a case study on the performance of embankments on treated soft ground (2003) - thesis (69(

A CASE STUDY ON THE PERFORMANCE OF EMBANKMENTS ONTREATED SOFT GROUND

by

Isabella Santini Batista

Diploma in Civil EngineeringUniversidade Catolica de Pernambuco - Brazil

Submitted to the Department of Civil and Environmental EngineeringIn Partial Fulfillment of the Requirements for the Degree of Master of Engineering

In Civil and Environmental Engineering

at the

Massachusetts Institute of TechnologyJune 2003 MASSACHUSETTS I

OF TECHNOLO

C 2003 Isabella Santini BatistaAll rights reserved LIBRARI

NSTITUTEGY

003

ES

The author hereby grants to MIT permission to reproduce and to distribute publicly paper andelectronic copies of the thesis document in whole or in part.

Signature of A uthor........................................ ...... ............. ..Department of Civil and Environmental Engineering

A - . t May 09, 2003

Certified by................................................... .....Andrew J. Whittle

Professor of Civil and Environmental EngineeringThesis Supervisor

Accepted by....................... .---..- OOral Buyukozturk

Chairman, Department of Committee on Graduate Studies

A CASE STUDY ON THE PERFORMANCE OF EMBANKMENTS ONTREATED SOFT GROUND

by

Isabella Santini Batista

Submitted to the Department of Civil and Environmental Engineeringon May 9, 2003

In Partial Fulfillment of the Requirements for the Degree of Master of EngineeringIn Civil and Environmental Engineering

ABSTRACT

The proposed construction of a new high-speed rail link between Amsterdam and Brussels

involves construction of embankments overlying soft clay and peat deposits with very stringent

requirements on the time frame of construction and allowable settlements. A field program of

instrumented test embankments, referred to as No-Recess, has compared the performance of five

schemes for stabilizing the soft ground behavior. This thesis summarizes the No-Recess project

and compares the measured performance for two instrumented using finite element analysis. The

results confirm the benefits of geotextile encased sand columns as a practical technique for

stiffening underlying soft soils, accelerating consolidation and reducing settlements.

Thesis Supervisor: Andrew J. WhittleTitle: Professor of Civil and Environmental Engineering

A Case Study on the Performance of Embankments on Treated Soft Ground

Acknowledgments

I would like to dedicate this thesis to the memory of my beloved grandfather Gracio

Barbalho and my cousins Mairson Bezerra and Eduardo Carvalho.

First and foremost, I would like to thank Professor Andrew Whittle, my thesis supervisor

and academic advisor, for introducing me to this topic and for guiding me through my work.

My warmest appreciation also extends to Dr. Lucy Jen and Professor Herbert Einstein,

who introduced me to the 'art' of soil and rock mechanics; to Cynthia Stewart, for helping me on

several occasions regarding academic and bureaucratic problems; to Robert F. Woldringh for

providing me the data for this research; and to all my friends at MIT, especially to Tina, Kartal

Toker, and Patrick Lee.

I am grateful to my parents, brother and sisters for their continuing love throughout my

life and to my parents-in-law for their encouragement in this challenging phase of my life.

Jozeane Bezerra helped me take care of 'our' Guiga - many thanks!

Finally, but most importantly, I offer my deepest gratitude to my beloved husband

Valdner Batista and my son Antonio Guilherme for 'everything': without their endless love and

support, finishing up this work would not have been possible.

3

M.Eng 2002 - 2003

A Case Study on the Performance of Embankments on Treated Soft Ground M.Eng 2002 - 2003

Table of Contents

I Introduction .......... .............................................. 82 No-Recess Project..................................................................................................................10

2.1 Objectives of No-Recess Project .............................................................................. 102.2 Site Characteristics ..................................................................................................... 112.3 Ground Improvements Options ................................................................................. 16

2.3.1 Conventional Test Embankment (HW 1)............................................................ 162.3.2 Stabilized Soil Columns (HW 2)....................................................................... 192.3.3 Stabilized Soil W alls (HW 3)............................................................................ 212.3.4 Geotextile Coated Sand Columns (HW 4) .......................................................... 232.3.5 Stabilized Soil Test Embankment on Piles (HW 5) ............................................ 25

2.4 M onitoring ..................................................................................................................... 283 Performance of an Unimproved Embankment ................................................................... 30

3.1 Site Characteristics ..................................................................................................... 303.2 Soil Parameters and in situ Stresses ......................................................................... 313.3 One-dimensional Settlement Predictions................................................................... 34

3.3.1 Calculations for undrained, initial settlement.................................................... 343.3.2 Calculations for Consolidation Settlement........................................................ 343.3.3 Calculation for Secondary Compression.......................................................... 353.3.4 Total Settlement................................................................................................. 37

3.4 Simulation using Plaxis .............................................................................................. 383.4.1 Finite Element model..........................................................................................393.4.2 M aterial Properties ............................................................................................ 413.4.3 Calculations ....................................................................................................... 42

4 Ground Improvement Options.......................................................................................... 494.1 Vertical W ick Drains ................................................................................................. 49

4.1.1 The effects of the smear zone in wick drains ..................................................... 524.2 Geotextile Coated Sand Columns...............................................................................55

5 FE Simulations of No-Recess Instrumented Embankments...............................................595.1 Conventional Design, HW I....................................................................................... 595.2 Predictions for Geotextile Coated Sand Columns (HW 4)........................................ 62

6 Conclusion ............................................................................................................................. 66

4

List of Figures

Figure 2-1: The Hoeksche Waard (HW) test site (Geotechnical Engineering for TransportationInfrastructure, 1999).............................................................................................................. 11

Figure 2-2: Soil profile and in-situ properties at HW site ............................................................. 13Figure 2-3: Surcharge sequence for constructing embankment HW 1, high section .................. 17Figure 2-4: Plan view of the embankment HW1.......................................................................18Figure 2-5: Cross-section LL of embankment HW1................................................................18Figure 2-6: Cross-section AA of the high embankment section of HW1 ................................ 19Figure 2-7: Plan view of the trial embankment HW2 .............................................................. 20Figure 2-8: Cross-section LL of the embankment HW2..........................................................20Figure 2-9: Cross-section AA of the embankment HW2 .......................................................... 21Figure 2-10: Plan view of the embankment HW3.....................................................................22Figure 2-11: Cross-section LL of the embankment HW3 ....................................................... 22Figure 2-12: Cross-section AA of the embankment HW3 ....................................................... 23Figure 2-13: Plan view of the embankment HW4.....................................................................24Figure 2-14: Cross-section LL of the embankment HW4........................................................24Figure 2-15: Cross-section AA of the embankment HW4 ........................................................ 25Figure 2-16: Plan view of the embankment HW5.....................................................................26Figure 2-17: Cross-section CC of the embankment HW5........................................................27Figure 2-18: Cross-section AA of the embankment HW5 ....................................................... 27Figure 2-19: Instrumentation of test embankments (Barennds, 1999)......................................28Figure 3-1: Initial effective stresses and piezometric head in the FE model.............................40Figure 3-2 Cross-section of a road embankment on soft soil...................................................41Figure 3-3 Tim e vs. height of fill .............................................................................................. 43Figure 3-4: FE meshes for unimproved embankment and location of 3 reference points on the

ground surface ....................................................................................................................... 43Figure 3-5: Total displacement after construction = 1.04m. .................................................... 44Figure 3-6 Total displacement after consolidation = 1.62m......................................................44Figure 3-7 Excess pore pressures at the end of construction = 87.OOkN/m 2 .............. . . . . . . . . . . . . . . . 45Figure 3-8: Settlement and Surcharge vs. time for the unimproved embankment....................46Figure 3-9: Horizontal Displacement in the toe of the embankment for the unimproved

em bankm ent...........................................................................................................................46Figure 3-10: Relationship between maximum horizontal displacement and maximum settlement

........................................................... 47Figure 3-11: Excess pore pressure at three elevations...............................................................48Figure 4-1: Em bankm ent with wick drains .............................................................................. 49Figure 4-2 Triangular pattern of installation of vertical drains (Terzaghi et al., 1996).............50Figure 4-3 Vertical drain of radius r., smear zone of radius r,, and consolidating soil cylinder of

radius r, (Terzaghi et al., 1996)......................................................................................... 53Figure 4-4 Relations between degree of consolidation and time factor for radial flow into a

vertical drain. Loading is assumed to be instantaneous and excess pore water pressureconstant with depth. (a) different drain spacing; (b) different mobilized discharge capacity(T erzaghi et al., 1996)............................................................................................................55

Figure 4-5 Geotextile coated sand columns .............................................................................. 56Figure 4-6 Process of installation of geotextile encased sand columns ................................... 58

5

. n 02-20MEng 2002 - 2003

Figure 5-1: FE model of surcharge vs. time and comparison between measured and predictedground settlem ents for H W l.............................................................................................. 59

Figure 5-2: Comparison between horizontal displacement predicted and that measured in thefield for H W I.........................................................................................................................6 1

Figure 5-3: Relationship between maximum horizontal displacement and maximum settlementfor H W ................................................................................................................................. 6 1

