Characterization of Permanent Deformation

Of Subgrade Soils Using Cyclic Loading

Test
Nahla Y. Ahmad Ali A. Alwash
Babylon University - Civil Engineering
Anmar Faleh Degan
Engineering Researcher
Abstract
Permanent deformation of subgrade soils is the accumulation of plastic deformation in soils due to
continuous repeated loads from traffic loading. The main objective of this study is to investigate some of
the factors that affect the permanent deformation of subgrade soils such as soil types, compaction moisture
content, applied stress and dry density. Also, to formulate regression models which have been obtained to
predict the permanent strain in subgrade soils as related to moisture content , time of loading, dry density
and applied stress by using STATISTICA software.
This research also established a test procedure to use cyclic load test to measure the permanent
strain of subgrade soils. The permanent strains were monitored during 60 load cycles, and the accumulated
permanent deformations were determined. The test procedure and test results conducted on two types of
soils that cover an important geographical area of Ira, A! and A6 soils are presented.

!" #$ %
&
'! #$ !! ()* + (" , -%.! /0 1 20 (%
3 " 45 65 67
89: ;( < (= ; >0?@ A! (7'. B2 -() (C1D EFG>
H* (1I !5 JK < ;(= +L A1 6!
(7'; () >0?@. A! MN O9 P- "TATI"TI#A Q
! , 2 R S TU. V?@ W: %X MN Y !" #%'!
1Z - (S FX [\ #%.>
#%' . " E (1! , -(!! V?, %X] JK +1K9 +
89: +G2 1I^G (ZI* (1^? (!0 + _5 !. 89 PP A! . PP
A6 -
Intr!d"cti!n
The major function of subgrade soils is to provide support to pavement
structures. $nder heavy traffic loads, subgrade soils may deform and contribute to
distress in the overlying pavement structure. In asphalt pavement, this distress normally
ta%es the form of crac%ing and rutting. It has been well documented that the subgrade
soils plays a critical role in the initiation and propagation of permanent deformation of
pavement structures and directly influences the pavement performance #$"ang%
&''().The subgrade soils, in particular the wea% soft subgrades, contribute a significant
portion &above !0 '( of the total pavement rutting #*a+id,adeh% et al. &'-.). . All
pavement layers contribute to the permanent deformation of the pavement structure. )et,
subgrade does not attract as much attention as do the asphalt surface and granular base.
The last are man*made products in some sense and their properties are relatively well
%nown and have been e+tensively studied for several decades. Actually, with the
implementation of &super level I mi+ture design method(, the rutting of asphalt concrete
can be minimi,ed through the proper selection of materials. As a result, subgrade
becomes a wea% point in the pavement structure #/lli!tt% et al. &''.).
0esearch 1b+ecti2es
-. Investigate the main factors related to the permanent deformation of two Irai
subgrade soils including the compaction moisture content, applied stress, dry density
and some soil types.
.. /evelop models for the permanent deformation prediction of two %inds of local
subgrade soils. Also, Presenting a procedure for the prediction of the permanent
deformation of some subgrade soils.
Sam3ling
The subgrade soil samples that are used in this research were collected from two
sites that cover an important geographical area of Ira. The first subgrade soil sample was
ta%en from the north shoulder of 0erbala*1illa roadway near new 0erbala university,
while the second sample was ta%en from Al*jadrreya in 2aghdad near university of
2aghdad .
C4clic 5!ading Test
/ue to the limitation of the available testing euipments, the possible local
method which could be used to characteri,e the subgrade soils permanent deformation is
the unia+ial cyclic load compression test. The consolidation apparatus for soils
manufactured by &3eonor, 4slo * 5orway( is employed to perform this test. The tests
were performed on cylindrical specimens, 60.7 mm & . in( in diameter and -0-.6 mm & !
in( in height, giving a height to diameter ratio of &.( # 6"33ala% et. al.% &'''). Three
stresses are used for the cyclic load tests 8 .7 0Pa, !. 0Pa and 69 0Pa.
*echanism !7 6ermanent de7!rmati!n
:hen a shear stress is sustained on a soil specimen the stress will be carried in
the first place by the non*rigid clay matri+ which e+hibits translations and rotations of the
particles. As a result the load will be transferred to the more rigid contacts causing
significant decrease in creep rates. Important changes in the structure happen during this
stage, the number and si,e of voids are reduced, fabric defects partially closed, spacing
between clay si,e particles decreases. As a result the structure grows stronger leading to a
stable euilibrium. After this stage, the soil structure is behaving macroscopically as a
viscous material with a viscosity which increases with time due to a stiffening of the
