Correlation between soil bearing capacity and modulus of subgrade reaction

Apurba Tribedi
Senior Product Manager
Bentley Systems Inc.
Yorba Linda, CA, US
apurba.tribedi@bentley.com

The author is a Senior Product Manager at Bentley. He has been involved in architecting
and coding structural software for more than 18 years. He is one of the core developers
of the STAAD.Pro program and currently manages the STAAD Foundation product. After
graduating from Calcutta University, he joined Research Engineers as a software
developer and has since worked in different areas including graphics, user interface,
database, analysis and design engine.

Abstract: Engineers increasingly using software to design mat as flexible foundation to
save concrete. Instead of soil bearing capacity these software programs often ask for a
property called modulus of subgrade reaction. Why this soil property is needed? Is
there any relationship between these two parameters? Can one parameter be estimated
from the other? This paper digs dip to explain the significance of these parameters and
how one parameter relates to the other.


Introduction

Probablythemostwidelyusedvalueinasoilreportissoilbearingcapacity.Theobviousreasonis
thebasicexamplesgiveninmosttextbooksalmostalwaysusebearingcapacitytocalculatethe
plan dimension of a footing. Because of simplicity and ease of use, that method is still the
fundamental soil parameter for foundation design. However, that simplicity assumes that the
footing will behave as a rigid body. That assumption works well in practice for small and single
column footings. But for large and multi column foundations, most engineers prefer flexible
analysis. Manual computation of flexible analysis could be challenging and in almost all cases
software programs such as STAAD, SAFE, GT STRUDL etc. are used. However, these computer
programsoftenaskforaninputcalledmodulusofsubgradereaction.Manyengineersarenot
familiar with this term and often try to compare it with bearing capacity. As more and more
engineers will use software to design foundations, it is more essential now than ever for
engineerstohaveafundamentalunderstandingofthissoilparameter.Isthereanyrelationship
betweenbearingcapacityandmodulusofsubgradereaction?Herewewilldiscusstheconcepts
andpossiblerelationship.

Modulusofsubgradereaction(Ks)

This term is measuredand expressed as load intensity per unit of displacement. For the English
unitsystemitisoftenexpressedinkip/in
2
/inandinSIsysteminkN/m
2
/m.Someoftenexpresses
thisterminkip/in
3
(orkN/m
3
)whichcouldbemisleading.Numericallykip/in
3
iscorrectbutdoes
notproperlyrepresentthephysicalsignificanceofthemeasuredvalueanditcouldbemistaken
asdensityunitoravolumetricmeasurement.

Mathematically,thecoefficientofsubgradereactionisexpressedas:
(1)

Ks =
p
s
cq. 1

wherep=contactpressureintensityands=soilsettlement

As Terzaghi mentioned,
(2)
proper estimation of contact pressure for a flexible foundation could
be very cumbersome, so it is assumed that Ks remains constant for the entire footing. In other
words, the ratio between pressure and settlement at all locations of a footing will remain
constant. So the displacement diagram of a footing with a load at center will have a dishing
effect. A point at the center of the footing will experience the highest displacement.
Displacementreducesasitmovesawayfromthecenter.Figure1a,showsasimpleslabongrade
foundation. It was modeled and analyzed in STAAD Foundation as Mat, which is a flexible
foundation,andthesoilwasdefinedusingcoefficientofsubgradereaction.Forthisexercise,the
software default value for the modulus of subgrade reaction was used. The displacement
diagramshowsadishingeffectasdiscussedearlier.Figure1bshowsthesoilpressurecontour.It
isalsoobviousthatthepressureintensityatthecenterismaximumandreducesastheelements
(ornodecoordinates)movesawayfromthecenter.So,itistoassumethattheratioofpressure
intensityandsettlementisconstant.

Figure1DeflectiondiagramandSoilpressurecontour

Let us investigate some of the numbers from the same example. Soil pressure, corresponding
displacementandtheratioislistedinTable1below.Thepointsarerepresentedonadiagonalto
illustratethevariationofpressureanddisplacementasthepointsmoveawayfromthecenterto
themostdistantpointinthecorneroftherectangularfooting.Figure2showsthepointsonthe
matslab.

