysis and numerical techniques. Preliminary DSM columns
were design on the basis of simplified and semi-empirical analysis including consolidation
analysis, liquefaction analysis and pile capacity analysis. Three dimensional finite difference
analysis (FLAC 3D) was used to evaluate the DSM performance of designed ground
improvement. Figure 2 shows the DSM layout developed for the North and South Approach
Embankments.
Soil Parameters. Soil parameters were developed using the published correlations with blow
counts for sand and correlation between CPT test data and engineering properties developed by
National Cooperative Highway Research Program (NCHRP) Synthesis 368 (2007). Table 1
below presents the soil parameters used for design of the DSM system. The compaction grout
columns were not explicitly modeled in our DSM design analysis but the effects of compaction
grouting was considered to develop the improved soil parameters by only considering the
unconfined compressive strength of the grout column based on the as-built column diameter.
The densification effects of the compaction grouting were ignored because of the inability to

© ASCE

Grouting 2017
 Grouting 2017 GSP 289 282

inject th
he design grrout volumee due to ex
xcessive groound movem
ment and/orr early returrn of
compactiion grout to the ground surface.
s
Downloaded from ascelibrary.org by Tufts University on 03/17/18. Copyright ASCE. For personal use only; all rights reserved.

Figure 2. North (Top) and South (Bottom)

( Ap
pproach DS
SM Column
n Layout.
Table 1. Soil Param
meters
Unit Weig
ght Moodulus of
So
oil Unit 3 Shear Strrength
(kN/m ) Elastticity (MPa)
Sand (aabove GWT) 19.6 φ = 38o, c = 0 kPa 38.3
28.7 (UUnimproved))
φ = 36o, c = 0 kkPa (Static)
Sand (b
below GWT) 18.8 47.9 (Improved)d
φ = 0o, c = 23.9 kP
Pa (Liquefiedd)a
5.7 ((Liquefied)e
φ = 0o, c = 35.9 kPaa (Unimproveed)b 57.5 (U
Unimproved))
Claayey Silt 15.7 o
φ = 0 , c = 67.0 kPPa (Improvedd) c
71.8 (Improved)d
o
Rock
R 21.2 φ = 0 , c = 4478.8 kPa −
Notes:
a. The residuaal shear strength of the liquefiable laayer is determined using the correlattion between equivvalent clean sand S SPT blow counts ((N1)60-
CS and residu ual undrained sheaar strength (Seed and
a Harder, 1990).
b. The undraiined shear strengthh of the clay is deteermined using the correlation: (qt – σ vo) / Nkt, where qt = cone tip penetrration resistance, σvo =
vertical stresss and Nkt = 15.
c. The improv ved (compaction grouting)
g shear streength was calculatted based on undraained shear strengtth of the compactioon grout = 345 kPa.
d. The improv ved (compaction grouting)
g modulus of elasticity was calculated
c weighted average vvalue of the modullus of elasticity of the
as the w
compaction grout
g of 300 x qu = 207 MPa, and thee modulus of elastiicity of the soil.
e. The liquefiied Modulus of Elaasticity was back calculated
c using th
he liquefaction indu
duced settlement siimplified method aanalysis results.

© ASCE

Grouting 2017
 Grouting 2017 GSP 289 283

DSM and Structural Slab Parameters. The DSM column properties were determined using the
results of the unconfined compressive strength laboratory tests and the plate load test. Filz et al (2005)
presents the relationship between Young’s Modulus, E, and the unconfined compression strength, qu, to
range between 75 and 1,000. Based on the results of the unconfined compressive strength tests performed
on the DSM column core samples, we estimated the DSM column stiffness to be about 190 qu. Two plate
load tests were also completed, one at each approach, and the DSM column stiffness was back-calculated
to be range from 500 to 1,000 qu. Based on the results of the laboratory tests and plate load tests, we
completed our numerical modeling assuming a DSM column stiffness of 150 qu and 500 qu to account for
the uncertainties associated with the DSM column stiffness. The structural slab properties were
Downloaded from ascelibrary.org by Tufts University on 03/17/18. Copyright ASCE. For personal use only; all rights reserved.

determined using the ACI semi-empirical equations for static conditions, this assumes that the slab
remains elastic (uncracked conditions). For seismic conditions, we used 50 percent of the static modulus
(cracked conditions) based on discussions with the project team. Table 2 below presents the DSM
column and structural slab properties used in our analyses.

