You are on page 1of 6

IOSRJournalofMechanicalandCivilEngineering(IOSRJMCE)

eISSN:22781684,pISSN:2320334X,Volume12,Issue4Ver.III(Jul.Aug.2015),PP3741
www.iosrjournals.org

LinearRegressionEquationforSubgradeLateriticSoil

J.A.Ige
DepartmentofCivilEngineering,LadokeAkintolauniversityofTechnologyOgbomosoNigeria

Abstract:
The results of a study that considered the use oflinear regression equationsto have a correlation
between index properties and California bearing ratio of some lateritic soil within Osogbo. South Western
Nigeria have been presented. For an appreciable conclusion to be established, lateritic soil samples were
collected from eight (8) differentborrow pits within the townandvariouslaboratorytests includingAtterberg
limit, Sieve analysis, California bearing ratio, Compaction and Specific gravity were performed on the soil
samples. Various linear relationships between index properties and California bearing ratio (CBR) of the
samples were investigated and predictive equations estimating (CBR) from the experimentalindexvalueswere
estimated. The findings indicate that goodcorrelationexists between the two groups. Consequently,thevalues
of theCBR computedfrom the models areonly tobe usedfor preliminaryassessmentinviewoftheoretical and
economy and not acceptable alternatives to laboratory testingbecauseofthe anisotropicnatureoflateriticsoil
anditsheterogeneity.
Keywords:
Borrow pits, Carlifornia bearing ratio, Lateritic soil, Linear regression equation, Models,
Predictiveequation

I.

Introduction

Index properties have wide applications in geotechnical engineering practice and a number of index
properties are easilyrecognized in soil mechanics have been outlined by many authors (Lambeand Whitman,
1979 Holtz and Kovacs, 1981 Das, 1998). It is normal practice totry to predict the engineering behaviourof
soils from their index properties. The properties form the basis for soilclassification intogroups where each
group is said to have similar behaviour.Some authors (Lambeand Whitman, 1979 Holtz and Kovacs, 1981)
havegivenreasonsforthe useofindexpropertiesinpredictingsoilbehaviour.Accordingly, manyresearch have
made attempt at predicting some engineering properties of soil suchascompaction parameters (Korfiatis and
Manikopoulos, 1982 Alkhafaji, 1987 Howell et al., 1997 Boltz et al., 1998), permeability (Benson et a!..
1994 Benson and Trast, 1995 Boadu, 2000 Osinubi and Amadi, 2006), swelling pressure (Komornik and
David, 1969), consolidation parameters such as compression index (Azzouz et al., 1976 Alkhajaji and
Andersland, 1992), California bearingratio(AdegbolaandAlajede,2003)usingindexproperties.However,soil
composition affects index propertiesofsoils.Mostofthesepredictionsoftenutilizeatterberglimit values which
areaffectedessentiallybysomesoilcompositionalaswellasenvironmentalfactors,
The strength of subgrade is the main factor in determining the thickness of the pavement. Subgrade
strength is expressedintermsofitsCaliforniaBearingRatio(CBR)value.TheCBRtestisgenerallycarriedout
in theLaboratory on remoulded samples of the subgrade,asdescribedinBS1377of 1990. Thesamplemustbe
compacted at the equilibrium moisture content to the dry density likely to apply after the road has been
constructed. It is to be noted that CBR test is widelyaccepted internationally asareliablemethodofpavement
design and is alsousedinsoilclassification ofbase andsubbase(road base)materials forhighwaydesignsand
construction,thusitbecomesnecessarytoaddtothe existingknowledgeby usinglinearregression topredictthe
rangeofCBR values for usewhere constraints as to the levelof expertise and equipment arise (Adegbolaand
Alajede,2003).
Laterites (or lateritic soils) as a soil group rather than welldesigned materials are most commonly
found in the leached soils of the humid tropicswhere they were first studied. According to wooltorton (1975)
these soils are formed under weathering systems productive of theprocessof laterization, the most important
characteristic of whichis the decomposition of ferroalumino silicate mineralsandthepermanentdepositionof
sesquioxides (i.e. oxides of iron and aluminium Fe
O
O
) within the profile to form thehorizon of
2
3 and Al
2
3
materialknown as laterite.Gidigasu andKuma(1987)usedthe term"laterization"todescribe theprocessesthat
produce lateriticsoils. Lateritic soils are being used inthe construction of roads, highways, airfields, andearth
damsandasthefoundationofstructures(Gidigasu,1976).
It should be noted that researchers have done a lot in using statistical parameters such as regression
analysis in solving geotechnical problems. Adegbola and Alajede (2003) established a relationship between
DOI:10.9790/168412433741www.iosrjournals.org1|Page