contact bonds #8a9er% &':') as cited by #;emm!nr% &'.:).
At low and moderate stress levels, the deformational energy stored in the
specimen during the loading process is of a small order, and dissipated within a short
period of time. $nder high stress levels, the specimen e+hibits greater primary and
secondary creep. The stresses will be so high that microcrac%s of different orientation are
liable to develop, old fissures grow and new ones appear. As a result, the specimen
undergoes accelerating creep at a considerable rate leading to failure #6"sch < Felthan%
&'.=).
Fabricati!n !7 S3ecimens
&. *i>ing
The soils were mi+ed with the desired amount of water to control the moisture
content and dry density. This means that the compacted soil specimens were prepared at a
various moisture content to simulate the range of moisture e+pected in the subgrade
within a pavement.
?. C!m3acti!n
At first, the standard Proctor compaction tests were conducted on the two soil
types. These tests yielded dry unit weight versus moisture content relationships. The
results of the standard Proctor compaction test used to establish a various moisture
content and dry unit weight levels at which soil samples prepared and tested with the
cyclic load tests.
A special split of three pieces of steel mold with & 60.7 mm ( in diameter and
&-0-.6 mm( in height was designed.
The strain &;
p
( is calculated by applying the following euation8
@
3
A 3
d
B h
:here8
@
3
< a+ial strain &mm=mm(
Pd < a+ial deformation &mm(
h < specimen height &mm(
In order to investigate the effects of moisture content on the accumulation of
permanent deformation , specimens were fabricated at si+ different moisture contents
corresponding to >0' , -00' , -06' , --0' , --6' and -.0' of optimum moisture
content to simulate the field conditions of pavement subgrades.
"pecimens with the same moisture content and dry density were subjected to
various levels of a+ial stress &.7 0Pa, !. 0Pa and 69 0Pa( to e+plore the role of stresses
in the development of subgrade deformation. The effect of limited variation in density on
deformation behavior was also investigated where the specimens are compacted in
different degree of dry density to investigate the effects of dry density on a permanent
deformation.
Classi7icati!n Tests
The results of Atterberg limits tests conducted on fractions passing sieve 5o. !0
indicate that the 0erbala*1illa roadway soil samples are ?ow plastic soils while Al*
jadrreya soil samples are @edium plastic soils. the soil specimens can be classified
according to the unified soil classification system &u.s.c.s( as &clayey silt soil( for both
0erbala*1illa roadway soil sample and Al*jadrreya soil sample. :hile the AA"1T4
classification indicates that 0erbala*1illa roadway soil sample is &A!( soil and Al*
jadrreya soil sample is &A6( soil. The physical properties of the two soils are summari,ed
in Table &-(.
Table C&D S"mmar4 !7 S!il 6r!3erties
S!il ;erbalaE$illa AlE+adrre4a
Sam3ling ;erbalaE$illa 0!adwa4 AlE+adrre4a E Faghdad
5!cati!n
Near New Gni2ersit4 !7
;erbala
Near Gni2ersit4 !7
Faghdad
55 (=.- (H.H
65 ?(.: ?I.?H
6I -.& &&.?H
8ater C!ntent CJD ?: ?=
S3eci7ic Kra2it4 ?.:' ?.-(
13tim"m *!ist"re C!ntent
CJD
&..?H ?=
*a>im"m Dr4 Densit4
CgmBcm
(
D
&.-H &.-&
J !7 Fines ...I '(.:
AAS$T1 Classi7icati!n AI A:
Gni7ied Classi7icati!n Cla4e4 Silt S!il Cla4e4 Silt S!il
Test 0es"lts and Disc"ssi!n
Permanent deformation of subgrade soils is characteri,ed by a rapid increase in
deformation during the first cycle, followed by gradual stabili,ation. The amount of
deformation depends on the characteristics of the material and the applied load.
/77ect !7 *!ist"re C!ntent !n 6ermanent De7!rmati!n
the permanent strain is plotted against the time of loading for all degrees of the
moisture content under several levels of stress &.7 0Pa(, &!. 0Pa( and &69 0Pa(.
#yclic load testing results for 0erbala*1illa roadway soil &A! soil(, are presented
in Aigures &-( to &9(. Aigure &-( illustrates test results with moisture contents
corresponding to &>0', -00', -06', --0', --6', and -.0'( of optimum moisture
content and dry densities corresponding to >6' of ma+imum dry density with stress of .7
0Pa. In this Aigure it can be seen that when moisture content increases from >0' to
--0' of optimum, the permanent strain increases by .B0'.
:hile cyclic load testing results for Al*jadrreya soil &A6 soil(, are presented in
Aigure &!( to &6(. Aigure &!( illustrates test results with moisture contents corresponding to
&>0', -00', -06', --0', --6', and -.0'( of optimum moisture content and dry
densities corresponding to >6' of ma+imum dry density with stress of .7 0Pa. In this
Aigure it can be seen that when moisture content increases from >0' to --0' of
optimum, the permanent strain increases by ..0'.
0
0.005
0.01
0.015
0.02
0.025
0 500 1000 1500 2000 2500 3000 3500 4000
Time (second)
A
c
c
u
m
u
la
t
e
d