Figure 5-4: Excess pore pressures dissipation at three elevations in HW1 ............................... 62Figure 5-5: Imposed surcharge sequence and comparison of predicted settlement by plaxis and

that m easured in the field................................................................................................... 63Figure 5-6: Horizontal displacement predicted by Plaxis .......................................................... 64Figure 5-7: Relationship between the maximum horizontal displacement and maximum

settlem ent in H W 4 ................................................................................................................. 65Figure 5-8: Predicted excess pore pressure in three locations beneath the high embankment

(H W 4 )....................................................................................................................................6 5

6

. n,202-20MEng 2002 - 2003

A Case Stud on the Performance of Embankments on Treated Soft Ground

List of Tables

Table 2-1:Table 3-1:Table 3-2:Table 3-3:Table 3-4:Table 3-5:Table 3-6:Table 3-7:

M onitoring program ................................................................................................. 29Total head for each layer.......................................................................................... 31Engineering properties from lab and field testing at No-Recess site.......................31Engineering design parameters used to model soil behavior...................................31In-situ stress at No-Recess site.................................................................................33Vertical stresses due to embankment construction based on elastic theory.............33One-dimensional calculations of settlement for the high embankment...................38Material properties of the road embankment and subsoil........................................42

7

.M lng 2nn2 - 2003

1 Introduction

The new high-speed rail link between Amsterdam and Brussels will require new techniques

for construction of embankments overlying very soft soils. The principal factors affecting the

design of the embankments are related to the high train velocities, short construction time, and

very strict requirements on residual settlements. After analyzing these requirements, a research

team initiated a program in 1997 to evaluate different techniques for stabilizing railroad

embankments constructed on very soft clay and peat foundation soils.

The program, formally entitled "New Options for Rapid and Easy Construction of

Embankments on Soft Soils" (No-Recess), was a joint venture of the railroad project manager

High-Speed Rail South, the Ministry of Transport, Public Works and Water Management, Road

and Hydraulic Engineering Division (RWS/DWW), and the European Community (EuroSoilStab

lime-cement columns project).

The No-Recess project comprises a series of test embankments constructed in the Hoeksche

Waard polder near Gravendeel in the Netherlands. Five ground improvement techniques were

selected during a workshop in 1997 in Delft based on advice from a panel of international

experts. Direct comparisons of field performance were intended to demonstrate the viability of

the new methods for constructing the railway embankments for the new high-speed line. Since

there was no prior experience of these construction methods in the Netherlands, the instrumented

embankments were monitored for two years after the end of construction.

This thesis presents a summary of the No-Recess field project. A two-dimensional finite

element model is developed to represent the in-situ conditions and soil properties at the site for

one case of an unimproved embankment and two of the five ground improvement schemes used

8

M.Engz 2002 - 2003

in the No-Recess project: (1) using pre-fabricated conventional drains, and (2) reinforcement of

the soft soil with geotextile encased sand columns.

9

M.Engz 2002 - 2003

2 No-Recess Project

2.1 Objectives of No-Recess Project

The goal of the No-Recess research project was to investigate alternatives for constructing

embankments on soft clay foundations, which would provide a cost effective alternative to slab

track (i.e. concrete structure). The techniques in the No-Recess project are compared with a

reference technique of embankment construction using with a traditional free-draining sand fill

and pre-fabricated, vertical wick drains, together with surcharge loading to accelerate

consolidation. Since the purpose of the No-Recess project was to address design requirements

imposed by the high speed train, the following specification were formulated:

" Short construction time (less than 18 months)

* Low residual settlement (less than 30mm in the 30 years after 24 months after starting

construction)

* Minimum surplus of soil

" Sufficiently stiff behavior of the construction under dynamic loading by high "speed trains"

(mitigate potential problems associated with surface waves in the fill and sub-grade soils)

* Minimum impacts caused by widening of existing rail or road constructions

The purpose of selecting the new techniques was to fulfill the requirements to accomplish the

specifications above. Five instrumented embankments were constructed in the Hoeksche Waard

as follows:

" HW 1: Conventional test embankment with PV wick drains

" HW2: Lime stabilized soil columns

" HW3: Stabilized soil walls

10

A Case Study on the Performac ofEbnmnso rae oft Ground M.Ena 2002-- 2003

- -- a------~-- -- .N, %, %a J i jVUii .n1 1r 2 - LUJA Case Study on th~ P~rform~nr~ Of Fmh~nL~npntQ Tr~t,.A ~rd'~ ~

" HW4: Geotextile encased sand columns

* HW5: Stabilized soil test embankment on piles

The site characteristics and embankment construction methods are described in the next

sections.

2.2 Site Characteristics

The test site has a plan area of 400x125m, with level ground surface at EL -0.75m +

NAP*. Each embankment was constructed with a high part, 5m above the ground level, and a

low part, im above the ground level. Figurel illustrates the lay out of the five embankments

area.

- . 7C :- JE -

Figure 2-1: The Hoeksche Waard (HW) test site (Geotechnical Engineering for Transportation

Infrastructure, 1999)

* NAP: Niew Amwsterdam Piel, datum

11

ME 2

A Case Study on the Performance of Embankmet nTetdSf rud

The underlying soft foundation soils are typically 9.6m thick and comprise the following

main layers (Figure 2-2):

" Clay 1: average thickness of 3.1m of clay, very silt, moderately organic, grey, with sand

lamination;

* Peat 2: average thickness of 2.Om of peat, slightly clayey, wood fragments, brown, grey

(Hollandveen);

" Clay 2: average thickness of 3.Om of clay, extremely silty, moderately organic, peat traces,

wood fragments, grey;

" Peat 2: average thickness of 1.6m of peat, highly organic, brown, with sand lamination

(Basisveen).

12

A~~~~~~~~~ ~ ~ ~ ~ ~ ~ Cas S vothPefracofEbnmnsoTraeSotGud NA~~~~ 1 )A, n lMEn 2002 2003

A tfnt W1I I" %, m IU . n11 V - I 3J

Profile y(kN/m 3) Stresses and0-1

-2

-3

-4

z

c6 -

-7

-8

-9

I1 -WP Wn W1

Pore Pressure(kN/m 2)Wn, WP and W,(%)

50 100 150 200 250 300

0

e0

- 0

0

0

1011 12131415161718

Figure 2-2: Soil profile and in-situ properties at HW site

13

I

0 50 100 150

-0 (7' (kN/m2

-U-- u(kN/m 2)A ' P(kN/m 2)

-y- ' ,(kN/m 2

Peat 1

Peat 2

200

ME 2002 200

A

A

A

F

All of these layers are Holocene deposits. Below the Basisveen peat layer is a thick sand

layer comprising medium-fine, slightly silty sand, which is considered to be relatively

incompressible and free draining.

Many in-situ and laboratory tests were carried out before the construction of the

embankments. Fugro Ingenieursbureau B.V. (Masterbroek, 1998) performed the geotechnical

soil investigations.

In-situ tests carried out in the site consisted of the following considerations.

0 35 piezo-cone penetration tests (PCPT) with measurements of sleeve friction and pore

pressure to a depth of 15m below ground level.

0 35 dissipation tests at the end of each PCPT

The tests were performed with the electrical Fugro-sleeve friction cone unit (with a

capacity of 200kN). The dissipation tests were performed at the maximum depth, which are

usually done when the penetration process is stopped to allow the excess pore pressure to

dissipate.

0 6 mechanical soil borings with undisturbed continuous sampling, 12m below ground level.

The borings have been carried out using a truck/crawler type mounted drilling rig (according

to the Dutch standard NEN 5119). Undisturbed samples have been taken during the borings to a

maximum depth of 12m below ground level. After the samples were taken in the thin-walled

Shelby tubes with a diameter of 70mm and a length of 400mm, the samples have been

transported to Fugro testing laboratory in Amhem and stored in a temperature conditioned room

before laboratory testing.

14

M.Eng 2002 - 2003

0 36 field vane tests (FVT) at four locations, 9 meters below ground level, at Im vertical

intervals, performed close to the locations of the piezocone tests beneath embankments 1

through 4.

* 16 cone pressuremeter tests (CPM) at four locations, 9 meters below ground level, depth

interval of 2m, performed close to the locations of piezocone tests (DKMP 1-1, 2-1, 3-1 and

4-1) CPM 1-1, 2-1, 3-1, and 4-1.

0 Installation of two fixed cone points to a depth of 15m below ground level in the south-

western and south-eastern corner of the test site, coded (FP 6-1 and FP 6-2)

* Installation of 6 standpipe piezometers at two locations with 3 depths of filter screen (each

Im long) indicated as B 6-1 to B 6-3 and B 6-4 to B 6-6, the standpipes were installed at 2

locations (3 in each location) closed to the Fixed Cone Points. The filters were located at

depths of 3.5m, 5m and at 10m with respect to the ground level. The filters were isolated

with clay pallet packers and the ground water was measured after a period of time.

The laboratory test program was carried out on samples at lm vertical intervals and

consisted of:

" Description of the soil

" Photographs

" Water content

* Bulk density

" Density of solids

" PH values of the soil

" Atterberg limits

* Particle size distribution (one per boring)

15

Some more specialized laboratory tests also were carried out in the site, including:

" Organic content (Im intervals)

* Carbonate content (1 m intervals)

" Fall cone strength (tm intervals)

* Sulfate content (Im intervals)

" Isotropic consolidated undrained multistage triaxial (3 per boring)

* Oedometer compression tests (3 per boring)

The in situ vertical stresses that are shown in Figure 2-2 take into account the artesian pore

pressures in the underlying sand layer.

