Jusae, 1953

Some Recent Foundation Research

and its Application to Design*
By G . G . Meyerhof, Ph.D., M.Sc.(Eng.), A.M.I.C.E.; A.M.I.Struct.E., A.M.ASCE.
The total bearing capacity of a foundation of base
Summary area A b and side (or shaft) areaA , is then given by
The bearing capacity of shallow and deep foundations
under a central load is briefly +cussed, and the effects
of eccentricity and inclination of the load are indicated.
The total and differential settlements of foundations are
estimated in relation to the soil and foundation charac- . .
teristics and the stiffness of the structure. The settle- h
ment stresses in foundation and superstructure are
considered, and recent observations on brick walls and'
claddedbuilding framesare analysed. The lateral
movement and tilting of foundations are estimated and
illustrated by the results of some loading tests on model
portal frames on clay and sand. D

In the past,
Introduction
foundation design was essentially empirical
and basedlargelyonpermissiblebearingpressuresor
loads to avoid failure of the ground. The study of
foundation behaviour in the field amplifiedby laboratory
research has led to a more rational approach in which
l
the results of a site exploration and soil tests form the
basis of an analysis of the stability and probable move-
ment of the proposed foundation and superstructure. Fig. 1;-Theoretical failure zones at bearing capacity
The present paper is mainly concerned with recent of foundations
researchon foundation behaviour and indicates how
the results may be applied to structural analysis. I t .has been shown2 that for a foundation of width B
The main purpose of the paper is to showhow the and depth D in uniform soil
soil, the foundation and the structure canbe considered
B
as one integral unit and to discuss in general terms the
main aspects of this approach. Since foundations form q = cN~Q + Y-
2
Nya . . . . . . (2)
the link between the superstructure and the soil, they
have to fulfil two main interrelated functions : First,
thestability of the foundation soil mustbe assured where y = density of soil
with an adequate margin of safetyand, second, the and NcQand N Q yare bearing capacity factors, depending
movement of the foundation must be within limits that mainlyon 0, depth and shape of the foundation and
can be tolerated by the superstructure. I t is convenient the inclination of the load. The base resistance and
to discuss these aspects separately first and then to show skin friction canbe estimated from the mechanical
how they may be interrelated. properties of the soil and ground water conditions and
from the physical characteristics and method of instal-
1. Stability of Foundation lation of the foundation.
In order to ensure the stability of the structure, the (a) Shaltow Foundations
loads transmitted to the foundation must be within the
ultimate load bearing capacity of the soil. An estimate For footings and rafts when D / B SI, the above ex-
of this bearing capacity at which shearing failure of the pressions can be simplified because the skin friction is
soil occurs can be obtained from plastic theory', 2, using small and may be ignored. For clays (purely cohesive
a number of simplifyingassumptions. Thus the zones materials, 0 = 0 ) the bearing capacity is therefore
of shearing failure in the soil near the foundation are given by the first term of eq. (2),while for sands and
idealised as indicated in Fig. r. In addition, the gravels(cohesionless materials, c = 0 ) the,bearing
shearing strength of the soil is represented by capacity can be represented bythe second term of
S = c + 9 tan o (Coulomb-Mohr's theory of eq. (2). The results of this analysis, which is supported
by extensive researchonmodelfootings and limited
rupture),
where c = cohesion, field data2, are shown in the lower parts of Figs. 2a
9 = . normal pressure on shear plane, and b for a vertical load ona horizontal foundation.
and 0 = angle of internal friction.
In the case of clays the original overburden pressure yD
Further, the effect of the deformation of the soil on the at base level can be added to the graphed values (Fig.
bearing capacity is taken into account by an empirical za). Full submergence of sandsand graveIsreduces
reduction of the shearing strength1, 2. the effective density and thus the bearing capacity by
about one-half (Fig. zb) ;the bearing capacity is however
. *Crown Copyright Reserved. not affected by a water table at a distance greater than
152
about 1.5B below the base so that the bearing capacity The bearing capacity of an eccentric inclined load is
for intermediate positions of thewatertable can be readily estimated by combining the methods of analysis
interpolated accordingly. summarised above, i.e., the reduction factors for inclina-
In practice foundations are frequently subjected to a tion are used in combination with the effective contact
horizontal thrust H and bending moment M in addition width or area for eccentricity. In the case of pad
to a vertical load V . This condition reduces the bearing footings the whole structural H , M , and V are trans-
capacity and can best be considered in two stages :- mittedtothe horizontal bearing area since the soil
(i) the effect of inclination, and (ii) the effect of eccen- resistanceon the vertical sides is negligible. Inthe
tricity of the load on the bearing capacity. The following case of deep blockfootings ( D / B approaching I) with
results have been obtained from a theoretical analysis, large eccentricities (e/B approaching 0.5) however the
which is supported byloading tests on model footings4. resistance on the sides enables somewhat larger forces
For a load that is inclined at an angle a = tan-' (H/V) to be applied to the top of the footings, especially in
to the vertical, the factors Ncg and N y q in eq. (2) are cohesionless soils.
reduced, and for any position of the load on horizontal So far as building frames on single column footings
foundations the reduction factors given in Table I have are concerned, it follows that in the absence of structural
been obtained. failure the maximum H,M ,and V which can be imposed