Figure2:Selectedpointstocomparebasepressure,deflectionandratio

Nodenumber Soilpressure(p) Nodedisplacement() Ratio(p/)

(kN/m
2
) (mm) (kN/m
2
/m)
1(topleftcorner) 58.38282 5.377 10858
41 61.94684 5.70524 10858
51 65.56358 6.03834 10858
61 69.19262 6.37257 10858
71 72.64874 6.69087 10858
81(middle) 75.31719 6.93664 10858

Table1:soilpressure,nodedisplacementandtheirratio
Nowthisishardlyasurpriseas,bydefinition,modulusofsubgradereaction(Ks)isaconstantfor
the entire footing and the programused Ks as its soil property. It is also important to note that
the software default Ks value (10858 kN/m
2
/m) was exactly the same as the constant ratio
calculatedintable1.

Base pressure was calculated from the support reaction. So, one might think that the ratio of
supportreactionandcorrespondingdisplacementwillalsobeaconstant.Letusexaminesomeof
the numbers as listed in table 2. Obviously the ratios are not constant for all but for most. This
bringsustoournexttopiconhowKsvalueisusedinsidetheprogramandthebasepressureis
calculated.

Nodenumber SupportReaction(P) Nodedisplacement() Ratio(P/)

(kN) (mm) (kN/m)
1(topleftcorner) 1.313609 5.377 244.3
41 5.575193 5.70524 977.2
51 5.900749 6.03834 977.2
61 6.227366 6.37257 977.2
71 6.538362 6.69087 977.2
81(middle) 6.778522 6.93664 977.2

Table2:Supportreactionanddisplacement

Tributaryarea/influencesurfacearea

Oftenanassumptionismadetocalculatehowmuchareaofaplatecanbeattributedtoanode
or,inotherwords,theinfluenceofeachnodeonthesurfaceareaofaplate.Itdependsonthe
shape of the plate. For a perfect square or rectangular plate, each node will influence exactly
1/4
th
oftheplatesurfacearea(Figure3a).Butforageneralizedquadrilateral,thebestpractice
would be to calculate the center of the mass of the plate and then draw lines from that center
pointtothemiddlepointsofeachside.Theshadedarearepresentstheinfluencesurfaceareaof
thecorrespondingnode(Figure3b).

Figure3:Nodetributaryarea

Calculationofspringsupportconstant

Theabovedescribedtributaryareacalculationisthekeyprocedureusedinternallybythe
programtocalculatethelinearspringconstant.Theprogramfirstcalculatesthetributaryarea
foreachnodeofthefootingandthenmultipliesthemodulusofsubgradereactionbythe
correspondingtributaryareaforeachnodetogetthelinearspringconstantateachnode.

Ky

= Ks Io

cq. 2
where
Ky

isthespringconstantatithnode
Io

istheinfluenceareaofithnode
Ksisthemodulusofsubgradereaction

Foraconcretefoundationanalysis,thosespringshavetobedefinedascompressiononlyas
concreteisassumednottocarryanytensileforce.Thebasepressureiscalculatedateach
supportnodebydividingthesupportreactionwiththecorrespondingnodetributaryarea.Ifwe
lookattheaboveexample,node1hasamuchsmallertributaryareathantherestofthenodes.
ItcanalsobenotedthatallothernodeshavesametributaryareawhichexplainsTable2asit
showsratiofornode1isdifferentthanothernodes.Figure4showsthetributaryareafor
differentnodes.Node1hasatributaryareawhichis25%ofNode81.Table3isanextensionof
Table1andTable2whichshowshowconstantratioisachievedforallnodes.