P
e
r
m
a
n
e
n
t

S
t
r
a
in

(
m
m
/
m
m
)
MC= 0! o" #MC
MC= 100! o" #MC
MC= 105! o" #MC
MC= 110! o" #MC
MC= 115! o" #MC
MC= 120! o" #MC
0
0.002
0.004
0.00$
0.00%
0.01
0.012
0.014
0.01$
0.01%
0.02
0 500 1000 1500 2000 2500 3000 3500 4000
Time (second)
A
c
c
u
m
u
la
t
e
d

P
e
r
m
a
n
e
n
t

S
t
r
a
in

(
m
m
/
m
m
)
MC= 0! o" #MC
MC= 100! o" #MC
MC= 105! #& #MC
MC= 110! o" #MC
MC= 115! o" #MC
Aigure &-( Cffect of @oisture #ontent on the
Delationship between ?oading Time and
Accumulated Permanent "train &A! soil, stress8 .7
0Pa(
Aigure &.( Cffect of @oisture #ontent on the
Delationship between ?oading Time and
Accumulated Permanent "train &A! soil, stress8 !.
0Pa(
0
0.002
0.004
0.00$
0.00%
0.01
0.012
0 500 1000 1500 2000 2500 3000 3500 4000
Time (second)
A
c
c
u
m
u
la
t
e
d

P
e
r
m
a
n
e
n
t

S
t
r
a
in

(
m
m
/
m
m
)
MC= 0! o" #MC
MC= 100! o" #MC
MC= 105! o" #MC
MC= 110! o" #MC
0
0.002
0.004
0.00$
0.00%
0.01
0.012
0.014
0.01$
0.01%
0.02
0 500 1000 1500 2000 2500 3000 3500 4000
Time (sec)
A
c
c
o
u
m
u
la
t
e
d

P
e
r
m
a
n
e
n
t

'
e
"
o
r
m
a
t
io
n

(
m
m
/
m
m
)
MC=0! o" #MC
MC=100! o" #MC
MC=105! o" #MC
MC=110! o" #MC
MC=115! o" #MC
MC=120! o" #MC
Aigure &9( Cffect of @oisture #ontent on the
Delationship between ?oading Time and
Accumulated Permanent "train &A! soil, stress8 69
0Pa(
Aigure &!( Cffect of @oisture #ontent on the
Delationship between ?oading Time and
Accumulated Permanent "train &A6 soil, stress8 .7
0Pa(
0
0.002
0.004
0.00$
0.00%
0.01
0.012
0.014
0.01$
0.01%
0.02
0 500 1000 1500 2000 2500 3000 3500 4000
Time(sec)
A
c
c
u
m
u
la
t
e
d