LinearRegressionEquationforSubgradeLateriticSoil
CBRand index properties and concluded that astrong correlation exist betweenthetwotestsand suggeststhat
the method of a approach is sufficient to predict the range of CBR values expected through mathematical
analysis without actually performing CBR laboratory tests on soil samples within the locality. Adedimila and
Usifo(1994) carried out a comparative study of CBRand Unconfined CompressionTests(UCT)intheareaof
characterizingcohesivesoilsforpavementdesignandanalysisinviewoftheoreticaldevelopmentandeconomy,
Thus, Osun State government has embarked upon mass construction of roads withinthestatecapital,
Osogbo to ease the means of transportation within the stale capital and the surrounding towns.Most of these
roads consisted mainly of transported materials.As a result of this, there is a need toreplace them with good
paving materials. Hence, eight borrow pits that have large quantities of good constructional materials for the
pavement constructionwere identifiedafter geotechnical investigation with respectto quality and quantityhad
been carried out. As a result of this, the studynowconsidered theuse of linear regressionequation to have a
correlationbetween index properties and California bearing ratioof lateritic soilwithinOsogbo,Southwestern
Nigeria.

II.

MaterialsandMethod

The study commenced with deskwork and reconnaissance survey of the project site. Then, sampling
wascarriedoutthroughtrialpittingwhichpermittedacloseexaminationofthesites.

Preparation of specimens:
Samples that have beencollectedviatrialpittingwerepreparedinaccordancewith
BS 1377, 1990 and Head,1992.Prior topreparingthe test specimens,thematerialswereairdriedand broken
intosmallerfragments,carebeingtakennottoreducethesizesoftheindividualparticles.

Test Procedures:
The following tests viz:Atterberg limits,Particlegrainanalysis,Californiabearingratio,and
Compactiontestwerecarriedoutoneachofthedisturbedsamples.Theproceduresoftheirtestsareasfollows:

Grain size analysis:


Representative sample of approximately 500kg was used for the test afterwashing and
ovendried.Thesievingwasdonebymechanicalmethodusinganautomaticshakersandasetofsieves.

Liquid limit determination:


Soil sample passing through 425um sieve, weighing 200g wasmixed withwater
to form a thick homogeneous paste. Thepastewascollected insidetheCasangrade'sapparatuscupwithagrove
createdandthenumberofblowstocloseitwasrecorded.

Plasticlimitdetermination
:soilsampleweighing200gwastakenfromthe
material passing the 425um test sieve and then mixed with watertill it become homogenous andplastic to be
shaped to ball. The ball of soil was rolled on u glass plate until the thread cracked at approximately 3mm
diameter.

California BearingRatio (CBR


): Fresh sets of 3kgairdriedsoilweremixed withsuitableamountofwaterof
about5% of Its weight of water. The sample was completedfollowingthestandardprocedure(BS1377,1990).
Thesample was put inCBRmouldin3layerswitheachlayercompactedwith55blows using2.5kghammerat
a drop of50mm (standardproctor test).The compactedsoil andthemouldwereweighedandplacedunder CBR
machine following the standard procedure. Load was recorded at penetration of 0.625, 1.9,2.25, 6.25, 7.5,10
and12.5mm.