2.3 Ground Improvements Options

The five embankments were designed to have a high part (5m above initial ground surface)

and a low part (1 m above initial ground surface) with side slopes of 1:2 (v:h) with a transition

zone of 10 m located at the end of the high part. Each embankment was to be constructed within

eighteen months (after the working platform was made), followed by a period of six months prior

to assume surface construction (rail bed installation). The residual settlements expected after 24

months from the start of the construction must be less than 30mm over a 30-year period.

2.3.1 Conventional Test Embankment (HW1)

The conventional test embankment (with pre-fabricated, vertical wick drains) was

constructed to provide a base reference on performance compared with the other test

embankments. The conventional embankment was constructed in stages within a 6-month

period. The surcharge of the low embankment was 1.8m of sand whereas for the high

16

M.Engi 2002 - 2003

embankment was 2.5m of sand, both for a year. In both, low and high embankments, the vertical

drains were installed to a depth of 1 m above the sand layer and were designed with 1 m triangular

grids. Figure 2-3 shows surcharge sequence used to construct the embankment with vertical

drains. Figures 2-4, 2-5, and 2-6 illustrate the details of the embankments and also show the

locations of the important instruments stated earlier in this chapter.

E

0)

(0

5

4

3

2

010 Time(days) 100 1000

Figure 2-3: Surcharge sequence for constructing embankment HW1, high section

17

M.Engz 2002 - 2003

1

1

A Cni 4Ztimd nn thi- Prfnrmi~r, AfF m..sp~Q ~ ~ ~ N AV kAl Ir A6 Q 4,l fI A~l~ KAI CJUUI ilL I ASt I.AMI.

!-5

.6 o0-, 1 -21 '-24 ~ 33 5 3 -

1- A'- I 11-317 41-741 0l-20 21-1-12 -,/ 9- -

/ V TO.: 71K 1- 22 12513

-23 -3

a ~12

'-40

Figure 2-4: Plan view of the embankment HW1

2a~C1 200002000C 4

j SEIT

ittIQ

SE~~h V.R1KALE DRAJNt HO.- 1'rv JE E SRAII N

Figure 2-5: Cross-section LL of embankment HWI

18

1-6

*,-ZS

-3

-6

A Cae Sudy n te Prforanc ofEmbakmets o Trate Sof GrundM Ency 2A(Y - IQA(Y

S0020=2 200

Figure 2-6: Cross-section AA of the high embankment section of HW1

2.3.2 Stabilized Soil Columns (HW2)

The stabilized soil columns method used in HW2 is the Scandinavian lime-cement

mixing method of stabilizing clay soil mixing 200kg/m 3 of new binder mix comprising 80% of

blast furnace cement and 20% anhydrite (instead of the combination of Portland cement and

unslaked lime that would be used in Scandinavia). This mixing method of stabilization

technique (dry mix technique) uses dry air to transport the binder. The reason for the change in

the binder mix is to have better performance with this new mixture in the very soft Dutch soils.

The embankment was constructed in 1 month. For both, high and low embankment, a

preload of respect to 1.5m and 1.Om for a period of 260 days was needed to accelerate the

consolidation process. The lime columns were 600mm in diameter and were placed in 1.6m

square grids for the low embankment whereas for the high embankment they were placed in the

range of 1 m to 1.2m square grids. For both cases, the columns extend 0.5m into the underlying

sand layer. The low embankment uses stabilized blocks (overlapping short columns) of 1.5m,

which were produced with a speed of 500m per day. More details can be seen from Figures 2-7,

2-8, and 2-9.

19

A Case Study on the Performance of Embankments on Treated Soft Grud

A Case Study on the Performance of Embankments on Treated Soft Ground M Eno, 2002 - 2003

-------

r unrT s

9

Figure 2-7: Plan view of the trial embankment HW2

~r~rF1rrrFnr7~tF nir ~nr~rri F

L~ L L it

Figure 2-8: Cross-section LL of the embankment HW2

20

A C~i'~e Afl thc~ PPrfArm~riPP r~fFmnvi..nt~~ t~.i Trj~x~ti~.A Qr.4~ flrr.,,..A- - - ~**n "* 'J-sue n Utnu 1V.ndlg~J~ - ~.U

PW 7C~-NA

Figure 2-9: Cross-section AA of the embankment HW2

2.3.3 Stabilized Soil Walls (HW3)

The technique used in the test embankment HW3 comprises the stabilized soil walls

installed by the FMI process (Barends et al., 1999), a process of constructing a soil mix wall

using a specialized cutting tree device. The machine has a speed around im/min with a

maximum depth of 9m and width of 500mm. The cutting tree device is inclined up to 80 degrees

and is dragged behind the FMI machine. Due to the cutting blades rotated by two chain systems,

the soil is not excavated but mixed in place with cement slurry. The embankment was built in

two weeks.

In the high embankment, a load transfer platform was used, which is placed at the height

of 0.5m with 3 geotextiles (tensar geogrid type SS20, SS30 and 80RE with crushed rock). The

stabilized walls, which have dimensions of 2 x 500 mm, are placed at an interval distance of 1m

to 2.5m in both embankments, but in the case of the stabilized block, it is placed at a depth of

1.5m only in the low embankment. The binder dosage used for both stabilized walls and

21

ME 2002 2003

A Case Study on the Performance of Embankments on Treated Sof Grud M Pnv ~ -

stabilized block is 150 kg/m3, where 80% consists of blast furnace cement and 20 % of

anhydrite. Figures 2-10, 2-1 1,and 2-12 show the details of the embankments.

3-4 -

1443-

3-, - 34

--2--

I4 i a3- 11 2.2 3 37 3-V 3

~44- 44 4-3 -~ * 1-4 31 414 ~ 4- ~ ~ ~'4- -54 -4 44' 3.4

_______ ~loon--- O1

3-3 3~~~ ~~~~- 1 4 5 - ~ 1 ~ ~ 3

SE11 3- 3414x -3

3-' -3 34 54- 0 342

1-4 3-4

3-1 33 ~4 344 3-V 3-33RN " "

3-34Dal 3- 41

33353-

Figure 2-10: Plan view of the embankment HW3

Figure 2-11: Cross-section LL of the embankment HW3

22

I OCCO

if (q.O Mve Cno( 60CC- + M)

voe' (C 340

z

I tS"2X

MV'

Dr3 2000 CCC0 2000 CC, 2000 000 C0CICCC~

De:Wceo~

2 NAP G W S 700 - MV

NAP

PR 4AP

inc 1030_1000 2200 1000 2000 -000, 2000 0CC 200

_ _ _ __T


2.3.4 Geotextile Coated Sand Columns (HW4)

Geotextile coated sand columns (GCC) system is the technique used in HW4. This

process consists of vibrating a casing with two valves at the bottom until it reaches the sand

layer. A 40-cm thick layer of sand is made, a geotextile stocking is installing and filled with I m

of betonite/sand mixture to create a geo-hydrological barrier (12%). Then, the rest of each

800mm diameter column is filled with sand to the top of the casing, which is vibrated and pulled

upwards, compacting the sand. This system can have a rate of construction of 40 columns per

day. The embankment with a load transfer platform of geotextiles in the HW4 test was

constructed in 6 weeks.

For both embankments, the depth of the columns is 9m below ground level and the

distance between columns is 2.4m to 3.4m triangular grids for the low embankment, whereas for

the high embankment, it is in the range of 2.Om to 2.4m. For the low embankment, a surcharge

23

Sr-S

I

-A P nl IA Case Study on the Performance of Embankments on Treated Soft Ground MEne 2002 2003

A Case Study on the Performance of Embankments on Treated Soft Ground M.Engz 2002 - 2003

of I m of sand for one month and one geogrid Fortrac 80/80-10 were used, whereas for the high

embankment, they used three geogrids Fortrac 200/30-30 and no surcharge was necessary. More

details can be seen in Figures 2-13, 2-14, 2-15.

XT. 3 VD__ No -

C~0 0.0 0 C Ca ~ ~ ~ ~ ~ ~ W 'a .'0C~c ~ hm..~

0 0 W-3 a~ ~2~,t w" t'~00 0

0 0~~ 0)

*1 4) ~ ~ 0- -~ C Pi0 C C 0 .0 t 11 0k

C2 .-'-2 0 PI -

Figure 2-3 Plnve0fteebnmn W

V G-- 0 A-

Figure 2-14: ~~ CrsssctinL f h manmn W

24

10000 100CC

000 -

7m '00- - \P

970 170

31520


2.3.5 Stabilized Soil Test Embankment on Piles (HW5)

The piled foundation in the technique used in the HW5 trial is made with wooden piles

and AuGeo pile system (Barends et al., 1999), and the sand used in the embankment is replaced

by stabilized soil from another site, which is made by using mix-in place plant installation

(ARAN installation). The embankment was built in six weeks, with a production rate of 200

piles a day.

The AuGeo pile system basically consists of a Cofra stitcher that pushes a steel plate into

the ground with a casing (1 80x 1 80x6.3mm) using maximum pressure of 25tons. After that, a

PVC pipe, which is sealed at the bottom, is inserted in the casing and filled with the foamed

concrete with unit weight of 1200kg/m3. The last step of this process, after having sufficient

hardening of the concrete, is to lift the casing, cut off the PVC pipe, and cover the top with a

concrete tile of 300x300mm.