P
e
r
m
a
n
e
n
t

S
t
r
a
in

(
m
m
/
m
m
)
MC=0! o" #MC
MC=100! o" #MC
MC=105! o" #MC
MC=110! o" #MC
MC=115! o" #MC
0
0.002
0.004
0.00$
0.00%
0.01
0.012
0.014
0.01$
0 500 1000 1500 2000 2500 3000 3500 4000
Time (sec)
A
c
c
u
m
u
la
t
e
d

P
e
r
m
a
n
e
n
t

S
t
r
a
in

(
m
m
/
m
m
)
MC=0! o" #MC
MC=100! o" #MC
MC=105! o" #MC
MC=110! o" #MC
Aigure &6( Cffect of @oisture #ontent on the
Delationship between ?oading Time and
Accumulated Permanent "train &A6 soil, stress8 !.
0Pa(
Aigure &6( Cffect of @oisture #ontent on the
Delationship between ?oading Time and
Accumulated Permanent "train &A6 soil, stress8
690Pa(
/77ect !7 Stress !n 6ermanent De7!rmati!n
"pecimens with the same moisture content and dry density were subjected to
various levels of stresses to e+plore the role of stress in the development of subgrade
deformation.
Aigure &B( shows the accumulated permanent strain of A! soil at >0' of the
optimum moisture content at different stages of loading procedure for three levels of
applied stress &.7 0Pa(, &!. 0Pa( and &69 0Pa(. it can be seen that when the applied
stresses increased from .7 0Pa to 69 0Pa, the permanent strain increased &7-'( for soil
A!, while in Aigure &7(, when the applied stresses increase from .7 0Pa to 69 0Pa, the
permanent strain increased &-00'(, for soil A6.
it can be seen that at a lower applied stress there is a gradual stabili,ation, while at
a higher applied stress, a rapid increase in strain is observed for both soils used in this
research, indicating possibility of rutting accumulation over time.
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
0 500 1000 1500 2000 2500 3000 3500 4000
Time (second)
A
c
c
u
m
u
la
t
e
d