Compactiontest
: Samples that were crushed to pass through 4.76mm(BS No.4) sieve apertureasoutlinedby
head (1992)of about 3kg was used. Thesample was mixed with suitable amount ofwater of 5% atthe initial
stage and later increased to 7%, 9%, 10 % and 13% on subsequent tests. The soil was compacted using BS
mould of 105m diameterand115.5mmheight.Thecompactionwasdonein5layers.Eachlayerwascompacted
with4.5kgrammer at 10 blows fromadroppingheight300mm.ThismethodisknownasWest AfricaStandard
(WAS). Maximum Dry Density (MDD) and Optimum Moisture Content (OMC) were determined from the
graphofdrydensityagainstmoisturecontent.(Bowles,1988andBS1377,1990)

III.

ResultsandDiscussion

Sieve Analysis:
The percentage of the lateritic soil samples passing BS Sieves 2mm, 425m and 5m are
shown in Table l. The percentage passing through No. 200(75m) BS sieve ranges between 10.5% and24.7%
showing that the soilsamplesarccoarsematerialsaccordingtoUnifiedSoilClassificationSystem(USCS).The
soil samples can be basically suitable forsubgrade construction as theirpercentage byweight finer than No.
DOI:10.9790/168412433741www.iosrjournals.org2|Page

LinearRegressionEquationforSubgradeLateriticSoil
200 BS sieveis less than 35%, according to Federal Ministry of Worksand Housing(1972)specification.The
lateriticsoil samples also belongto A26 subgrade according to AASHTO method ofclassificationasshown
onTable1.

Atterberg Limit:
The liquid limits of the soil sarnples range between 28.6% and 45% and plastic limits
between 18.8% and 25% while plasticity index is between 14.2% and 20.7% as shown in table 1. Soils with
liquid limit less than 30%are consideredto be oflowplasticity, those withliquidlimitbetween30%and50%
exhibit mediumplasticity and those withliquidlimitgreaterthan50%exhibitmediumplasticityandthosewith
liquid limit greater than 50% exhibit high plasticity. All samples exhibit medium plasticity except sampleT5
which exhibit low plasticity. This type of liquidlimit can be expected forsiltysoilswhichusuallyhavetypical
values of 25% to 50% (BS 1377. 1990). Also, a liquid limit of 50 (maximum) and a plasticity index of 15
(maximum)havebeenrecommended,thussamplesT2,T3andT5havefallenoutoftherecommendation.

Specific Gravity:
The values of the specific gravityfor thesamplesrangedbetween2.66and2.77asshownin
Table 1. The specific gravity values are within the range recommended in BS 1377 of (1990). Thus, lower
specificgravityvalueindicatesacoarsesoil,whilehighervaluesindicateafinegrainedsoil.
Compaction:
British standard light compactive effort was used. The Maximum Dry Density (MDD) ranges
between1.98Mg/m3
and 2.21Mg/m3
whiletheoptimummoisturecontentrangesbetween 10.81%and12.52%as
depictedinTable2,

California bearing ratio:


The values of CBR have been shown in Table 3. lt has unsoakcd CBR ranges
between 64 and 85% which that of its corresponding soaked samples range between 26 and 33%. The
percentagedecreasesfromunsoakedCBRtosoakedCBR.Thisimpliesthataswateris absorbedintocompacted
specimen, the resistance to penetration becomes drastically reduced. It has been recommended by Federal
Ministry of Works and Housing (1972) that the valuesofCBR forroadbase,subbaseandsubgradeshouldnot
be less than 80%, 30%and 10% respectively under soaked condition. It can beseenthatsamplesTlandT5 do
notsatisfytheconditionforbothroadbaseandsubbase.