Table 2. DSM and Structural Slab Parameters

Unit Weight Unconfined Compressive Strength - Modulus of Elasticity
Element
(kN/m3) qu (MPa) (MPa)
DSM Columns 19.6 2.1 310 – 1,034
Structural Slab 23.6 27.6 24,897

Global Stability. Global stability analyses were completed using the computer program
SLOPE/W (GEO-SLOPE International, Ltd., 2005). SLOPE/W evaluates the stability of the
critical failure surfaces identified using vertical slice limit-equilibrium methods. This method
compares the ratio of forces driving slope movement with forces resisting slope movement for
each trial failure surface, and presents the result as the FS. Based on our analyses, we conclude
that the factor of safety of global stability is significantly higher than 1.5 and 1.1 for the static
and seismic conditions, respectively.

Performance and Internal Stability Evaluations. Numerical analysis was completed to

evaluate the performance of the ground improvement system under the static conditions using
the computer PLAXIS 3D, as shown in Figure 3. FLAC 3D models were developed to evaluate
the performance of the ground improvement system under the seismic conditions, as shown in
Figure 4. Three dimensional (3D) models were used to assess the variable subsurface soil
conditions and the non-uniform ground improvement configurations. Prior to completing the
seismic analyses using FLAC 3D, the static FLAC 3D results were compared to PLAXIS 3D
results, to make sure that results from both models compare reasonably well.
Based on the results of our numerical analyses, we concluded that the ground
improvement system will reduce the foundation soil settlement to meet the specified settlement
criteria for this project. Table 3 below summarizes the settlement computed based on the

© ASCE

Grouting 2017
 Grouting 2017 GSP 289 284

numericaal modeling results. Alsso presented d in Table 3 are the rangge of settlem
ment measurred at
the end of
o the emban nkment consttruction.
The
T internal stability an nalyses wass completedd to evaluatte the stressses of the DSM
columns (i.e. axial force,
f shear and momen nt) using thee results of PLAXIS 3D D and FLAC C 3D
results. Lateral
L pile analyses
a werre also comppleted using the computeer program L LPile. The reesults
of the internal stab bility analysses confirm that minim mum requirred unconfinned compreessive
strength of the DSM M columns of o 2.1 MPa will providee a factor of safety highher than 2 uunder
nditions and higher than 1.1 for the seismic
static con s and ppost-earthquuake conditioons.
Downloaded from ascelibrary.org by Tufts University on 03/17/18. Copyright ASCE. For personal use only; all rights reserved.

(102 to 192 kPaa)

(102 to 192 kPa)

(300 m
mm)
(150 mm)

(1500 mm)
(30
00 mm) (3
( to 12 m)

(15.2 to 22.9 m)

(1500 mm)
(150 mm
m)
(2.4 to 6.1 m) (2.4 to 12 m
m)

(6.11 to 12.2 m)
(2.4 to 6.1 m)

North Approach
A (1.2 to 1.5m)

S
South Approach
h

Figure 3. PL
LAXIS 3D M
Models.

Norrth Approach South Ap

pproach

Figure 4. FLAC
F 3D M
Models.

Table 3. Su
ummary of Numerical
N Modeling
M R
Results and S
Settlement S
Surveys
Post-improvem
ment Static Post-improvem
P ment Seismic Measured E
Embankment
Settlementt (mm) Settlemennt (mm) Settlement (mm)

North Approach
A 2.5 to 7.6 2.5 to 227.9 2.5 too 17.8
South Approach
A 2.5 to 17.8
1 2.5 to 35.6 5.0 too 22.9

© ASCE

Grouting 2017
 Grouting 2017 GSP 289 285

DSM GR
ROUND IM
MPROVEME
ENT SYSTE
EM CONST
TRUCTION
N

DSM Installation. DSM D column ns was constructed by innjecting cem mentitious grrout into thee soils
as the drrill rig advaances drill stteel with a radial mixinng paddle (ssingle axis) located neaar the
bottom of
o the drill string
s to the design deptth. The radi al mixing ppaddle mechanically cuts and
blends th
he soil with h the injected grout, wh hich creates in-situ soil--cement inddividual coluumns.
Figure 5 shows the seelected consstruction pho otos of DSMM installationn at the project site.
Downloaded from ascelibrary.org by Tufts University on 03/17/18. Copyright ASCE. For personal use only; all rights reserved.