Figure4:Influenceareaofselectednodes

Node
number
Support
Reaction(P)
Influence
area
BasePressure
(p)
Displacement() Ratio(p/)
(kN) (m
2
) (kN/m
2
) (mm) (kN/m
2
/m)
1(topleft
corner)
1.313609 .0225 58.38282 5.377 10858
41 5.575193 .09 61.94684 5.70524 10858
51 5.900749 .09 65.56358 6.03834 10858
61 6.227366 .09 69.19262 6.37257 10858
71 6.538362 .09 72.64874 6.69087 10858
81(middle) 6.778522 .09 75.31719 6.93664 10858

Table3:Reaction,basepressure,displacement,Ksconstant

BearingCapacitydependencyonallowablesettlement

Bearingcapacityisthemeasurementofthesoilpressurewhichsoilcansafelybear.Inother
words,bearingcapacityisthepressurewhichsoilcanwithstandbeforeitfails.Thetwomost
importantsoilfailurecriteriaare:

1) Shearfailure
2) Maximumallowablesettlement

Amongmanyfactors,foundationwidth(B)caninfluencefailurecriteria.Normally,shearfailure
governsforsmallerfoundationsandsettlementfailuregovernsbiggerfoundations.The
followingtableisatypicalexamplewhichshowstherelationshipamongdifferentfoundation
sizesandfailurecriteria.

Shape B
m
L
m
q
a
(kPa) Governing
Criteria
Square 1 1 113 Shear
2 2 117 Shear
3 3 111 Settlement
4 4 92 Settlement
6 6 75 Settlement
10 10 64 Settlement

Table 4: Final allowable bearing capacity for allowable settlement = 25 mm. and a given
embedmentdepth

Toestimatesettlementfailure,anallowablesettlementvalueisassumed(normally25mmor1
inch).Whensoilsettlesmorethanthatallowablevalue,thesoilfails.So,evenforabearing
capacitycalculation,anallowablesoilsettlementisusedandstructuralengineersshouldbe
awareofthatvaluewhiledesigningafooting.Theallowablesoilsettlementvalueistypicallyan
integralpartofanysoilreport.

Whyusethemodulusofsubgradereaction

Itwaspreviouslystatedthattodesignaflexiblematfoundation,themodulusofsubgrade
reactionisusedinsteadofbearingcapacityofsoil.Butwhyisitso?Theanswerliesinthe
underlyingassumptionsofhowafoundationmightbehave.

Foundationscanberigidorflexible.Bearingcapacityisusedtodesignrigidfoundationsbut
subgradereactionisusedforflexiblefoundations.Theveryassumptionofarigidfoundationis
thatthedistributionofthesubgradereactionpoverthebaseofthefoundationmustbeplanar,
becausearigidfoundationremainsplanewhenitsettles
(3)
.Letusconsiderasimplysupported
beamloadedatcenterasshowninthefigure5a.Bystatics,wecanobtainR1=P/2andR2=
P/2.Ifthesamebeamisloadedeccentrically,reactioncanbecalculatedasshownin5b.

Figure5:Reactionsforasimplysupportedbeam

Thesameconceptisextendedforrigidfoundationdesign.Butinsteadoftheendsupports,the
wholefoundationissupported.Itisalsoassumedthattherelativestiffnessoftheconcreteslab
ismuchhigherthanthesoilstiffness.So,theslabisassumedtoremainplanarevenafterthe
applicationofload.

Figure6ashowsafootingloadedatthecenter.Fromarigidwidebeamanalogy,P=RxL.
Similarlyforaneccentricallyloadedfootingthereactionwillvarylinearlyfromoneendtothe
P
R1
R2
P
R1
R2
L
a
R2 = P x a / L

P = R1 + R2

R2 = P R1
(a)
(b)
otherasshowninfigure6c.Equations3and4canbesolvedtofindendreactions.Butnoneof
theequationscontainmodulusofsubgradereaction(Ks).So,thedistributionofsubgrade
reactiononthebaseofarigidfootingisindependentofthedegreeofcompressibilityofthe
subgrade
(4)
itisrestingon.Asmanyauthorsconcluded,arigidfoundationcanbesafely
designedusingbearingcapacityasinmostcasesthismethodyieldsmoreconservativeresults.