Table1:
SummaryofGrainsizeanalysis,AtterberglimitandSpecificgravity
SampleNo
Tl
T2
T3
T4
T5
T6
T7
T8
SampleNo
Tl
T2
T3
T4
T5
T6
T7
T8

Grading %passingBSsieve2mm
425^m75|im
68.951.721.5
83.165.624.7
88.250.018.6
69.433.413.1
54.832,210.5
84.853.823.7
77.642.420.3
80.452.721.8

AtterberglimitLL(%)PL(%)PI(%) Specific
gravity
33,018.814.2
2.68
36.719.117.6
2.70
38.020.617.4
2.66
45.024.320.7
2.74
28.617.411.2
2.77
36.018.917.1
2.67
36.018.016.0
2.66
40.025.015.0
2.71

Table2:
SummaryofCaliforniabearingvalueandCompaction
CBR(Unsoaked%)CBR(Soaked%)'
66.7
29
79.3
33
78
30
85
32
64'
26
73
31
77
34
84
33

3
MDD(Mg/m
)
2,13
2.09
2.18
2.20
1.98
2.16
2.04
2.21

AASHTO
Classification
A26
A26
A26.
A26
A26
A26
A26
A26

OMC(%)
11.58
11.98
U.15
10.90
12.52
11.23
12.34
10.81

Computational analysis of linear regression equations:


Table 3 presents the summaryofthecomputationof
linear regression equations for different categories of tests results while table 4 presents the summary of the
experimental and calculated values oftheCaliforniabearingratio,compactionandotherindexvalues.Attempts
were made to establish correlations between California bearing ratio values and index properties of the soil
samplesusinglinearregressionanalysis(Loveday,1989),

Table3:
Linearregressionequationfordifferentcategoriesoftest
DescriptionLinearregression
DOI:10.9790/168412433741www.iosrjournals.org3|Page

LinearRegressionEquationforSubgradeLateriticSoil
CBR(unsoaked)vs.LiquidlimitCBR
CBR(unsoaked)vs.Plasticlimit
CBR(unsoaked)vs.Specificgravity
CBR(unsoaked)vs.MDD
CBR(soaked)vs.Liquidlimit
CBR(soaked)vs.Plasticlimit
CBR(soaked)vs.Specificgravity
CBR(soaked)vs.MDD

CBR=0.031LL+83.19
CBR=0.8P.L+65.31
CBR=10.43S.G+56.19
CBR=8.66MDD+65.88
CBR.=0.22L.L+28.87
CBR=1.04P.L+13.56
CBR=9.42S.G+10.91
CBR=50.28MDD70.22

DOI:10.9790/168412433741www.iosrjournals.org4|Page

LinearRegressionEquationforSubgradeLateriticSoil
Table4:
Summarytableforexperimentalandcalculatedvalues

SampleNo.
Liquidlimit
ExperimentalvalueofCBR(unsoaked)
CalculatedvalueofCBR
Plasticlimit
ExperimentalvalueofCBR(unsoaked)
CalculatedvalueofCBR
Specificgravity
ExperimentalvalueofCBR(unsoaked)
CalculatedvalueofCBR
Maximumdrydensity
ExperimentalvalueofCBR(unsoaked)
CalculatedvalueofCBR
Liquidlimit
ExperimentalvalueofCBR(soaked)
CalculatedvalueofCBR
Plasticlimit
ExperimentalvalueofCBR(soaked)
CalculatedvalueofCBR
Specificgravity
ExperimentalvalueofCBR(soaked)"
CalculatedvalueofCBR
Maximumdrydensity
ExperimentalvalueofCBR(soaked)
CalculatedvalueofCBR

Tl
33
66.7
84.19
18.8
66.7
80.4
2.68
66.7
84.14
2.13
66.7
84.32
33
29
36.13
18.8
29
33.1
2.68
29
36.19
2.13
29
36.88