Fig
gure 5. DSM
M Installatioon Photos.
Grading g and Load Transfer Platform
P Coonstruction. Upon comppletion of thhe DSM coluumns,
w graded such that the top of the DSM colum
the site was mns are at thhe design ellevations priior to
placing the
t gravel laayer and con nstruction off the load traansfer reinfoorced concrette slab. Figgure 6
shows th he constructiion photos ofo the gradiing and connstruction off the load trransfer slab. As
shown in n Figure 6, many
m DSM exposed
e durring the gradding processs was damagged where reepairs
were mad de prior to placing
p the gravel
g layer. In additionn, a layer of geotextile w
was placed oon top
of subgraade soils wiith the DSM M columns prior
p to placcing the gravvel layer. TThe intent oof the
geotextile is to creaate a separaation betweeen gravel annd the nativve soil and also reduce the
potential shear force transfer from
m the reinforrced concrette slab to thee DSM colum mns.

DSM Construction n QA/QC. A Quality Control proogram was implementeed to verifyy the
installed DSM colum mns meet: 1) the specified DSM uncoonfined com mpressive streength of 2.1 MPa
at 28-dayy; and 2) Thee minimum percent
p reco
overy from thhe core sampples of 85%.. Coring of D DSM
column wasw completed using a Soilmec SM M-14 drill w
with a triplee-barrel wireeline rock ccoring
system outfitted
o withh a 82.6-mm meter face ddischarge suurface set coore bit. Figure 7
m inside diam
shows thhe core recovvered for a DSM
D columnn SG-31 locaated in the South Approaach, which sshows
highly vaariable recovvery for the DSM colum mn constructeed using thee same installlation proceedure.
The high hly variable core recoveered could be attributed to the variaable unconfiined compreessive
strength of the DSM M column and a the coriing proceduure used. Inn order to confirm thaat the
installatio
on procedurre can proviide a contin nuous DSMM column, p--wave veloccity of the D DSM
column was
w measureed using the suspension logging geoophysical tecchnique. Figure 8 show ws the
p-wave velocity
v profile measureed for the DSM
D columnns with the variable recovery show wn in
Figure 7.7 Using th he results of
o the p-waave measureements, we were able to correlatee the

© ASCE

Grouting 2017
 Grouting 2017 GSP 289 286

unconfin ned compresssive strength h of the DSMM column aand verify thhat the minimum unconnfined
compresssive strength h was met th hroughout thhe entire lenggth of DSMM. DSM saamples, incluuding
both the wet in-situ DSM samplles (collected from the ddrilled holess) and core ssamples obtained
were testted to determmine the 3, 7,
7 21 and 28--day unconffined compreessive strenggth in accorddance
with AST TM D1633. The results of the uncon nfined comppression testss are presentted in Figuree 9.
For this projeect, a static load test waas also comppleted on twwo selected D DSM colum mns to
verify thee design parrameters useed in the num merical moddeling. Figuure 10 showss the photo oof the
static loaad test set up
u and the load test results for DSM M Column DG-16 locaated in the N North
Downloaded from ascelibrary.org by Tufts University on 03/17/18. Copyright ASCE. For personal use only; all rights reserved.

Approach h. The DSM M test colummns were loaaded in 5 perrcent incremments up to thhe design looad of
689 kN. Each load increment
i was
w held for 4 minutes w with the creepp displacemment recordedd at 1
minute in ntervals. Affter the test columns
c weere tested at the design lload, the DS
SM columns were
then unlo oaded in 20 percent load d increments down to thhe alignmennt load. Thee load test reesults
show thaat the DSM parameters
p used
u in our an
nalyses are aappropriate.

Figuree 6. Grading
g and Load Transfer S
Slab Constru
uction Photos.

Figure 7. DSM Corees with High

hly Variablee Recovery ((Column SG
G-31).

© ASCE

Grouting 2017
 Grouting 2017 GSP 289 287

Depth (m)
0.0 5.0 10.0 15.0 220.0
1000

p-wave velocity (m/s)

1400

1800

2200

2600

3000
Downloaded from ascelibrary.org by Tufts University on 03/17/18. Copyright ASCE. For personal use only; all rights reserved.

Fig
gure 8. P-w
wave velocity
y measurem
ment (Colum
mn SG-31).