25

.A Case Study on the Performance of Embankments on Treated Soft Ground MPng 2MW -2003

~ - ~ %1 i m - nt' 5I l 4U Cl iuUEU M.Eng 2 - .3

The depth of the piles for the low and the high embankment is 12m below the ground

level but the pile type used for the low embankment was AuGeo pile system only, whereas for

the high embankment wooden piles also were used. The distance of piles in the case of the low

embankment is Im square grids, but for the high embankment a 0.8m square grid is used.

For the low embankment, Im of stabilized soil is used with a binder dosage of 120kg/m3

where 80% is blast furnace and 20% is anydrite, whereas for the high embankment 5m of

stabilized soil is used. The load transfer platform consists of two geogrid Fortrac with a thickness

of 30cm and with a fill of crushed concrete for both cases. The embankments can be better seen

in Figures 2-16, 2-17, and 2-18.

howter paler SchuimoeOnpolen

5-7 5-9 5-11

eta A55-5-95-1721

5-42

5-14 -210000 0 0

5- 2 3 5-8 5-

50005- 1000 5 8 5-0 10g551

'Ice 3 $Ic10-000

Figre2-6:Pln ieCo the emakmn H 51-

5-23

5-26-

3~4 5-24

5-220000 109

26

002 200

ACase Study on the Performiance nf Fmhnnk~mpnt,_ nn Trp,,tpi 4Ztff I-gr% nAl---------...~-.~...... iV.L~ "Y'J - v~

:aer pen

4 crvor £'A C

/1jL47

OC0 Asets

-m ' -. --

KQ

1%>

750025000

4 50C 6c00 4 50

500--, zoo3:C0--- -lso. -530- :a:

9"C - "t ''

et

r

300--N3

4 4,- P

Figure- V w x B Z 21 Ct

Figure 2-18: Cross-section AA of the embankment HWS

27

Figure 2-17: Cross-section CC of the embankment HW5

R243,'rW, R32LH/5!v

XM= ---------- P"

:30 m 3 g G cp r

40C,-na

7

ME 2002 2

A ae Sud n t fomane o Emankent onTretedSof GrundiAt.- ')A

2.4 Monitoring

Each of the five embankments was monitored to provide detailed information on vertical

and horizontal soil displacements, pore pressure, and vertical stresses. Each design is expected

to satisfy stringent rail requirements for high-speed embankments (with reduced construction

time and long-term deformations). The monitoring was carried out from the start of construction

through the end of 1999. The instrumentation for monitoring the embankments was installed

after the construction of the foundation and before the construction of the embankments. The

monitoring program for each embankment is shown schematically in Figure 2-19, while the

rotation for the instruments is given in Table 2-1.

54M Z(NAP)

t~c xL

~=iL1L=rFigure 2-19: Instrumentation of test embankments (Barennds, 1999)

28

-A IQ I ~ nA Case Study on the Performance of Embankments on Treated Soft Grud MEn 2002 2003

High

A Case Study on the Performance of Embankments on Treated Soft Ground MEt 02-20

Since each technique has its own features, each embankment trial has a different

monitoring starting time in 1998 as follows: February 20 for HWI, June 29 for HW2, March 27

for HW3, July 17 for HW4, and September 29 for HW5. The description and location of the

monitoring program for each embankment can be seen from previous figures shown in this

chapter.

The monitoring time is based on the average construction time of 3 months. A

monitoring time of 100, 300 or 600 days corresponds with a construction time of 6, 12, or 24

months. The settlement S =(Z - Z,) will be monitored and evaluated. The logarithmic rate of

settlement LRS = can be used as an evaluation parameter for the residual settlement

log(

(Barends et al., 1999).

Description Instrument Location

SETT Settlement top layer 5x5m grid at the top of embankment

SETH Settlement hose longitudinal section and low and high cross-section

SETP Settlement plate longitudinal section every 10m and 2 cross-section

EXTM Extensometer 5 levels in soft layer under high embankment

INCP Inclinometer 3 in cross-section high embankment

PWSP Pore water stand pipe 1 at ground level under high embankment

PWPT Pore water pressure transducer 3 levels in soft soil layer

SPCL Soil pressure cell 1 beside the column and 1 at the top

Table 2-1: Monitoring program

29

M.Enix 2002 - 2003

3 Performance of an Unimproved Embankment

This chapter considers the behavior (or a hypothetical) unimproved embankment with the

same site characteristics, soil parameters, and in-situ stresses as the No-Recess project, in order

to provide a basis for evaluating the performance with the ground improvement solutions.

3.1 Site Characteristics

Since this analysis is being done for the same site in the Netherlands, an analysis of this

site was made and a general profile was designated for the study of the unimproved

embankment. This site basically consists of 9.7m of soft soil underlying a relatively

incompressible, artesian sand layer. The ground surface is at EL.-O.9m and the ground water

table is located at the elevation -2.2m.

Three piezometers were installed in the high section of the embankment, at elevations of

-2.7m, -5.Om and -7.5m. The pore pressure was calculated from the results given by the

piezometer installed in the top of the sand layer (EL.-7.5). The piezometer gives the total head in

the top of the sand layer as +1.25m and the hydraulic conductivity in the clay and peat layers, is

k=8.64x1 0 5m/day and 8.64x 104 rm/day, respectively. After some calculations, it has been

inferred that there is minimal head loss in the peat layer and the head loss in the clay layer is

0.72m/m, the total piezometric heads in the layers are summarized in Table 3.1.

30

Elevation Stratum Total Head

-2.2m (wt) Clay 1 -2.2m

-4.Om Clay 1 -0.9m

-6.Om Peat 1 -0.9m

-9.Om Clay 2 1.25m

-10.6m Peat 2 1.25m

Table 3-1: Total head for each layer

3.2 Soil Parameters and in situ Stresses

The soil parameters were based in the parameters given in the report of the No-Recess

Project (Masterbroek, 1998). Calculations and correlations were done to find other parameters to

be able to analyze the unimproved embankment.

Stratum Thickness(m) wn(%) wI(%) wD(%) suFV(kN/m 2) eClay 3.1 66.4 160 81.9 38 1.9Peat 2 98.5 267 186 45 4.8Clay 2 120 160 81.9 37 1.9Peat 1.6 131.6 267 186 36 1.9

Stratum cc CR RR c c(m 2/year) k(m/day)Clay 0.27 0.093 0.0093 0.017 2 8.64E-05Peat 1.33 0.23 0.023 0.034 3 8.64E-04Clay 10.33 0.113 0.0113 0.011 2 8.64E-05Peat 0.45 0.155 0.0155 0.01 3 8.64E-04Table 3-2: Engineering properties from lab and field testing at No-Recess site

Stratum 0'(0) Ko(0 ) c'(kPa) E(kPa) V' 4Clay 18 0.69 40 650 0.35 0Peat 15 0.74 38 295 0.35 0Clay 20 0.66 38 610 0.35 0Peat 15 0.74 37 530 0.35 0Table 3-3: Engineering design parameters used to model soil behavior

31

M.Engz 2002 - 2003


The drained young's modulus (E') and apparent cohesion (c') parameters were calculated

from the available soil data as follows:

. Drained young's modulus.

By definition the one-dimensional modulus,

compressibilityCR

M, = 0.435 ,' , and G'vave is

D = -, where m, is the coefficient of volumeMv

the average vertical effective stress during the

loading event.

Equating D with the constrained modulus from elastic theory.

E'(l-v') -045CR(1 + 0.43v5 - ) , where CR is the measured(I + v')(1- 2V) (71'A

compression ratio, u' is the possion's

ratio, and E' is the drained young's modulus.

* Apparent cohesion

The Mohr-Coulomb, the undrained shear strength su (in plane strain), can be related to the

drained shear strength parameters (c',$') as follows:

s, = c'cos#'+p'sin #'

where p'= -(1 + K, )o',,, and2

KO = coefficient of earth at rest

Assuming su is reliably measured in the field (by the field vane), a'v, is also well known,

and $' has been reported from laboratory tests (Masterbroek, 1998), then:

S s14 p'sin#'cos ' Cos b'

32

M.Eng 2002 - 2003

Elastic stress analyses were used to estimate the vertical effective stresses in the subsoil

due to embankment loading at ground level.