TABLE 1.-Reduction Factors for Vertical Bearing Capacity of Shallow Horizontal Foundation under Inclined Load
___ - . _ . ~
Foundation Inclination of load with vertical cc
Depth --_I_--

Soil Type
Width O0 IO" 20" 30" 15" 60" go"
DIB
--
Clays ......... 0 to I I .o 0.8 0.6 0.4 i 0.25 0.15 0
I
Sandsand 0 I .o 0.5 0.2 ofor OC = 0 - -

Gravels ......... I I .o 0.6 0.4

0.25 I 0.15
0.05 0

The vertical component of the bearing capacity at any on a column at footing level are interdependent and are
given inclination of the load is then found by multiplying governed bythe bearing capacity of the soilfor the
the bearing capacity for a vertical load (e.g.,Figs. za particular ratios of H / V and M / V . Thus for a given
and b for a central load) by the appropriate reduction H / V and a single eccentricity of the load the maximum
factor. The horizontal component of the bearing bending moment at the column base is
capacity is equal to tan a times the vertical component ;
to this result can be added the passive resistance on the M = Qce (I -ze/B) . . . . . . . . (6)
vertical face of the toe. where QC = total bearing capacity of central load
If the load on a foundation acts at an eccentricity (vertical or inclined). In conjunction with the previous
e = M / V , the maximum toe pressure has hitherto been relationships it can be shown that for clays the absolute
taken to govern the bearing capacity. Thisapproach maximum M = QC B/8 independent of depth, while
is generally safe for small eccentricities but not for for sandsand gravels the maximum M variesfrom
large ones4. A more rationalmethodisobtained by QC BI13.5 for D / B = o to about QC B / I Ofor D/B = I.

assuming that the contact pressure distribution under Similarly, for a given M / V the maximum shearing
an eccentric load is identical with that of a centrally force H at thecolumn base depends on a ;and it can be
loadedfoundation of reduced width B'. Thus for a shown that for clays the absolute maximum H varies
strip foundation with single eccentricity e of the load from about o.zQv for D/B = o to o.4QV for D/B = I,
acting in any direction while for cohesionless soils the maximum H varies from
about o . I Q ~for D / B = o to o.zQv for D / B = I, where
B' = B (I -2423) . . . . . . . . . (3) Qv = bearing capacity of vertical load (central or
eccentric).
so that for clays the bearing capacity The whole analysis in this section applies strictly only
to singlefootings but is alsosufficiently accurate for
q = qc (I -ze/B) . . . . . . . . . (44 groups of shallowfootings and for rafts. Where the
while for cohesionlesssoils soil is variable with depth, theaverage shearing strength
within a depthof about IQ B below the base, controls the
q = qc (I -ze/B)2 . . . . . . . . . (4b) bearing capacity at foundation level provided
strength at a depth of 14 B is not less than about one-
the