T2
36.7
79.3
84.3
19.1
79.3
80.6
2.7
79.3
84.4
2.09
79.3
83.98
36.7
33
36.9
19.1
33
33.4
2.7
33
36.34
2.09
33
33.8

T3
38
78
84.37
20.6
78
81.8
2.66
78
83.3
2.18
78
84.76
38
30
37.2
20.6
30
34.98
2.66
30
35.97
2.18
30
39.39

T4
45
85
84.6
24.3
85
84.8
2.74
85
84.77
2.20
85
84.93
45
32
38.8
24.3
32
38.8
2.74
32
36.72
2.20
32
40.4

T5
28.6
64
84.1
17.4
64
79.23
2.77
64
85.08
1.98
64
83.03
28.6
26
35.16
17.4
26
31.66
2.77
26
37.0
1.98
26
29.33

T6
36
73
84.3
18.9
73
80.4
2.67
73
84.04
2,16
73
84.59
36
31
36.8
18.9
31
33.2
2.67
31
36.06
2.16
31
38.38

T7
36
77
84.3
18.0
77
79.7
2.66
77
83.93
2.04
77
83.55
36
34
36.8
18.0
34
32.28
2.66
34
35.97
2.04
34
32.5

T8
40
84
84.4
25
84
85.3
2.71
84
84.46
2.21
84
8^.02
40
33
37.67
25
33
39.56
2.71
33
36.44
2.21
33
40.9

IV.

Conclusion

TheconstructionofhighwaypavementrequirestomeetregulatoryminimumCBRwhichshouldnotbe
less than 80%, 30% and 10% for base, subbase and subgrade materials thus requires careful selection of
materials. The suh presented an empirical approach for calculating or predictingCaliforniabearingratiowhich
servesasthemostimportantfactorduringroadpavementconstruction.
The results of the experimental analysis show that California bearing ratio for the unsoaked samples
ranges between 66.7% and 85% while that of the calculated value ranges between 79.2% and 85.3%. In the
same vein, the results of the experimentalvalue for the soaked California bearing ratio shows that the values
range between 26% and 34% while that of the calculated value ranged between29.3% and 40.9%. Thus, the
results of the analysisindicatethatthereisarelationshipbetweenCaliforniabearingratiocompactionandindex
properties.The results of the experimentalvaluesshowthatsampleTlandT5donotsatisfytherequirementfor
both road base and base materials can only beused for subgrade material. In the other way, when looking at
angle of the calculated value there isslight contradiction because the values satisfy the conditionsfor the base
andsubbasematerials.Also, forthesoakedCaliforniabearingratio,samplesTlandT5havevaluesthatareless
than30%fromtheexperimentalpointofviewbutgreaterthan30%fromthecalculatedvalue.
It can thus be safely concluded that linear regression analysis provides a sound background for
predicting California bearing ratio of samples for preliminary assessment andnotan acceptable alternative to
laboratory testing. It is recommended for future research that regression based motels such as two ways
ANOVA and computer based reliability analysis be carried out on a wider variety of soil samples so as to
specifytherangeofapplicabilityofthederivedmodelvizaviztheinputvariable.

References
[1].
[2].
[3].

[4].
[5].

th
AASHTO(1986).Standardspecificationfortransportationmaterialsandmethodsofsamplingand
testing,14
Ed.,Washington,
D.C.
Adedimila, A.Sand Usifo, J.E.F (1994). 'A comparativestudy ofCalifornia BearingRatioandUnconfined compressiveTestsfor
StabilizedLalerile.NSI:TechnicalTransactions.Vol.29pp.
6174.
Adegbola,A.A.(2003).EstablishingarelationshipbetweenCBRandIndexpropertiesofroadbase materials. Science Focus
Vol.4pp.610.AlKhafaji,A.N.(1987).Asimpleapproachtothe estimationofsoilscompaction parameters.Q.J.Engrg.Geol.,
20:1530.
AlKhafaji, A.N. and Andersland, O.(1992). Equations for compression indexapproximation. J. Geotech. Engrg.,
ASCE,118(1):148153.
AlKhafajiandAndersland,1992),Benson,C.H.andTrast,J.(1995).'Hydraulicconductivityofthirteen
compacted
clays'Clayandclayminerals43(6),669.