3 2 2 G 'p(kN/M2)_Elevation(m) Stratum Y(kN/m 3) u(kN/m2) ao(kN/m2) a'ydkN/m2) Lab

-0.9 Clay1 14 -12.7 0 12.7-1.2 Clay 1 14 -9.81 4.2 14.01-2.2 Clay 1 17.5 0 21.7 21.7 18-3 Clay 17.5 13.3 35.7 22.4-4 Clay 1 17.5 30 53.2 23.2 44-5 Peat 1 10.5 40 63.7 23.7-6 Peat 1 10.5 50 74.2 24.2 42

-7 Clay 2 14.8 66.6 89 22.4-8 Clay 2 14.8 83.3 103.8 20.5 50-9 Clay 14.3 100 118.1 18.1-10 Peat 2 11.8 110 129.9 19.9 96-10.6 Peat 2 11.8 116 136.98 20.98Table 3-4: In-situ stress at No-Recess site

&Vf(kN/m 2) -1D O'f(kN/M2) -1D & 'f(kN/m 2) -2D G'A(kN/m 2) -2DLow High Low High Low High Low High

20 134 20 134 33.1 133.8 45.8 146.520 134 20 134 34.3 133.4 48.3 147.420 134 20 134 35.5 133.0 57.2 154.720 134 20 134 36.5 131.9 58.9 154.320 134 20 134 38.9 131.2 62.1 154.420 134 20 134 42.6 130.6 66.3 154.320 134 20 134 46.3 130.1 70.5 154.320 134 20 134 48.9 129.9 71.3 152.320 134 20 134 51.4 129.8 71.9 150.320 134 20 134 53.8 129.0 71.9 147.120 134 20 134 54.8 128.0 74.7 147.920 134 20 134 55.9 126.9 76.9 147.9Table 3-5: Vertical stresses due to embankment construction based on elastic theory

33

M.Eniz 2002 - 2003

3.3 One-dimensional Settlement Predictions

3.3.1 Calculations for undrained, initial settlement

The calculations for the elastic settlement were based on the elastic theory of Janbu et al.

(1950) and corrected by Christian & Carrier (1978). In the case of multi-dimensional

deformations and settlements for more complex layer structures and/or complex load

configurations, finite element programs employ realistic stress deformation relationships. The

elastic settlement are computed from

qBEU

where q = applied load (considering the crest of the high embankment)

p and i are factors accounting for the foundation embankment depth and compressible layer

thickness, respectively.

q = applied load

B = width of the foundation

The undrained young's modulus, Eu, can be estimated approximately from the drained

modulus assuming vu = 0.5

3.3.2 Calculations for Consolidation Settlement

Calculations for consolidation settlement were based on the following equation:

a' p o'vf~Pcf=X Hi RR log , +CR log ,

a vo th P

where Hi = thickness of the layer

34

.MEng 2002 - 2003

RR = recompression ratio

CR = compression ratio

C'R = initial effective vertical stress

Y'vf= final effective vertical stress

T', = pre-consolidation pressure

3.3.3 Calculation for Secondary Compression

The calculation for the secondary compression were based on the following equations:

tA = Ca log-

t,

where ca = secondary compression index

t = duration of load

tp= end of primary

Tables and graphs based on simple differential equations usually predict consolidation

behavior under a fill area as a function of time. One-dimensional consolidation problems can be

solved using general calculations, but computer programs are required in the case of two-

dimensional consolidation problems.

In the case of one-dimensional consolidation, Terzaghi's consolidation theory explains how

saturated soils compresses with time under some restrictions such as

" the soil is homogeneous and saturated,

" the granular material and water are incompressible,

35

M.Eng 2002 - 2003

" a linear relationship exists between compression and increasing of effective inter-granular

stress,

" the consolidation coefficient is constant throughout the consolidation process,

" the compression is small related to the thickness of the layer.

Su32u

, &2

where, cv = KV

u = excess pore pressure (kPa)

t = time (sec)

c, = vertical coefficient of consolidation (m2/day)

z = coordinate in z direction, depth (in)

kv vertical permeability coefficient (m/s)

mv= vertical coefficient of volume compressibility (m2/day)

7w unit weight of water (kN/m 3)

T cAt(H,1)2

where T = time factor

t = time duration (sec)

a = drainage constant, a = 1 for one way drainage and a = 0.5 for two way drainage

h = layer thickness (in)

The relationship between the average degree of consolidation U and time factor T can be

expressed graphically or by numerical expression.

36

. n,202-20MEng 2002 - 2003

C~i~i~ ~tiuI~, nn th~ Pprfr~rn-i~.np~. ~ m,~j o~.

-*-,*..-,~, .. *..-, . ~L~UIJNIEi~jIL~ U~ I IV4L~U 3UIL ~IIUUIIU~ ~" %1 %.~ .II C.~L11s m mmemS on I eae S:>on Gwroud M.Eng 2002 - 200.3

The end of consolidation happens when all excess pore pressure has totally dissipated, in

theory t = oo. For a practical end of consolidation, an average degree of consolidation U=90% is

widely used. Therefore, the end of primary can be calculated as:

0.848(Hd)2

P Cv

where, tp = end of primary time

cv = vertical coefficient of consolidation

Even though the calculations for one-dimensional consolidation can be analyzed

numerically, calculation for two-dimensional consolidation process requires computer programs.

A finite element program has the ability of giving the solution for plane strain consolidation

problem, simulation of a non-homogeneous soil structure, and force, displacement calculation

together with a flow rate or water pressure.

3.3.4 Total Settlement

The total settlement is calculated by adding the elastic undrained, primary and secondary

settlements. It is done by sub-dividing the soil profile into four layers, with soil parameters cited

in Table 3.2. Table 3.6 summarizes the hand calculations of one-dimensional settlements for the

high part of the embankment. The height of the fill assumed for this calculations is 6.7m and the

unit weight used is ytiI= 20kN/m 3.

37


1-D High Embankment

p _ 1 q(kN/m 2) Bm) Ea(kPa) P,(M)1 0.1 134 35 520 0.902

Layer CR RR thickness(m) a'yo(kN/m 2) a',(kN/m2 ) a'p(kN/m 2 ) pD(M)

Clay 0.093 0.0093 3.1 21.7 121.7 31 0.176

Peat 0.23 0.023 2 23.7 123.7 42 0.227

Clay 0.113 0.0113 3 22.4 122.4 50 0.144

Peat 0.115 0.0115 1.6 19.9 119.9 96 0.030__ _ _0.577

Layer thickness(m) Hd cv(m 2/year) tp(years)Clay 6.1Peat 3.6 5.55 3 8.66

Layer co eo thickness(m) t(years) t,(year) log t/tQ p0 (m)

Clay 0.017 1.9 3.1 30 8.66 0.540 0.010Peat 0.034 4.8 2 30 8.66 0.540 0.006

Clay 0.011 1.9 3 30 8.66 0.540 0.006Peat 0.01 1.9 1.6 30 8.66 0.540 0.003

0.025

pt(m)=pe+pp+ps= 1.500Table 3-6: One-dimensional calculations of settlement for the high embankment.

3.4 Simulation using Plaxis

The one-dimensional hand calculations suggest a total settlement of 1.50m. Two-

dimensional analyses of the embankment can be studied using finite element program. As stated

earlier, in order to perform two-dimensional calculation, it is necessary to have the tools of a

finite element program. The FE program can also account for real time consolidation occurring

during phases of embankment construction. During the process of the dissipation of excess pore

pressure, the soil obtains the shear strength needed to continue the construction stage process.

38

M.Eng 2002 - 2003

3.4.1 Finite Element model

This part of the chapter covers the consolidation analysis of the embankment on soft soil.

Figure 3-2 illustrates the cross-section of the embankment. The embankment is 46m wide, and

6.7m high (this accounts for the actual height of the fill placed) with a side slope of 3:1 based on

the settlement vs. time graph in the No-Recess project report.

The embankment itself is composed of compacted sandy fill. The subsoil comprises

9.7m of soft soil sub-divided into four layers. The underlying artesian sand layer is not included

in the model but is treated as a rigid base with prescribed piezometric head, H=+1.25m. The

model calculations assumed that phreatic surface coincides with the original ground surface, and

the total head calculations were based on the results of the piezometer located on the top of the

sandy layer. This model assumes a steady upward flow through the soft clay and peat layers.

Figure 3-1 summarizes the initial effective stresses and piezometric head in the FE model.

.39

A Cae Study An th~ PPrfArm2ncP nfFmhnn1m~nrc ~ Tr~t~A ~,-.44 fl..~-....A- Ca-s - - -S onn r 'Ui VJuII M.EJng 2.0U02 - 201J3

Effective Stresses(kPa)

-1

-2

-3

Profile

0 5 10 15 20 25 -1.5 -1.0 -0.5 0.0 0.5

Figure 3-1: Initial effective stresses and piezometric head in the FE model

40

- o',(kPa)-- a',.(kPa)

Total Head(m)

-4

Peat 1-5

-6

CL

z

C0

(VE

-7

-8

-9

Peat 2-10 -

-11

K1.0 1.5

i ,

A Cae Sudyon te Prfomanc ofEmbnkmnts n Teatd Sot Goun M~n 202 -20.

46m

27m 27m

Cs B 41Say I 9

11LR I I ClayII

.7m

-qn-H- 1H- -

Figure 3-2 Cross-section of a road embankment on soft soil

The embankment was analyzed as a plane strain FE model using 15-node (cubic strain)

triangular elements. Lateral boundaries are set 27m from the embankment where deformations is

expected to be negligible, the lateral boundaries are also specified as closed flow boundaries,

imposing one-dimensional vertical flow conditions far from the embankment. In the initial

conditions, the embankment must not be present in order to generate the initial conditions. The

suggested Ko-values of the layers were calculated based on Jaky's formula.

Ko =1 - sin #

3.4.2 Material Properties

The properties of the different soil types are given in the Table 3-7 where five material

sets have been created. The type of 'undrained behavior' as set up for the clay and the peat

layers enables simulation of excess pore pressures due to embankment construction.