where qc = bearing capacity of a central load (e.g.' third of the average upper shearing strength ; if not,
Figs. 2a and b for a verticalload) at given D/B'. then the bearing capacity is governed by the weaker
With double eccentricities e, and e, of the load, the lower layer and can be estimated by spreading the load
minimum effective contactarea A'I, (withstraight to that layer at an angle of 300 to the vertical.
boundary across the base) is found by assuming that (b) Deep Foundations
the centroid of the area coincides with that of the load.
Then the total bearing capacity For piers and piles the skin friction (adhesion) may
formthe major contribution to the bearing capacity
Q = qA'b . . . . . . . . . . . . (5) depending on the soil type and the method of installing
the foundation. In clays, the adhesionvariesbetween
where q is given byeq, (2) but now depends on B' wide limits and, except in the case of piers, governs the
instead of B. bearing capacity. Thus fieldloading tests onbored
June, 7953 I53
in-sit% piles installed in an unlined auger borehole in not exceed one-third of the ultimate bearing capacity
stiff London clay showed that water from the concrete under normal loading conditions *nor one-half of the
softens the clay and reduces the adhesion to 300- ultimate under the most unfavourable conditions, when
700 lb./ft.2 ; on the other hand testson driven piles, even substantial local overstressing of the soil nearthe
in very sensitive clays, indicated that after a month or so foundation has to be avoided. These factors of safety
the adhesion may be taken as equal to the undisturbed .are of the same order as 'those normally used in struc-
shearing strength of the soil within a limit of about tural design ; since the most unfavourable loads act
2000 lb./ft.2(2). Theselower and upper limits of the only for a short time and do not affect the settlement,
bearing capacity are indicated in Fig. za. settlementestimates should always be basedon the
In sands and gravels, however, the skin friction can normalloads in order to decide whether a factor of
usuallybe
ignored andthe corresponding bearing safety of 3 against shear failure is adequate to keep the
capacity is shown in Fig. zb. Where piers or piles pass movements within tolerable limits.
through clay and rest in underlying cohesionless material, While the bearing capacity of most types of founda-
the skin friction of the clay (Fig. 2a) can be added to tions canbeassessedfrom simple soil tests with an
the base or point resistance of the cohesionless soil accuracy sufficient for most practical purposes, the

*J
0

' son
I

YIRY
l
I
sou
1

'
FIBM
I
I
S?JPP
I
I
HMD

COUDSTENCV
Fig. 2 (a).-Bearing capacity of. foundations in clays

corresponding tothe appropriate embedment ratio probable settlement, and evenmore, its distribution
(Fig. 2b).On account of the small point resistance in over the foundation area depends on so many factors
clays, piles should generally not be driven through sand that it can only be estimated within fairly wide limits
or gravel layers, which also assist in reducing settlement except in special cases. In general, the average settle-
by spreading the load. ment of a structure carrying a given load decreases as
Where the pier or pile spacing is less than about four the size and depth of the foundation increases. Since
times the base or point width, the bearing capacity of a stronger soils are usually less compressible than weaker
group is usually given by the base resistance of piers or ones, it follows that the settlement of a foundation is
piles and enclosedsoilplus the skin friction of the likely to .decrease as the bearing capacity of the soil
perimeter of the group ; for stratified soil the method of increases, other things remaining unchanged. For
estimating the bearing capacity is similar to that soils which are uniform with depth the above relation-
outlined above6. The bearing capacity of piers or pile ships canbeexpressed quantitatively within certain
groups under eccentric and inclined loads can be esti- limits.
mated as indicated for shallow foundations with allow-
ance for any lateral soil resistance1. (a) Total SettZemmt
The settlement of a structure due to foundation loads
2. Settlement of Foundation canbe divided into(i) immediate settlementdue to
If the foundation is located below the depth at which deformation of the soil without volumechange, and
the soil is subject to volume changes due to climatic (ii) consolidation settlement due to closer packing of the
or artificial influences, the movements due to structural grains. The immediate settlement occurs mainly on
loads form the criterion tothe permissible bearing application of the load and consists of an elastic (recover-
pressures or loads. To ensure a reasonablemargin of able) portion, which can be estimated from elastic
safety against shear failure of the soil, it has been theory', and a plastic (irrecoverable) portion, which can
suggested6 that the permissible foundation loads should be estimated from plastic theory5 In practice it is how-
154
everfrequentlysufficientlyaccurate to estimate the S = St+St
whole immediate settlementSI from elastic theory, using
a mean secant modulus of deformation Ea at one-half of
the compressive strength of the soil3. Then it can be
shown that for a foundation of width B and bearing
pressure fi where E‘, = effective modulus of deformation
Ea Kv
- . . . . . . .. (11)
0.75 Kv+ UEs
where C = influence value depending mainly on shape Preliminary examination of published data indicatesS
and depth of foundation and thickness of compressible that K , varies from about 25 to 80 times the cohesion
stratum. for normally-consolidated (i.e., recent and generally
soft) clays to from 70 to zoo times the cohesion for over-
consolidated (older and usually firm to hard) clays. In
the latter case the order of the observed total settlement
( U t I)
?
* S = (3 to 7%) B/F (F23). . . . . . (12)
For normally-consolidated clays the estimated settlement
a would be up to three times larger. Further, the limited
U
c field observations indicate that on the same clay the
P
total settlement isup totwo to three times the immediate
movement.
On account of the large permeability of cohesionless
soils, foundations on sands and gravels settle mainly
on application of the load, and for practical purposes
only thetotal movementis required. The results of
laboratory and field loading tests and settlement
observations on buildings with shallow foundations OR
sand indicates that approximately
B3