P
e
r
m
a
n
e
n
t
S
t
r
a
in
(
m
m
/m
m
)
ST= 2% (Pa
ST= 42 (Pa
ST= $3 (Pa
0
0.001
0.002
0.003
0.004
0.005
0.00$
0 500 1000 1500 2000 2500 3000 3500 4000
Time(sec)
A
c
c
u
m
u
la
te
d
P
e
r
m
a
n
e
n
t S
tr
a
in
(
m
m
/m
m
)
ST= 2% (Pa
ST= 42 (Pa
ST= $3 (Pa
Aigure &B( Cffect of "tress on the Delationship
between ?oading Time and Accumulated
Permanent "train &A! soil, @oisture #ontent
>0'(
Aigure &7( Cffect of "tress on the
Delationship between ?oading Time and
Accumulated Permanent "train &A6 soil,
@oisture #ontent >0'(
/77ect !7 Dr4 Densit4 !n 6ermanent De7!rmati!n
Aigure &>( demonstrates the effect of the dry density on the accumulation of the
permanent deformation for soil A!. "pecimens are fabricated to moisture content at
-00' of the optimum moisture content and the dry density increased from >9' to >6.6'
of ma+imum with applied stress of !. 0Pa, it can be seen that the permanent strain is
decreased by 9!'.
The effect of dry density on the accumulation of the permanent deformation for
soil A6 is presented in Aigure &-0(. "pecimens are fabricated to moisture content at -00'
of the optimum moisture content and the dry density increased from >!' to >6.!' of
ma+imum with applied stress of !. 0Pa, it can be seen that the permanent strain is
decreased by -B'.
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
0.0035
0 500 1000 1500 2000 2500 3000 3500 4000
Accumulated a)aermanent Strain (mm/mm)
Time (second)
''=3! o" M''
''=5! o" M''
''=$.$! o" M''
0
0.001
0.002
0.003
0.004
0.005
0.00$
0.00*
0 500 1000 1500 2000 2500 3000 3500 4000
Time (second)
A
c
c
u
m
u
la
te
d
P
e
r
m
a
n
e
n
t S
tr
a
in
(
m
m
/m
m
)
'.'.= $.4! o" M''
'.'.= 5! o" M''
'.'.= 4! o" M''
Aigure &>( Cffect of /ry /ensity on the
Delationship between ?oading Time and
Accumulated Permanent "train &A! soil, "tress8
!. 0Pa(
Aigure &-0( Cffect of /ry /ensity on the
Delationship between ?oading Time and
Accumulated Permanent "train &A6 soil,
"tress8 !. 0Pa
/77ect !7 S!il T43e !n 6ermanent De7!rmati!n
Arom Aigures &-(, &.(, &9( and Aigures &!(, &6(, &6(, it can be observed that soil A!
has more resistance to the permanent deformation than soil A6, specially for the moisture
content near to the optimum with all degree of stresses. 2ut when the moisture content
increased up to more than --0' of the optimum moisture content, the permanent
deformation of the two soils was appro+imately convergent.
2y a simple comparison between the permanent deformation of the two soils &A!
and A6( for the Aigures &-( and &!( with moisture content -00' and .7 0Pa applied
stress, it can be seen that soil A6 showed more permanent deformation than soil A!,
&about 66'(.
it can be observed that soil A! has more resistance to the permanent deformation
than soil A6, specially for the moisture content near to the optimum with all degree of
stresses. 2ut when the moisture content increased up to more than --0' of the optimum
moisture content, the permanent deformation of the two soils was appro+imately
convergent.
Keneral Trend !7 De7!rmati!n Feha2i!r
It can be seen that the permanent deformation under the first load application for
the two subgrade soils &A! and A6( constituted between &.B.6' to 60'( of the total
permanent deformation for all tests that was accumulated after 60 cycles. Table &.( and
&9( indicate that the accumulated permanent strain e+pressed as a percent of the total
permanent strain during the first, tenth, twenty five and fifty cycles for soil A! and for
soil A6 with the moisture content of -06' of the optimum moisture content and applied
stresses of .7 0Pa, !. 0Pa and 69 0Pa with dry density >6' of ma+imum dry density.