41

A Case Studv on the Performance of Embankmet onTetd9 rud M-Env 2002 - 2003

A Ca~ ~tiidv nn rh' P~rf~rm~nr~. nf Fm ~ ,~., ~ Q,..44 ~

Parameter Fill Clayl Peati Clay2 Peat2Material Model Mohr-Coulomb Mohr-Coulomb Mohr-Coulomb Mohr-Coulomb Mohr-Coulomb

Type of behavior Drained Undrained Undrained Undrained UndrainedUnit Weight (kN/m 3) 20 17.5 10.5 14.8 11.8Permeability (m/day) 10 8.64E-05 8.64E-04 8.64E-05 8.64E-04

Young's modulus (kN/m2)2.50E+04 6.50E+02 2.95E+02 6.1OE+02 5.30E+02Poisson's ratio 0.35 0.35 0.35 0.35 0.35Cohesion (kN/m2) 1 40 38 38 37Friction angle (degree) 35 18 15 20 15Dilatancy angle (degree) 0 0 0 0 0Table 3-7: Material properties of the road embankment and subsoil

3.4.3 Calculations

The embankment construction was simulated in 5 stages based on the reported sequence

used for the reference embankment (HW 1). Figure 3-2 shows that partial drainage was modeled

during each fill phase (using specified time increments for loading) with intervening

consolidation phases (no change in fill height).

'Undrained' - refers to the capability of modeling excess pore pressures within a soil layer in Plaxis; 'undrained'materials can represent the full range of drainage conditions within the soil depending on the type of analysis and/ortime frame specified in the calculation steps.

42

.L 111%,11 a VIj IVIM11- V JOU11 . 1jr,A Case Stud nn th,- P,-rfnrm5,n,, ,,f Pmhnl, + Ir A Q 4 t-- A %A U Inn') Infill

A Case Studv on the Performance of Embankments on Treated Soft GroundM n AY-9Af

80-

7.5 -

7.0-

6.5

6.0 -

5.5

5.0-

4.5.

4.0-.

-3.5-

i3.0 -

2.5-

2.0-

1.5.-

1.0 -

0.5 -

0.0 -

-0.5 -

Construction and Consolidation time(days)

0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000

Time(days)

Figure 3-3 Time vs. height of fill

Long-term embankment behavior was examined by allowing consolidation to proceed

until the minimum excess pore pressure, Au=lkN/m*

A

54 B

1

t/' 4 . 4 . tI 4 . . t~~ . 1~ 4 '++

4.4 4..44 ++494 4.4 4.4 +..44 +..4 +.4 +. +4 4. t 4.444 +..4 t..44 t..44 4.4 4.44.4.4.4.4.4.4.4.4.4.4.~~~~~~~"1 + +............ 4444444444444444444444444t..44 + .. 44 4.+ +.4 +..44 + .. 4 +. 4 +4.4++ t .. 4 ... 44 . 4. ... +..4 +. 4 4.4 +.. 44

A BFigure 3-4: FE meshes for unimproved embankment and location of 3 reference points on the ground surface

* The current analysis does not consider secondary compression and hence, does not simulate measured creepproperties of the clay or peat layers.

43

5

3

2

M'Fncr'?00') -')(InI

+ + +t + + t t +

+ + + +

A Cae Sudyon he erfrmane Q~mbnkmntson reaed Sft roud N 1~ )(A).

Figure 3-5 clearly shows that the settlements of the original surface and the embankment

increase significantly during the last phase due to the dissipation of all excess pore pressures,

which cause consolidation of the soil. It can be concluded that even though consolidation

already occurs due to the time interval needed for the construction of the embankment, there is

still a great amount of settlement occurring in the last phase (consolidation phase).

I

Figure 3-5: Total displacement after construction = 1.04m.

MV -Io6

I

Figure 3-6 Total displacement after consolidation = 1.62m.

44

-A P 1M I l

6 6

i

A Case Study on the Performance of Embankments on Treated Soft Ground MEne 2002 2003

Figure 3-7 illustrates the remaining excess pore pressures at the end of construction.

Au indicates how much consolidation has occurred during construction. So, it can be

concluded that 65% of consolidation occurred during the construction time.

Au = 87.OOkN/m2

Aav= Yt HfI = 134.OOkN/m 2 , therefore;

Au 87- =65%

Aoa 134

Figure 3-7 Excess pore pressures at the end of construction = 87.OOkN/m 2

Figure 3-8 illustrates the surcharge sequence used for FE analysis and settlement

behavior vs. time predicted in the centerline. Figure 3-9 shows the horizontal displacement

predicted in the toe of the embankment predicted during the construction and consolidation

phases.

45


A Cnctp 4Zn n.n th- DPrfnrmn,.p 4 m P ~ t. kT1 f r A Q,.A - K A 1C, Al' IAlJt~ . all mIs 1ILa '. .JJUUEI iI AU1 . "6 -

I--E

CL

a)0

8

6

4

2

0

-2

-410 100

Time(days)1000

Figure 3-8: Settlement and Surcharge vs. time for the unimproved embankment

0

-2 4--

-4

-6E)

-8 -

-10 +- -

-- - -- - - - -e - Day 25-0 Day 50-0 Day 85

- -e-Day 90,A- Day 120

M- Day 130

- 0 ; -a -- y -18--0-0 Day 5000

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Horizontal Displacement(m)

Figure 3-9: Horizontal Displacement in the toe of the embankment for the unimproved embankment

46

s I er

10000

-12

.

ME 2002 2003

I I


E

EE

O D0 0.2 0.4 0.6 08 1.0 1.2 1.4 15 1.8

Madimun Settlement (m)

Figure 3-10: Relationship between maximum horizontal displacement and maximum settlement

Figure 3-10 illustrates the relationship between maximum horizontal displacement and

maximum settlement for the unimproved embankment. In this correlation for unimproved

embankment, significant drainage during initial loading causes less horizontal displacement than

predicted from undrained analysis (Ladd, 1991).

Figure 3-11 shows the excess pore pressures at three elevations where the piezometers

are installed such as EL.-7.5m, -5.0m, and -2.7m.

47

M.Eng 2002 - 2003


100

80Eze 60(DL..

L- 40CL

0

20Ch,

x

0

0.1 10 100

Time (Days)

Figure 3-11: Excess pore pressure at three elevations

48

1000 10000

. nQ X0

-- -EL. -7.5m--- A- EL. -5.OmA,

+ EL. -2.7m

MEng 2002 - 2003

1

A Cnupi Qtiidj nn thr- P~.rfnrm-ani-r rnf~ml n1, nt ' Tt A Q 4- A AJI~ I.Ai JU Inn InlJ.L " 111%111 a %il %I 1 % Jw uu . Hr -

4 Ground Improvement Options

4.1 Vertical Wick Drains

The primary consolidation of a compressive soft clay foundation may take place over a

long time depending on the drainage path length. At the HW site, Hd = 5.55m and hence, 90%

consolidation takes eight years and eight months to reach the end of primary consolidation (refer

to previous calculations). The purpose of installing vertical drains is to accelerate the primary

consolidation. When the soil has characteristics of stratified soil, in which the permeability is

larger in the horizontal than in the vertical direction, the vertical drains are more effective. At

the HW site the PV drains cannot be installed into the underlying sand due to artesian pressures

that exist in this layer.

Figure 4-1: Embankment with wick drains

49

ME 2002 200


The reason for not fully penetrating the consolidating layer to reach the permeable layer

is that the sand layer is an artesian layer (the total head in the sand layer is higher than that in the

compressible layer). As a result, the PV drains were installed to the depth 1m above the sand

layer (approximately to El. -8.7m).

The vertical drains are often installed in a triangular pattern at a spacing S in the range of

1 to 5m. The circular drains have normally the radius r, in the range of 80 to 300mm and the

radius re of the soil cylinder discharging water into a vertical drain is in an order of magnitude of

0.525S. To use wick drains with rectangular cross-section, the equivalent rw has to be calculated

such as

r (a+b)IT,

where, a = thickness of the drain, typically values between 3.2 to 4.0mm

b = width of the drain, typically ranges from 93 to 100mm

/VerticalD D Drain

105 DS

Figure 4-2 Triangular pattern of installation of vertical drains (Terzaghi et al., 1996)

The discharge capacity of drains is limited in many cases due to some soil and drains

characteristics. The discharge capacity of the drains can be calculated as

q 7 1 k4

50

M.Eng 2002 - 2003

where qw = discharge capacity

kw= permeability of the drain

r= the equivalent radius

The discharge factor D was created to determine whether the water can flow freely or if

there is any well resistance. This factor is dependent on the horizontal permeability of the

consolidating soil and the maximum drainage length of the drain (Terzaghi et al., 1996).

The factor D can be computed as

D = "k,, ,,

where q, = discharge capacity of the drain

1= maximum drainage length

kh= horizontal permeability of the consolidating soil

In the case of soft clay deposits, analyses of field performance point out that when D is

greater than five, the well resistance is no longer an important factor and the minimum discharge

capacity of the drains can be calculated as

q,(min) = 5k,.,72

where kho =the initial horizontal permeability

Since the minimum discharge capacity is calculated from the initial horizontal

permeability, the magnitude of discharge capacity changes as the permeability decreases during

51

M.Eng 2002 - 2003

A 1Pn 2)002 - 2)O3ri. '...-a~. 31UUY U II LII I..~ I L~I ILJI *II~II'.A Uk LIIIUaIINIII~IIL~ UII I I~4L~U ~3UIL '~Ji'Juu*..z

consolidation due to the factor that less water penetrates the drains during a given time,

therefore: qw(min) gets smaller with decreases in permeability.