2
a where B is measured in feet and the water table is at
more than B below base level ; if the .water table is at
base level, the settlement is doubled.
The above relationships (eqs. 12 and 13 with F = 3)
enable a rough preliminary estimate to be made of the
probable settlement of a structure on shallow founda-
tions. After the effect of this movement on the structure
has been ascertained along the lines indicated below, a
more accurate assessmentcan then bemadefrom a
detailed investigation of the particular site.
(b) Differential Settlement
Analysis of field observations on spread foundations in The total settlement is mainly important in relation
clays indicatess that E , varies from about 50 to zoo to access and services of a structureand is sensibly
times the cohesion and approximately unaffected bystructural stiffness so thatit rarely
influences structural design. The differential settlement
however depends not only on the factors governing the
total movement but also on the layout and rigidity of
where F = factor of safety against shear failure (F2-3). the structure and the variation of the compressibility
The magnitude of consolidation settlement St can be of the soil in plan ; it is therefore moredifficult to
estimated from consolidation theory1 and may be estimate and is more important because it affects the
represented by behaviour of the structure under working conditions.
Analysis of field observationss indicates thatthe
CpUB (maximum) differential settlement of footings onnor-
St = . . . . . . . . . (9) mally-consolidated clays at customary bearing pressures
may be very large (several inches) but for over-consoli-
Kv
dated clays rarely exceeds 2 in. For rafts on clay the
where K, = mean modulus of compressibility of soil differential settlementisabout one-half of thetotal
and o < U S 1 is the degree of consolidation at any movement, but the slopes of the bowl-shaped deforma-
given time. tion are so gentle that thedifferential movement between
adjacent columns never exceeds a small fraction of the
The maximum time to be considereddependson the differential settlement of the whole raft.Settlement
economic life of thestructure,andin preliminary records of uniformly loaded strip footings and equally
estimates U can be taken as unity. loaded pad footings of similar size on sands show that
Combining eqs. (7) and (g) the total settlement the (maximum) differential movement is unlikely to
J w e , 1953 G5
exceed three-quarters o€ the maximum settlerhentb. For rigidity E'g I B of the superstructure, which has to be
rafts on sand the differential settlement between adjacent added to E ' p I p of the foundation to obtain the total
columns'may be equal to the differential settlement of E'I onwhich the differential settlement depends, For
the whole raft, which is about three-eighths of the total this purpose two main types of superstructures, multi-
movement. storey building frames and load bearing walls, will be
If the total settlement hasbeen calculated, then on considered.
the basis of these field indications the differential Consider an open multi-storey building frame of
settlement of a structure on shallow foundations can be length L withapproximately equalbays of length I
estimated approximately ; and it hasbeensuggested6 deflecting intotheshape of a trough with a similar
that most ordinary structures can withstand a differ- curvature at each floor leveland a maximum differential
ential movement of $ in.between adjacent columns, settlement A.
fromwhichpermissible bearing pressures havebeen As the beams (including floor slabs in an encased
selected. The author considers however that the steel or reinforced concrete frame) are generally stiff
movements which can be tolerated by a structure can
only be decided in each.particular case on the basis of a
structural analysis. In an earlier paper a methodwas
proposedg in which the foundation and settlement
characteristics of the soil at a site could be interrelated
with the loading, layout and rigidity of building frames
on single footings. This method, which treats the super-
structure, its foundations and the underlying soil as one
complete statically-indeterminate system, willnowbe
extended to give at least a qualitative indication of the
behaviour of different types of structures on different
types of soil.
The present analysis is basedon elastic theory for
bothstructureand soil,whichwould appear to be
reasonable for working stresses of one-third or less of the
ultimate. Thestructure is assumed to beuniformly
loadedwith an effective flexural rigidity E'I which
allows for elastic and creep deformations. Thesoil is
assumed to be an isotropic half-space with a constant
effective modulus of deformation E', which includes
elastic, plastic and consolidation characteristics (eq. 11).
This analysis applies therefore mainly to foundations
on cohesive soils ; in cohesionless materials E', increases
with depth.so that theresults are thenonly veryapproxi-
mate.
When a uniformlyloadedfoundation of width B
rests on an elastic medium, it has been shownlo)n that
the distribution of contact pressure on the base depends
on the relative stiffness Kr of the foundation (E'I per
unit width or length) to that of the soil (expressed by
E'J33). Using the corresponding general equations for
circular and long strip foundations and the finite differ-
enceapproximationsuggested12for other shapes, the
author has computed the settlement and the bending
moments and shearing forces(seebelow)along the
principal axes of foundations of different shapes and
relative stiffness Kr. Details of this analysis are outside
the scope of this paper, but the general results obtained
for the maximum total and differential settlements are
indicated in Figs. 3a and b. The totalsettlement (b) DifferentialSettlement
increases with the length/width ratio LIB of the founda- Fig. 3.-Effect of relative stiffnesson total and difler-
tion, but for any given LIB the (maximum) total settle- entia1 settlements ofuniformlyloadedstructures
and foundations
ment is not very much affected by the relative stiffness
K , (Fig. 3a). It canbeseenfrom Fig. 3b thatthe
(maximum) differential settlement however decreases compared with the columns, the points of contraflexure
rapidly as the relative stiffness Kr increases and, for a in the columns cannow as a first approximation be taken
given Kr, increasessomewhatwith the ratio LIB. The at mid-storey height. Then it can be shown (Appendix
settlementdistribution along the principal axes of in any storey of height h the effect of the
foundations, W+ fqund to vary from an approximately is to increase the flexural rigidity E'Ib of the
circular form for Xr = o to a parabolic form as Kr de beam line to
increased. '