Table &.( Accumulation of Permanent "train at Earious ?oad #ycles C+pressed as a
Percentage of /eformation at 60 #ycles for A! "oil &@#<-06' of 4@#, // <>6' of
@//(
"tress &0Pa( -st -0th .6th 60th
.7 96.7 6-.6 BB.. >!
!. 9B.. 6!.7 7-.! >6.!
69 !B.6 6>.B 7..6 >6.6
Table &9( Accumulation of Permanent "train at Earious ?oad #ycles C+pressed as a
Percentage of /eformation at 60 #ycles for A6 "oil &@#<-06' of 4@#, //< >6' of
@//(
"tress &0Pa( -st -0th .6th 60th
.7 9- !7 66.B >-.6
!. 9B.7 60.9 B6.6 >6
69 !9.. 6!.7 B7.7 >!.!
Statistical *!del
2y using STATISTICA software, these variables will be obtained and
determined simultaneously under each moisture content using nonlinear regression
analysis on the obtained tests results.
Statistical Anal4sis
In a statistical modeling the overall objective is to develop a predictive euation
relating a criterion variable to predictor variables. The criterion variable is the permanent
strain while the predictor variables include the test conditions &stress level(, and some of
the soil properties &dry density, moisture content( and soil types.
*!del AdeL"ac4 Assessment
There are two approaches generally used to assess the adeuacy of the proposed
regression models, the first one is based on the e+amining goodness of fit measures,
where the second approach is based on the graphical analysis of the residuals, &also called
diagnostic plotsD #Alba4ati% ?==:).
6redicti!n *!dels 7!r 6ermanent Strain
The adopted models structure and the terms included in these models are shown
in Table &!(, while in Table &6(, &B( and &>( the statistical summary of these models are
presented. Also Tables &6(, &7( and &-0( shows the nonlinear analysis for the soils.
Table CID Keneral *!dels Str"ct"re 7!r the 6ermanent Strain
Criteri!n
Mariable
*!del Str"ct"re *!del N!.
@
3
C
!
#CT DD *C STD
C&
) &
@
3
C
!
N CT
C&
D N CST
C?
D N C*C
C(
D N CDD
CI
D ?
@
3
C
!
N C
&
T N C
?
ST N C
(
*C N C
I
DD (
@
3
C
!
#CT
C&
D CST
C?
D C*C
C(
D CDD
CI
D) I
:here8
@
3
< accumulated permanent strain &mm=mm(
T< time of loading &second(
//< dry density &gm=cm
9
( as a percentage of ma+imum dry density
"T< applied stress &0Pa(
@#< moisture content &'( as a percentage of optimum moisture content
Table CHD Statistical S"mmar4 7!r 6ermanent Strain *!dels 7!r S!il AI
5!ss F"ncti!n Final 5!ss 0 Mariance
/>3lained C0
?
D
*!del
N!.
C1FSE60/DDOO? =.=?I''='=( =.&::&= ?.-H.'J &
C1FSE60/DDOO? =.=='I(-==I =.-'HI' :(.?.=J ?
C1FSE60/DDOO? =.==.:&:&=' =..&H(? ::.I-IJ (
C1FSE60/DDOO? =.==&:??I-I =.':-'? '(.:.-J I
Table C:D N!nlinear Anal4sis 7!r S!il AI
*!delP @
3
A C
!
OCCTOOC
&
DOCSTOOC
?
DOC*COOC
(
DOCDDOOC
I
DD
Final l!ss A =. ==&:??I-I
Sta. N!nlin.
/stimati!n
C
I
C
(
C
?
C
&
C
=
N A &=?=
E&H.H=?I.- &I..H:&.H =..&-&:H =.&IH.-. =.===?H= /stimate
:here8
5 < numbers of valid cases
According to Tables &6( and &6(, the final form of proposed model for soil A! is8
@
3
A ?.HO&=
EI
&T
=.&IH.-.
( &ST
=..&-&:HD
C*C
&I..H:&.HD
CDD
E&H.H=?I.-
( FF..&-*-(
In this model the numbers of valid cases &5( are &-0.0( which represent a large
number of data in order to give the model high accuracy.
it can be seen from Aigure &--( that the residuals in the first plot are normally
distributed, and the second plot yields points close to the &!6(G line, then the proposed
regression function gives an accurate prediction of the values that are actually observed.
&re+uenc, 'istri-ution. /esiduals
01)ected 2ormal
30.0030
30.0025
30.0020
30.0015
30.0010
30.0005
0.0000
0.0005
0.0010
0.0015
0.0020
0.0025
0.0030
0.0035
0
20
40
$0
%0
100
120
140
1$0
1%0
200
220
240
2
o