P =
1
2
I(R
1
+ R
2
) cq. S

P o =
1
6
B
2
R
1
+
1
S
B
2
R
2
cq. 4

Figure6:Subgradereactionsforanisolatedfooting

Butamatfoundationisoftendesignedasaflexiblefoundationasitcanbelargeinsizeandthere
maybemanyloadapplicationpointsandothercomplexities,suchasholesandgradebeams.
WidespreadavailabilityofFEAsoftwarecontributestothistrend.Butaflexiblefoundation
cannothavelinearsubgradereactionunlikerigidfoundations.Rather,itdependsonthe
compressibilityofthefoundationaswellasthestructuralrigidity.Aflexiblefoundationwillbe
subjectedtointernalbendingandrelativedisplacementsbetweentwoslabpoints.Thegreater
thestructuralrigidityis,thelesstherelativedisplacement.Theauthortestedthecasewithvery
highelasticityoftheslabelementsanditresultedinanearlyplanarsurfaceaftertheapplication
oftheload.Similarly,thegreaterthemodulusofsubgradereactionis,thelessthepressure
distribution.InotherwordshigherKsvaluewillabsorbmorepressureattheloadapplication
point.Hence,themodulusofsubgradereactionwhichisthefunctionofsoilsettlementand
theexternalpressureisusedforflexiblefoundation.

P
R
P
R2
R1
(a)
(b)
(c)

Correlationbetweenbearingcapacityandmodulusofsubgradereaction

Themostcommonandprobablythesafestansweristhatthereisnocorrelation.Butthere
shouldbeone,asbotharethemeasurementsofsoilcapacitiesandanyofthesetwoparameters
canbeusedtodesignaregularfoundation.

LetuslookatthedefinitionofKsagain,whichisthepressureperunitsettlement.So,inother
words,soilcapacitytowithstandpressureforagivendisplacement.Fromearlierdiscussions,itis
alsoclearthatevenbearingcapacityhasanallowablesettlement.So,itistemptingtoconclude
thatmodulusofsubgradereactionisthebearingcapacityperunitsettlement.

ThisconclusionisverysimilartotheequationpresentedbyBowles.
(5)

SI: Ks = 4u(SF)q
u
kNm
3

FPS: Ks = 12(SF)q
u
kt
3

whereSF=Safetyfactorandq
a
istheallowablebearingcapacity.

Intheaboveequations,theallowablebearingcapacityisfirstconvertedtoultimatebearing
capacitybymultiplyingwithasafetyfactor.Theauthorassumedoneinchor25mmsettlement.
Thefinalequationisthenformulateddividingtheultimatebearingcapacitybytheassumed
settlement.

Themoregenericformoftheequationcanbewrittenas:

Ks =
Iq
u
o
strcssJisploccmcnt

where

I=Safetyfactor
q
a
istheallowablebearingcapacity
oistheallowablesoilsettlement

Fromaboveequations,itisevidentthattheappropriatesafetyfactormustbeusedandtheKs
valuecanbebettercomparedwithultimatebearingcapacityratherthantheallowablebearing
capacity.Thesafetyfactorcanvarydependingonprojectsandgeotechnicalengineers.Theother
importantfactoristheassumedallowablesettlementforthecalculatedbearingcapacity.

Howevertheabovementionedequationshaveitslimitations.Itcanbeappliedtothefootings
wheresettlementfailuregovernsbutcannotberelatedtothefootingswhereshearfailure
occursbeforereachingallowablesettlementlimit.So,Engineersmustexercisecautionbefore
usingtheseequations.

Conclusion

Thecorrelationbetweenbearingcapacityandmodulusofsubgradereactionisatbest
estimation.ItcanbeusedforestimationbutKsvaluedeterminedbyaplateloadtestshould
alwaysbeusedifavailableorshouldberequestedwheneverpossible.However,theabove
discussiongivesinsightintothesevaluesandhelpsengineerstounderstandthephysical
significanceofmodulusofsubgradereaction.

References:

(1),(2),(3),(4)
Soil Mechanics in Engineering Practice (Third Edition) Terzaghi, Peck, Mesri
(5)
Foundation Analysis and Design (Fifth Edition) J oseph E. Bowles