In practice the flexural rigidity of the whole foundation

is frequently zero(i.e.,where a building rests on a
number of individual footings) or is at least small
comparedwith that of the superstructure, especially
where the latter consists of a rigid frame with cladding
E'I
K b
K'
+ K1 + Ku)
+

3
average stiffness of beams,
. (14)
or of load-bearing walls. It is therefore necessary to average stiffness of lower columns
obtain at least an approximate estimate of the flexural average stiffness of upper columns
 The Structural Bngineev
for a rigid frame or the corresponding equivalent stiff- the wholeframebehaves like a soliddeepbeam, the
n e s s in a semi-rigid frame@, flexural rigidity of the framework itself can frequently
Hence for the total superstructure be ignored compared withthat of the wall. The problem
is then similar to that of load bearing walls,when in
E'IB = CE'Z'b of each storey level obtained from the absence of openings
eq* (14).
E'b H 3
E'IB = -- . . . . . . . . . (16)
12

. W
where b = average thickness of wall
and H = height of wall or structure.
Openings in the cladding or walls reduce the flexural
rigidity and require a more elaborate analysis.

3. Settlement Stresses in Foundation and

Superstructure
After the differential settlement of the foundation
has been estimated, it is possible to obtain the stresses
in the foundation and superstructure from normaA
structural analysis. For a uniformly loaded foundation
..
the bending moments and shearing forces can however
U L A T l r t $TlMU€SS O f C o V y D I ~ I O ~ / S O l ~K?' j+ (..$L) be related directly with the relative stiffness K , by the
Fig. 4.-Effectof relative stiffness on maximum
analysis mentioned above. The maximum bending
bending moment and shearing force in uniformly moments and shearing forces increase rapidly with K,
loaded foundations from zero for a perfectly flexible foundation to a maxi-
mum for a perfectly rigid foundation, and they increase
As a check, this method has been applied to a 5-storey, with theratio LIB (Fig. 4). The variation of the
3-bay rigid frame, which had previously been analysed bending moment along the principal axes of foundations
in details. Using eq. (14)it was found that the columns follows that of the settlement, while the shearing force
increased the beam stiffnesses about five times and the increases from zero at the edges of the foundation to a
resulting differential settlement was within 5 per cent. maximum in the outer quarter sections and then de-
of that obtained before. creases to zero at the centre of the area. This variation
External building frames are generally stiffer than is shown in Fig. 5 for a rigid strip foundation for whicha
indicatedabove onaccount of wall cladding. Where closed solution has been obtained (Appendix za).