o
"

o
-
s
#-ser4ed 4ersus Predicted 5alues
30.002
0.000
0.002
0.004
0.00$
0.00%
0.010
0.012
0.014
0.01$
0.01%
0.020
0.022
0.024
Predicted 5alues
30.002
0.000
0.002
0.004
0.00$
0.00%
0.010
0.012
0.014
0.01$
0.01%
0.020
0.022
#
-
s
e
r
4
e
d

5
a
l
u
e
s
Aigure &--( /iagnostic Plot for Permanent "train &@odel 5o.!( "oil A!
Table C-D Statistical S"mmar4 7!r 6ermanent Strain *!dels 7!r S!il A:
5!ss F"ncti!n Final 5!ss 0 Mariance
/>3lained C0
?
D
*!del
N!.
C1FSE60/DDOO? =.=&H- =.(:? &(.&&.J &
C1FSE60/DDOO? =.==I( =..-? -:.&?IJ ?
C1FSE60/DDOO? =.==(. =...: -..::(J (
C1FSE60/DDOO? =.==&=' =.':' '(.'.:J I
Table C.D N!nlinear Anal4sis 7!r S!il A:
*!delP @
3
A C!OCCTOOC&DOCSTOOC?DOC*COOC(DOCDDOOCIDD
Final l!ssP =.==&='?.?=
Sta. N!nlin.
/stimati!n
C
I
C
(
C
?
C
&
C
=
N A &=?=
E'.&I((- ..:&H-(' =.:-:.H: =.?&H-= =.===I-H /stimate
According to tables &B( and &7(, the final form of proposed model for A6 soil is8
@
3
A I.-HO&=
EI
&T
=.?&H-=
( &ST
=.:-:.H:
D C*C
..:&H-('
D CDD
E'.&I((-
( F..F&-*.(
As in euation &-*-(, In this model the numbers of valid cases &5( are &-0.0(
which represent a large number of data in order to give the model high accuracy. Aigure
&-.( provides the comparisons between predicted versus measured plastic strain, shows
that good correlations are obtained, which indicate that the present formulation has
adeuately represented the plastic strain response of soil tested under the cyclic loading
test.
&re+uenc, 'istri-ution. /esiduals
01)ected 2ormal
30.002 0.000 0.002
/esidual
0
20
40
$0
%0
100
120
140
1$0
1%0
200
220
240
2
o

o
"

o
-
s
#-ser4ed 4ersus Predicted 5alues
30.002 0.000 0.002 0.004 0.00$ 0.00% 0.010 0.012 0.014 0.01$ 0.01% 0.020
Predicted 5alues
0.000
0.002
0.004
0.00$
0.00%
0.010
0.012
0.014
0.01$
0.01%
0.020
#
-
s
e
r
4
e
d

5
a
l
u
e
s
Fig"re C&?D Diagn!stic 6l!t 7!r 6ermanent Strain C*!del N!.ID S!il A:
Cuation &!*9( represents a model for both soils &A! and A6( where the results of
both soils were collected and dealing in the statistica software with &.0!0( number of
valid cases which represent a large number of data in order to represent the two soils in
one model. Also it can be seen from Aigure &-9( that the residuals in the first plot are
normally distributed, and the second plot yields points close to the &!6(G line, then the
proposed regression function gives an accurate prediction of the values that are actually
observed.
Table C'D Statistical S"mmar4 7!r 6ermanent Strain *!dels 7!r S!il A: and AI
5!ss F"ncti!n Final 5!ss 0
Mariance
/>3lained C0
?
D
*!del
N!.
C1FSE60/DDOO? =.=I&'-H.I& =.?H&I= :.(?==J &
C1FSE60/DDOO? =.=&H?.?.(I =..&&-I :H..'?J ?
C1FSE60/DDOO? =.=&I=&H?(- =..?.'. :..-?&J (
C1FSE60/DDOO? =.==H=?=(?: =.'I?(? ...-':J I
Table C&=D N!nlinear Anal4sis 7!r S!il A: and AI
*!delP @
3
A C!OCCTOOC&DOCSTOOC?DOC*COOC(DOCDDOOCIDD
Final l!ssP =.==H=?=(?:
Sta. N!nlin.
/stimati!n
C
I
C
(
C
?
C
&
C
=
N A ?=I=
E...(HI: &&.&''I& =.-=&.H. =.&.&H'H -O&=
E&=
/stimate
According to tables &>( and &-0(, the final form of proposed model for both A!
and A6 soils is8
@
3
A -O&=
E&=
&T
=.&.&H'H
( &ST
=.-=&.H.
D C*C
&&.&''I&
D CDD
EE...(HI:
( FF.F..&-*9(
&re+uenc, 'istri-ution. /esiduals
01)ected 2ormal
30.005 30.004 30.003 30.002 30.001 0.000 0.001 0.002 0.003 0.004 0.005
0
50
100
150
200
250
300
350
400
450
500
550
$00
2
o

o
"

o
-
s
#-ser4ed 4ersus Predicted 5alues
30.002
0.000
0.002
0.004
0.00$
0.00%
0.010
0.012
0.014
0.01$
0.01%
0.020
0.022
Predicted 5alues
30.002
0.000
0.002
0.004
0.00$
0.00%
0.010
0.012
0.014
0.01$
0.01%
0.020
0.022
#
-
s
e
r
4
e
d