DOI:10.9790/168412433741www.iosrjournals.org5|Page

LinearRegressionEquationforSubgradeLateriticSoil
[6].

[7].
[8].
[9].
[10].

[11].
[12].
[13].
[14].
[15].
[16].
[17].
[18].
[19].
[20].
[21].

Azzouz.a.S.,Krizek,R.J.andCorotis,R.B.(1976).Regressionanalysisofsoilompressibility.
Soils
and
Foundation. Japanese Soc.SoilMech.And FoundationEngineering. 16(2), 1929(inBenson, C.H.,Zhai, H.andWang,X.(1994)
Estimatinghydraulicconductivityofcompactedclayliners.
J,Geotech.Engrg.ASCE.120(2),366387.
Boadu,F.K.(2000)'Hydraulicconductivityofsoilsfromgrainsizedistribution:Newmodels',Jouenalof
Geotechnical
andGeoenvironmentalEngineering,ASCE,Vol.126,No.8,pp7,39746.
Blotz,L.R.,Benson,C.H.andBoutwell,G.P.(1998)Estimatingoptimumwatercontentandmaximum dry unit weight for
compactedclays.J.Geotech.AndGeoenvi.Engrg.ASCE,124(9),907912.
BritishStandardInstitution(1990),Methodsoftoilforsoilsforcivilengineeringproperties(BS1377) British StandardInstitution,
London,143Pp.
Das, B.M. (1998), Principles ofGeotechnical Engineering FourthEdition, PWSPublishingCompany,Boston, 712p. Federal
Ministryof Works and Housing (1997). (General specifications forRoads andBridges).FederalHighwayDepartment,Volume II
Ppl45284.
Gidigasu,M.D.(1976),LateriteSoilEngineering:PedogenesisandEngineeringPrinciples. Development
in
GeotechnicalEngineering9,Elsevier,Amsterdam.554p.
th
Gidigasu, M. Dand Kuma, D.O.K (1987)"Engineering significanceoflaterisation and profiledevelopmentprocesses."Proc,9

Regionalconf.ForAfricaonSoilMechanicsand foundationEngineering,Lagos,320.
nd
Head,K.H.(1992)."Manualofsoillaboratorytesting"SoilspecificationAndcompactiontests.
2
ed. Vol. 1.
Pentechpress,London.
Holtz,R.D.andKovacs,W.D.(1981).AnIntroductiontoGeotechnicalEngineeringPrenticeHall,Inc,Englewoods, Cliff, New
Jersey,733p.
Howel,J.L.,Shackelford,C.D.,Amer,N.H.,andStern,R.T.(1997)Compactionofsandprocessedclay
soil mixtures',
GeotechnicalTestingJournal,Vol.20,No.4,pp443458.
Korfiatis,G.P.andManikopoulos,C.N.(1982)correlationofmaximumdrydensityandgrainsize.J. Geotech. Engrg., ASCE,
108(9),11711176.
Kormonik,A.andDavid,D(1969)Predictionofswellingpressureofclays.J.SoilMech.Found.Div Proc. ASCE, 95 (1),
209225.
Lambe,T.W.andWhitman,R.V.(1979),SoilMechanics,SIVersion,JohnWileyandSons,NewYork,
553p.
Loveday,R.(1987).SecondCourseinStatistics.P140.CambridgeUniversityPress,London.
Osinubi,K.J.andAmadi,A.A.(2006).Regressionequationslorpredictinghydraulicconductivityin compacted lateritefly ash
mixtures.Nig.J.ofEngrg,13(1),6169.
Wooltorton, E.L.D. (1975) "Forward" Laterite soil engineering" Pedogenesic and engineering Principles, Elservier science,
Amsterdan.

DOI:10.9790/168412433741www.iosrjournals.org6|Page