0.4 0.3 0.1 0 .I 0 0.1 0 0 0.4 0.5

c a/& OtS1ANCI FROM CEAftRE '18 *
fd) ~ro*nSVCRS& UCIVDING MMgN7 (b) TUANSV€RSE SUEARlN6 FORCE

Fig. 5.-Bending moments and shearing forces in rigid strip foundation

this consists of solid panels between the beams and The above results, which are independent of founda-
columns and only shearing stresses can be transmitted tion depth, canbe applied directly tothe design of
from the framework tothe panels, it canbeshown uniformlyloaded rafts and solid load bearing walls
(Appendix Ib) that in any storey the flexural rigidity when the stiffness and thus the bending moments and
shearing forcescan vary withinwide limits. On the
E'pIpL2 otherhand single footings, unlessverywide, may be
(eq. 14) is increased further
by (15) taken as perfectly rigid (Kr = m). Thecorresponding
2 h2
bending moments and shearing forces in a strip footing
where E ' d P = flexural rigidity of panel (in vertical with a central line load (Appendix2b) are shown in
plane). If the wall cladding is fully continuous so that Fig. 5 and differ little from those obtainedby the
June, 7953 357

customary method based on a uniform distribution of open frames, and a maximum tensile stress of, say, two-
the contact pressure. thirds of the tensile strength of the concrete. The
Except in the case of solid walls just mentioned, the stiffening andstrengthening effect of wallpanels in
bending moments and shearing forces in superstructures frames is shown by some (horizontal) racking tests
we best ascertained from the estimated differential carried out at the Building Research Station" on about
settlements. The necessary equations were given IO ft. high and 12 ft. long cased steel frames with solid
previouslye for open (bare or encased) rigid and semi- brick, clinker and clayblockinfillingpanels.Cracking
rigid building frames. Analysis of a few typical frames of the panelswas first noticedwhen the (horizontal)
indicated that considerable settlement stresses could displacement varied from 0.3 to 0.4 in. (8/Z = 1/3oo to
be induced in all structural members and would belargest 1/400)compared with I in. (8/Z = I/IOO) at cracking of
in beams at external joints andin all but centrecolumns,
especially in the lower storeys. It was also shown that
in order to prevent overstressing (as distinct from
collapse) the pennissible bearing pressures or loads for a
given layout of the foundation had to be reduced as the
structure ,became stiffer. Panels and other wall cladding
stiffen exterior building frames appreciably, andthe
proportion of the load carried by the cladding and the
framework depends, on their relative flexural rigidity,
which in simple cases can be assessed from eqs. (14)and
(15); the corresponding bending moments and shearing
forces in the components follow then from the analysis
(Appendix I). In practice this compositebehaviour of
the structure ismuchmorecomplex, and in order to
obtain.some idea of the order of the stresses induced by
the deformation of structures, the results of recent
researchwillbebriefly analysed in relation to load
bearing brick walls and building ,frames. Y
W
E
Some vertical loading tests have recently been carried -x
out at the BuildingResearch Station1* on 8 ft. high a
solid brickwalls without supporting beamson an
IO ft. 6 in. effective span. Cracking started at a central
deflection of 0.004 in. ( A / L = I/ZOOOO) when the
maximum tensile stress was about 50 lb./in.2 ; failure
did not even occur at three times the cracking load. In
similar tests on walls without and with central openings
on shallowreinforced concrete supporting beams no
cracking wasnoticed at the previous maximum load
when the deflection varied from 0.02 to 0.03 in. ( n / L =
1/4000to 1/6000) and the maximum stress was about
IZO lb.lin.2 in the solidwalls. In loading testsfs on
IO ft. high and 30 ft. long lightly reinforced brick walls
withopenings on concrete strip footings,whichwere
jacked up either at the centre or the ends of the walls,
cracking began at a central deflection of 0.05 to 0.08 in.
( A / L = 1/450o to 1/7000) when the average tensile
stress (ignoring the effect .of the openings) varied from
40 to 50 lb./in.2 ; failure took place at about twice the
crackingload. It may beconcludedfrom these pre-
liminary experiments that brick walls on footings under
short-term loading are likely to crack if A / L exceeds
I / ~ O O Q(height/length ratio H / L from I to 3) and the
tensile stress exceed 50 Under
field conditions
the A / L ratio for a similar stress is likely to be much
greaterdue to creep of the brickwork. Thus from Fig. 6 (a).-Plan of WhitehallGardens,London,
observations on brick walls of buildings it was found16 showinglayout of.waterlevelling system
that the maximum stresses a t cracking wereof the
order of 60 lb./in.a andthe corresponding angular the encasement of a similar open frame ;the correspond-
distortion was 3.5 X IO-^ rad. with A/L of about ing average diagonal tensile stress in the panels appeared
I/IOOO. In order to have some margin of safety, to vary from about 80 to 100 lb./in.*. The maximum
a maximum differential settlement' A/L of I/ZOOO and loadwas about I& times the cracking load and for
tensile stress of 30 Ib./in.2 may thus be suggested for the claddedframes it was 14 to nearly 3 times that
buildings with load bearing brick walls or continuous of the openframe. In other racking tests18 on light
brick cladding. frames with and without wall cladding for two-storey
Very little information existsabout the distortion houses the framework took only a few per cent. of the
which encased steel and reinforced concrete frames can total applied load with a corresponding reduction of the
withstand without harmful cracking.Basedon the maximum horizontal deflection. Until further evidence
common limitations of deflections of structural members, is available, the differential settlement between adjacent
to about 1/300of the spanI to avoid cracking of finishes, columns of building frames with wall panels should
the limit 8/Z = 1/300may be suggested forthe permissible therefore be limited to 8/Z = I/IOOO and the maximum
differentialsettlement between adjacent columns of tensile stresses as suggested above for walls and concrete.
1513