5
a
l
u
e
s
Fig"re C&(D Diagn!stic 6l!t 7!r 6ermanent Strain C*!del N!.ID S!il A: and AI
C!ncl"si!ns
:ithin the limitations of materials and testing program adopted in this wor%, the
followings are concluded8
&D The moisture content has a significant effect on the permanent strain for A! and A6
soils. The increase of the moisture content from >0' to --0', has increased the
permanent strain by .B0' for A! soil and by ..0' for A6 soil. Therefore, with the
increase of the moisture content a considerable increase in the permanent strain is
highly e+pected.
?D The applied stress showed a considerable influence on the permanent strain values,
specially on the first cycle of loading, where the increase of stresses from .7 0Pa to
69 0Pa, has increased the permanent strain by &7-'( for A! soil and by &-00'( for
A6 soil. Therefore, 1igh permanent strain was recorded when soils were subjected
to high stresses.
(D The permanent strain results for soils A! and A6 compacted at wet of optimum
moisture content condition was higher than those compacted at dry of optimum
moisture content.
ID The permanent strain results showed that A! soil at the optimum moisture content
e+hibited lower plastic deformation &about -06'( compared with A6 soil for all
degrees of stresses.
HD At the optimum moisture content, the permanent strain has been increased by .6' for
soil A! and by 6>.B' for soil A6 for an increase in the applied stresses from .7
0Pa to 69 0Pa.
:D The effect of limited variation in dry density on deformation behavior produces a
slight increase in the permanent strain for both soils.
0e7erences
Albayati, A. 1. 0., &.006(. Q6ermanent De7!rmati!n 6redicti!n !7 As3halt C!ncrete
Gnder 0e3eated 5!adingQ, Ph. /. Thesis, $niversity of 2aghdad.
Clliott, D. P. /ennis, 5. /. and Hiu, )., &Ianuary ->>7(. Q6ermanent De7!rmati!n !7
S"bgrade S!ils C6hase IP A Test 6r!t!c!lDQ, @2T# AD -06>, /epartment of
#ivil Cngineering, $niversity of Ar%ansas.
1uang, ). 1. &->>9(. Q6a2ement Anal4sis and DesignQ, -
st
Cdition, Prentice 1all,
Cnglewood #liffs, 5ew Iersey, $"A.
0emmonr, A.A., &Ianuary, ->76(. QSt"d4 !7 Cree3 and /77ect !7 Strain 0ate !n the
Strength and De7!rmati!n !7 C!m3acted S!ilECement Sam3lesQ, @. "c.
Thesis, $niversity of 2aghdad.
@ajid,adeh, 0., 2ayomy, A., and 0hedr, ". &->B7(. Q0"tting /2al"ati!n !7 S"bgrade
S!ils in 1hi!Q, Transp. Des. Dec. 5o. 6B-, Transportation Desearch 2oard
5ational Desearch #ouncil, :ashington, /.#.
Puppala A. I., @ohammad, ?. 5, and Allen A., &->>>(. Q6ermanent De7!rmati!n
Characteri,ati!n !7 S"bgrade S!ils 7r!m 05T TestQ, Iournal of @aterials in
#ivil CngineeringJ, Eol. --, 5o. !, pp. .B!*.7..
Pusch, D. and Aeltham, P. &->70(.QA st!chastic m!del !7 the Cree3 !7 S!ilsQ,
3eotechniue, Eol. 90, 5o.!, pp. !>B*606.