4.1.1 The effects of the smear zone in wick drains

A smear zone or cylinder of disturbed soil is created when the drains are installed (due to

displacement of the surrounding soil) which forms external radius rs around the drains. In the

smear zone, the compressibility increases, whereas the permeability and the pre-consolidation

pressure decrease. The equation below illustrates one-dimensional vertical compression together

with vertical and radial flow.

C u + j(2u +I au = Du, DZ 2 ar 2 r ar at

where ch = , u is the excess pore pressure

Since the excess pore pressure u is a function of time as well as of vertical and radial

flow, the same result can be extracted by superimposing the solutions for vertical compression

by vertical flow and one by radial flow. The degree of consolidation, at any time, can be

calculated as

U =1-(1- UzXl-Ur)

where, U, and Ur are the vertical and radial degree of consolidation, respectively

The excess pore water pressure can be defined, at any time, as a function of vertical and

radial flow.

U uurU.i

52

where u, and ur are the excess pore water pressure for vertical flow only and for radial flow only,

respectively.

The vertical flow is often ignored in situations where drainage occurs principally in the

horizontal direction. It is necessary to take into account the contribution of the vertical flow

when the height of the drains is small and the spacing is large.

rwr

thouhot h at, DepthB ( 4zad e e f s a

I I Flow -ine

compressible layer with fully penetrating vertical drains as

U= I exp

Fni

53

M.Eng 2002 - 2003

where U = degree of consolidation for radial flow only

3Fn = ln(n)--n , where n is greater than 10

4

n =

Chtr 2

In expressions for vertical compression with radial flow into vertical drains for the smear

zone, the factor of radius as r, = sr,, has to be taken into account.

When the effect of smear zone is taken into account, the permeability is different from

the permeability of the undisturbed soil, whereas the compressibility is the same as that of the

undisturbed soil for the relationship between U and Tr. Permeability can be calculated as

ks = kS2

Figure 4-4 illustrates the small effect of n in the relationship between U and T, whereas

for the rate of consolidation, n plays a great role due to its influence on r, = nr,. in the time factor

Tr.

54

M.Eng 2002 - 2003


00 . , . .

(b)20 - 20

q, =q,(min) n & r,/rv 25

s =4 K. s , r, /rw 4

80 I0

f ch t

Figure 4-4 Relations between degree of consolidation and time factor for radial flow into a vertical drain.

Loading is assumed to be instantaneous and excess pore water pressure constant with depth. (a) different

drain spacing; (b) different mobilized discharge capacity (Terzaghi et al., 1996)

4.2 Geotextile Coated Sand Columns

Ground improvement has been developed very rapidly to suit a wide range of ground

conditions and foundation problems. It is now a major area of geotechnical engineering and has

become increasingly popular as a cost effective alternative to piling in areas with highly

compressive soils. Treatment of compressible soils is designed to improve the load carrying

characteristics of the ground and to mitigate the total and differential settlement of the structures

subsequently constructed on the ground.

55

M.Eniz 2002 - 2003

A Cae Sudyon te Prfomanc ofEmbnkmets n Teate Sot GoundM Fo 9~~.

Figure 4-5 Geotextile coated sand columns

The Geotextile coated sand columns technique involves the in situ joining of large

diameter, closely spaced, vertical sand columns within the compressible layers. This technique

was used in the No-Recess project (HW4) in order to stabilize the compressive soft soil and

minimize the long-term settlements. The geotextile confines the sand and acts as a filter to

prevent intermixing with adjacent clay and clogging. It also provides the stiffening effect to

ensure integrity of the sand piles. These features allow the columns to act as piles, considerably

mitigating settlement and deformations due to dynamic loads. Another advantage of the sand

column technique is its relatively rapid construction time.

Many foundation benefits have been found with the use of sand columns in soft soils

such as increasing the bearing capacity for overlying structures or embankments, accelerating the

consolidation process of the consolidating layer surrounding the granular columns, and

improving the load-settlement characteristics of the foundation (Davies, 1997).

56

M Fna 2002 -'2003A Case Study on the Performance of Embankments on Treated Soft Ground

Inclusion of the columns reduces soil drainage path length, and increases the rate of

excess pore pressure dissipation. Installation of the sand column by displacement techniques

may minimize the soil permeability due to formation of the smear zones along the column

boundary, but since the column are closely spaced, the dissipation of the pore pressure can still

be achieved quickly (Bredenberg, 1999). One of the major factors to be considered when

improvement of the treated foundation is to be predicted is the stiffness of the sand columns

relative to that of the soil in which they are installed.

The installation method consists of vibrating a steel casing into the ground to the load

bearing layers. The casing has two flaps at the bottom, which are forced closed during driving,

displacing the ground. The geotextile is connected to a funnel, which is put in the top of the

casing. The tube is filled with sand through the funnel and the casing vibrated out, causing

initial compaction of the sand and stressing the geotextile (Nods, 2002).

Sand columns generally have a diameter in the range of 0.8m, and they are normally

placed in a triangular pattern with a spacing of 1.7m to 3.4m. In the case of the No-Recess

project (HW4), the columns had a diameter of 0.8m and were installed in a triangular grid in the

range of 2.4m to 3.4m for the high embankment. A geogrid is laid over the columns in a load

transfer platform in order to give extra horizontal stability and help to transfers horizontal

railway loads. To avoid large long-term settlement, surcharge is used to accelerate movements

(Nods, 2002).

57

. n,202-20MEng 2002 - 2003

A~q~- (~ ~itu~ nn th- Perfnrm-nnr f P , 1 , t rtA Q -4 us 'ill w sit CAUI 1 aIUIU vI.EAng 2002 -003

Figure 4-6 Process of installation of geotextile encased sand columns

58

5 FE Simulations of No-Recess Instrumented Embankments

This chapter compares results of finite element simulations using the Plaxis program, with

measured performance for two of the embankments constructed at the No-Recess (HW) site: (1)

the conventional design with PV wick drains (HW 1), and (2) the geotextile coated sand columns

(HW4).

5.1 Conventional Design, HW1

The engineering properties and the design parameters used to model the PLAXIS finite

element code analysis in these two cases are identical to the case of the unimproved embankment

(chapter 3). Figure 5-1 illustrates the imposed surcharge sequence and the centerline ground

surface predicted settlement for the embankment with wick drains. This figure also shows the

good accordance between the measured settlements and those predicted by Plaxis model.

I I III

I II ~FTTTTii100

s c l _I I I I I I I

I 111H

-0--U--U-

settleme t

1000

Surcharge(m)Settlement(m)-PredictedSettlement(m)-Measured

liiiIf10000

Time(days)

Figure 5-1: FE model of surcharge vs. time and comparison between measured and predicted ground

settlements for HW1.

59

8

6

4

2

0

-2

-4

z

CU

10

I I IIIIII

I I L...I

i i . . . , i i ! 1 . . . I

-

A Case Study on the Performance of Embankments on Treated Soft Ground M.Ene 2002 - 2003

I I I

The total vertical displacement calculated for the high embankment with vertical wick

drains by the FE model was 2.15m. The settlement predicted by Plaxis in figure 5-1 shows

agreement with that measured in the field. This large amount of settlement in the first 4 months

is due to the wick drains, which help to accelerate the dissipation of pore pressure, thus

accelerate the settlement rate.

The development of the horizontal displacement is shown in figure 5-2. The predicted

horizontal displacements by FE model are not exactly plotted on the same days that the measured

results are plotted, but one can see that the predicted results have a small deviation from those

measured in the field.

The largest horizontal displacement occurs around the EL.-4m, which also happens to be

the same behavior with that measured in the field. Figure 5-3 shows the correlations between

horizontal displacement and vertical settlements that can provide useful insight regarding the

relative importance of undrained versus drained deformations within embankment foundations

(Ladd, 1991). In the HWl, the pore pressure was measured at three elevations (-2.7m, -5.0m,

and -7.5m). Figure 5-4 illustrates the predicted excess pore pressure dissipation in those

elevations. In the case of vertical drains, the highest pore pressure dissipation is at EL.-7.5.

60

M.Eng 2002 - 2003A Case Study on the Performance of Embankments on Treated Soft Ground

0-

-2

S-6 L K-

8A- Day 90 (Predicted)-8- -- - 7 - - Day 180 (Predicted)

- Day 840 (Predicted)Day 98 (Measured)

-10 - ------- ---- -- - -U-Day 182 (Measured)-4-- Day 850 (Measured)

-12

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35Horizontal Displacement(m)

Figure 5-2: Comparison between horizontal displacement predicted and that measured in the field for HW1.

E2.-5 -

Ca)

E

E2.0UCL)(U

1 0.02 0.0 0.1 0.2 0.3 0.4 0.5

Maximum Settlement (m)

Figure 5-3: Relationship between maximum horizontal displacement and maximum settlement for HWJ.