Thus, provided the permissible structural stresses are sections and willthereforebebrieflydescribed.This
not exceeded, the limit building, of which construction details have already
A / L (or all) = X beenpublished19,is a fairlystiffcladdedsteel-framed
structure on a raft foundation of variablethickness
where X = constant depending on type of superstruc- resting mainly on London clay (Figs. ba and b). Kcforc
ture construction a settlement estimate wasmade on the
and L = length of wallor I =: columnspacing, as assumptions thatthe whole structure was perfectly
the case may be. flexible and rests on a 145 ft. thick compressible stratum
I t has also been shown that whose settlement characteristics were the same as those
A/S (or 8/S) = I' determined at WaterlooBridge20.Onaccount of the
where Y = constant depending mainly on the relative great foundation depth,the foundationpressurewas
stiffness K , of structure to soil, practically balanced by the overburden pressure except
and approximately, the total settlement in the north-eastern portion (Fig. 6a). For the balanced
areas where a fairlyflexible raft wasprovided, it was
S = ZB/F estimated that during excavation the clay surface would
where 2 = constant depending mainly on soil conditions. rise by up to z in.,whichwouldlargelyberecovered
Hence, substituting, the rcq&red factor of safety on during construction ; subsequent movement due to the
adjacent net loaded area was expected to be small. In
the north-eastern portion, however, the superstructure
was built on an existing stiff raft foundation and pro-
duced an average net bearing pressure of 0.8 t/'L2 with
consequentlong-term settlement ; estimates for the
edge (minimum at point A , Fig. 6b) and centre (maxi-

(a) MOTTLEDC L A Y
-
- c
-. -
2_

Fig. 6 (c).-Minimum and maximum settlements of

points on north-wpstern portion

- -- . ... .. . mum at point B , Fig. 6b) of that areaare givenin

.* .-.
.. ................-
. *.
- e S

:
- e . _
THANET SAND Fig. 6c.
. .*I *

:
L L '
.
' '
;
-.
- 2/61 During and after construction of the building (First
Section) settlement observationsweremade,firstby
L%
"
SCALc
' l
CH A L.K
' ' - 1
7r ordinary levelling and later by means of a permanent
t v 1
0
I

15 50
I

75 too/c water levelling system, installed on the sub-basement

floor, with reference to outside benchmarks on existing
Fig. 6 (b).-Typical section through building and soil buildingsfor the absolute movement(Fig.ha). The
the ultimate bearing capacity of the soil to avoid installation consisted of in.dia.copper tubes along
excessive deformation of the structure is outside walls and principalcorridorswith 5 ft. high