North Approach – Compressive

C Stren
ngth Results South Approach - Compressive Strength Results
0 0

[6.1]
20
Depth (Feet) [m]

Depth (Feet) [m]

[12.2]
20 40
[6.1]

[18.3]
60
3-day strength 3-day strength
7-day strength 7-day strength
21-DAY STRENGTH 21-day strength
28-day strength [24.4] 28-day strength
40 80
[12.2] 0 200 [1.38] 400 [2.76] 600 [4.14] [5.52] 80
00 0 1000 [6.89] [13.79] 2000
Compre essive Strength (psi) [MPa] Comprressive Strength (psi) [M
MPa]

Figure 9. Summa
ary of DSM Unconfined
d Compresssion Strengtth.

Figure 10.
1 DSM Sta
atic Load Test and Tesst Results (C
Column NG
G-16).

CONCL
LUSIONS

Based
B on the results of our analyses, the DSM ccolumns withh load transffer structurall slab
system can significanntly lower thhe static setttlement and mitigate liqquefaction innduced settleement
under thee design earrthquake eveent. Using the load reiinforced conncrete load ttransfer slabb will
optimize the design n of the DS SM columnss. Combiniing with sim mplified andd semi-emppirical
engineeriing analysis, three dimensional simu ulation technniques can pprovide realiistic perform
mance
of the grround impro ovement sysstem. As presented
p in Table 3, settlement m monitoring dduring

© ASCE

Grouting 2017
 Grouting 2017 GSP 289 288

construction showed the static settlement ranged from 5.0 to 22.9 mm within the South Approach
Embankment and ranged from 2.5 to 17.8 mm within the North Approach Embankment, with the
highest reading from the area with highest embankment. The survey results were generally
consistent with the settlement estimated by the numerical modeling for static conditions.
The Airport Way viaduct was re-opened to traffic in December 2012, one month earlier
than the 14-month maximum closure requirement. The project also recovered over two months
of construction delays that occurred during the original compaction grouting design completed
by other consultant, and saved over $600,000 compared to the original design.
Downloaded from ascelibrary.org by Tufts University on 03/17/18. Copyright ASCE. For personal use only; all rights reserved.

REFERENCES

Filz, Dr. G.M., Duncan, Dr. J.M., Gutierrez, Dr. M.S., Kriz, Dr. R.D., Singh, Dr. M.P. 2005.
“Stability of Embankments Founded on Soft Soil Improved with Deep-Mixing-Method
Columns.” Navin, M.P. Dissertation submitted to the faculty of the Virginia Polytechnic
and State University in partial fulfillment of the requirements for the degree of Doctor of
Philosophy in Civil and Environmental Engineering.
GeoEngineers, Inc (2012). “Supplemental Geotechnical Explorations and Geotechnical Report,
Airport Way South Viaduct over ARGO Railroad Yard, Seattle, Washington.”
Itasca, 2009. Itasca Consulting Group, Inc. Fast Lagrangian Analysis of Continua (FLAC).
Imai, T. and Tonouchi, K. (1982). “Correlation of N-value with S-wave velocity and shear
modulus,” Proceedings, 2nd European Symposium on Penetration Testing, Amsterdam,
pp. 57-72.
Mayne, P. (2007). “Cone Penetration Testing, A Synthesis of Highway Practice,” National
Cooperative Highway Research Program, NCHRP Synthesis 368, Washington D.C.
Shannon & Wilson (2010). “Final Geotechnical Report, Plans, Specifications and Estimates
Phase, Airport Way South Viaduct Over ARGO Railroad Yard, Seattle, Washington.”
Youd, T.L., Idriss, I.M., Andrus R.D., Arango, I., Castro, G., Christian, J.T., Dobry, R., Finn,
W.D.L., Harder, L.F. Jr., Hynes, M.E., Ishihara, K., Koester, J.P., Liao, S.S.C.,
Marcuson, W.F. III., Martin, G.R., Mitchell, J.K., Moriwaki, Y., Power, M.S., Robertson,
P.K., Seed, R.B., and Stokoe, K.H. II. 2001. “Liquefaction Resistance of Soils: Summary
Report from the 1996 NCEER and 1998 NCEER/NSF Workshops on Evaluation of
Liquefaction Resistance of Soils.” J. Geotech. Geoenviron. Eng. 127, 817.

© ASCE

Grouting 2017