61

.MEng 2002 - 2003

A~~~~~~~~~~~~~~~~~~ CaeSuyo h efrac fEbnmnso rae otGon Eng 2002 - 2003

80

N)E.

I-,-a)V~

a-)

a)0)

60

40

20

0

10 100 1000 10000

Time (Days)

Figure 5-4: Excess pore pressures dissipation at three elevations in HW1

5.2 Predictions for Geotextile Coated Sand Columns (HW4)

In the Plaxis simulation of the geotextile coated sand columns (GCS columns), the

properties used for the sand columns were those used for the fill. Since the last meter of the sand

columns is filled with soft material, the sand columns in the Plaxis model were set up Im above

the sand layer, leaving this last meter of original soil (peat). By using geotextile coated sand

columns a reduction of the settlement of 1.05m is achieved under the high section of the

embankment, relative to the embankment with wick drains predicted by FE model.

Figure 5-5 illustrates imposed surcharge sequence and the comparison between predicted

settlement and measured settlement in the field. The results accomplished by FE model are close

62

- EL -7.5m (Predicted)-s-- EL -5.0m (Predicted)

- EL -2.7m (Predicted)

-f


A Cae Stdy n th Peformnce f Ebankent on reatd Sft Goun gA 1n ?A

to those measured in the field. In the case of sand

embankment, only the weight of the embankment.

6

E

0

U,

4

2

0

-2

-4

columns, no surcharge was used for the high

surch rge

- 0 Surcharge(m)

I II I I I III I II I--Settlement(m)-PredictedSettlement(m)-Measured

- I

1 10 100 1000 10000

Time(days)

Figure 5-5: Imposed surcharge sequence and comparison of predicted settlement by plaxis and that measured

in the field.

Figure 5-6 illustrates the results from the predicted horizontal displacement and that

measured in the field. The results are compared on different days but they are close enough to

show a small deviation between them. The predicted horizontal displacement is bigger than that

measured in the field. This deviation can be due to the soil stiffness provided by the geotextile

coated sand columns.

63

A Case Study on the Performance of Embankments on Treated Soft Grud MA 1n 2002l - 20M3


0

-2

-C

-6 - ~

-- Day 50 (Predicted)-8 -- Day 80 (Predicted)

A Day 840 (Predicted)0 Day 57 (Measured)

-10 - - -+- Day 82 (Measured)-,- Day 850 (Measured)

12

0.00 0.05 0.10 0.15 0.20 0.25 0.30

Horizontal Displacement(m)

Figure 5-6: Horizontal displacement predicted by Plaxis

Figure 5-7 gives the relationship between the maximum horizontal displacement and the

maximum settlement in the high section of the embankment (HW4). Figure 5-8 shows the

predicted pore pressure dissipation. The largest pore pressure dissipation in the case of sand

columns also occurs at EL. -7.5, but it is not too large due to the rigidity of the sand columns.

64

M Fna)2002 - 2IW3

A Case Study on the Performance of Embankments on Treat d SotGon

0.25 -

0.20

0.15 -

0.10 -

0.05

0.00 -0.0 0.2 0.4 0.6 0.8 1.0

Maximum Settlement (m)

Figure 5-7: Relationship between the maximum horizontal displacement and maximum settlement in HW4

(Pre dicte d)(Pre dicte d)(Predicted)

I* m---.in Vi

10 100 1000 10000

Time (Days)

Figure 5-8: Predicted excess pore pressure in three locations beneath the high embankment (HW4)

65

-A-

-U-

E

0

0CO

0N

0a:EEa)

-- -- - - - - ------- -

-- - -

20

15

E)

U)

CD

L)

00)LnaO

10

5

0

A~~~~~~~~~ ~ ~ ~ ~ ~ ~ Cas S vothPefracofEbnmnsoTraeSotGud NA~~~~ V In n1MEn 2002 2003

I I

-5.0 m-7.5m-2.5 m

EL.EL.EL.

6 Conclusion

In the Netheilands the western part of the country consists of soft, very compressible soils --

peat and soft clay. In this area, the organization HSL-South is involved in the construction of a

new high-speed railway link between Amsterdam and Brussels. Short construction time and

stringent requirements on long-term settlements are important issues in the design of the

construction of embankments overlying soft soils. Since conventional railway embankments

would not be appropriate regarding these requirements, the No-Recess project was created to test

methods for stabilizing soft ground behavior. The five ground improvement trials demonstrate

successful applications of recently developed geotechnical techniques.

This thesis analyzes one unimproved embankment test and two of the five reinforced

embankment trials in order to provide a basis for evaluating the performance of the five ground

improvement solutions. Calculations and engineering judgment were employed to develop the

proper parameters for the analysis of the Hoeksche Waard site conditions. It was clear from the

one-dimensional calculations that the embankment would have lateral spread, therefore a

computer program was necessary to calculate the two-dimensional analysis.

The results of the settlement, horizontal displacement, and pore pressure dissipation were

compared between the three embankments calculated by Plaxis model and the two improved

embankment trials used in the HW site. The results from the Plaxis model fit well in most of the

measurements done in the field. The pore pressure dissipation in HW4 is not too high due to the

rigidity of the sand columns, which prevent the soft layer from compressing, thereby preventing

the water from escaping so easily.

The large settlement achieved with the conventional embankment (wick drains) is due to the

vertical drains, which accelerate the dissipation of the pore pressure. Therefore, the large

66

M.Eng 2002 - 2003


amount of settlement is achieved much faster than with the other ground improvement options.

In the case of the geotextile coated sand columns, the settlement achieved was much smaller than

that with wick drains. The geotextile casing limits the movement of the sand, making the sand

stiffer than the surrounding soil, which allows the transfer of the embankment load to the bearing

layer.

In conclusion, the results achieved may help to validate the methods used in the design

project. No-Recess ground improvement options are offered to develop better ground

improvement techniques in soft soils, where construction and residual settlement are important

issues.

67

.Fn 202-20MEn 2002 - 2003


References

[1] Andrawes, K. Z. et al. (1983). "The behavior of a geotextile reinforced sand loaded by a

strip footing." Improvement of Ground, vol.1, A.A.Balkema, 329-334.

[2] Balasubramaniam, A.S. et al. (1994). "Embankment on sand columns: Comparison between

model and field." Predictions versus Performance in Geotechnical Engineering,

A.A.Balkema, 167-173.

[3] Barends et al. (1999). "Geotechnical Engineering for Transportation Infrastructure", A.A.

Balkema.

[4] Batereau, C. et al. (1983). "Improvement of ground through the application of geotextiles."

Improvement of Ground, vol.2, A.A.Balkema, 459-462.

[5] Bredenberg, Hakan et al. (1999). "The performance of stabilized soil columns in two Dutch

testsites." Dry Mix Methodsfor Deep Soil Stabilization, A.A.Balkema, 239-244.

[6] Charles J. A. (1997). "Treatment of compressible foundation soils." Ground Improvement

Geosystems, Thomas Telford Publishing, 109-119.

[7] Craig W. H. and Z. A. Al-Khafaji (1997). "Reduction of soft clay settlement by compacted

sand columns." Ground Improvement Geosystems, Thomas Telford Publishing, 218-224.

[8] CUR (1996). Building on Soft Soils, A.A.Balkema.

[9] CUR (2001). Evaluatie No-Recess - testbanen Hoeksche Waard, Data CD.

[10] Edil, T.B. (1983). "Improvement of peat: a case history." Improvement of Ground, vol.2,

A.A.Balkema, 739-744.

[11] Holm, G. (1983). "Lime columns under embankments - a full scale test." Improvement of

Ground, vol.3, A.A.Balkema, 909-912.

68

M.Eng 2002 - 2003

[12] Ilander, A. et al. (1999). "Eurosoilstab - Kivviko test embankment - Construction and

research." Dry Mix Methods for Deep Soil Stabilization, A.A.Balkema, 347-354.

[13] Jamiolkowski, M. B. et al. (1983). "Behavior of oil tanks on soft cohesive ground improved

by vertical drains." Improvement of Ground, vol.2, A.A.Balkema, 627-632.

[14] Jansen, H. L. (1997). "Eurosoilstab Test Site's Gravendeel NL Subsoil Characterization and

Soil Parameters." Fugro Ingenieursbureau B.V.

[15] Ladd, Charles (1991). The Twenty-Second Terzaghi Lecture, American Society of Civil

Engineers.

[16] Lambe, T. William and Robert V. Whitman (1969). Soil Mechanics, John Wiley & Sons,

Inc.

[17] Masterbroek (1998). "Soil Investigation No-Recess test site Hoeksche Waard report." Fugro

Ingenieursbureau B.V.

[18] Nods, Max (2002). "Put a sock on it." Ground Engineering

[19] Ovesen, N. K. et al. (1983). "Centrifuge tests of embankments reinforced with geotextiles on

soft clay." Improvement of Ground, vol.1, A.A.Balkema, 393-398.

[20] Terzaghi, et al. (1996). Soil Mechanics and engineering Practice. John Wiley & Sons, Inc.

[21] Wu C. S. and C. Y. Shen (1997). "Improvement of soft soil by geosynthetic-encapsulated

granular column." Ground Improvement Geosystems, Thomas Telford Publishing, 211-217.

69

M.Enji 2002 -_2003

a case study on the performance of embankments on treated soft ground (2003) - thesis (69(

Documents