F=
XL
YZB
- or (
. . . G)
(17)
stand-pipes where the water level could be measured in
graduated glass tubes to determine differential settle-
ments with an accuracy of 1/50in. The recorded total
settlement todateandthe relative settlement since
which depends on the characteristics of the soil, the substantial completion of the north-eastern and north-
foundation and the superstructure. This factor depends western portions are given in Figs.7a and b. The former
thus not so muchon the breadth as on the shape in portion ona relatively stiff raft has tilted
slightly without
plan of the structure. appreciable distortion onaccount of the great total
To provide information about the behaviour of framed stiffness. The maximum settlement of about 14 in.
buildingsundergoing settlement , some observations occurred near the northern end where construction began
have been made on a newGovernment Office Block, and the minimum of about g in. at the southern end ;
Whitehall Gardens, London; the results obtained, so thesemovements are of the same order as thosepre-
far, illustrate many of the points made in the last two dicted (Fig. 6c). The north-western portion on a fairly
June, 1953 159

flexible foundation has assumed the bowl-shaped settle- fixed to the steelwork. The stresses .observed since
ment distribution expected on theoretical grounds with substantial completion(Fig.7c) are generally of the
a maximum of about z in. as estimated. same signas those predicted from differentialsettlement,
At a section of the latter portion ( C A ,Figs. 6a and but they are on the whole much smaller (the maximum
h), where the maximum differential settlement between observed stress is,
however, aboutt/in.2 compared
adjacent columns since substantial completion is of the with an estimate of I tlin.2 a t the same section). This
order of 3/16 in. on a span of 25 ft. (Fig. 7b), an estimate difference may be explained by the effect of cladding of

SCALE 1

a' 4 G
, Art WINrER i950/51
f0 Wlh'TER l952/53 I WINTER 1949/50
TO WINTER I9.52/53

FIG. 7(a) O ~ S E R V ETOTAL

D SEfTL€MEw(IN.)
(WINTER 1952/53) FIG 7(b) OBSERVED DfFFERENrfAL ~€~EMEN~(IIV)

Ne rc

?I---

Fig. 7 (c).-Settlement bending stresses in basement storeys of section C-C (Fig. 7 (b)) (Winter
19!50/51 to Winter 1952/53)

has been made of the correspondingbending stresses the frame, which is particularly stiffin the first bay
induced in the structural members of the two bottom where 9 in. thick brick wall panels are provided. This
storeys on the assumption of rigidlyconnected bare composite behaviour is also borne out by the observa-
sections with stiffnesses ( K = Z/Z or Ilk) given in Fig. 7c. tions that on substantial completion of the .building
These estimated settlement stresses at the ends of the maximum stresses were of the order of 3 t/in.2, while
members, where they are largest, do not exceed about in a recent loading test on a beam (fourth span, Fig. 7c)
14 t/in.2 (Fig.7c),whichissmallcomparedwithcus- the measuredmaximumstresses and deflectionswere
tomary design stresses (8 t / k 2 ) for the dead and live only about one-third of those estimated for a bare
loadson a bare frame. I t is of interest to compare frame. I t canthereforebeconcluded that although
these estimates with the stressesmeasured by the the present settlement stressesark .~tnall,theyare of
Building Research Station using vibratihg hire gauges
, some importance, and so f a r have amounted to about
x60 The Stmctwal EHgirteev

one-quarter of the larger stresses due to normal struc- can be estimated, in a similar manner to that proposed
tural loads. for differential settlement, by a structural analysis in
conjunction with the deformation characteristics of the
4. Horizontal Movement and Tilting of Foundation soil based on the above relationships. Such an analysis
and Effect on Superstructure has been made for the lowest columns of building frames
Where a foundationissubjected to a horizontal on.square footings with hinged " and fixed feet to
I' "

thrust H and bendingmoment M inaddition to a determine the horizontal thrust H and bending moment
vertical load, horizontal movement and tilting must be M at footing level after movement (Appendix 3 ) .
expected in addition to settlement. According to The results (Figs. ga and b) show that the thrust H
elastic theory' the stresses in the direction of a load and moment M (expressed in terms of the customary
on an isotropic half-space with effective modulus E', values H, and M,, respectively, for no footing move-
areindependent of the load inclination. For practical ment) decrease rapidly as the relative stiffness K', of the
purposes this is also sufficiently accurate for the dis- column ( E ' I / h ) to that of the soil (E',B3) increases, and
placements so that for a horizontal foundation the as the footing width (B/hl) increases. The decrease in
total movement S in the direction of a pressure 9 bending moment is particularly marked because it also
inclined at a to thevertical can, as a first approximation, depends on the lateral movement as well as tilt of the
be estimated from eqs. (7) to (12) ; the horizontal and footing; since K', can vary in practice within wide limits
vertical components of the movement are then S. sin a (about o to IO or more), it will be seen that unless the
and S. cos a ,respectively. Similarly, for a vertical footings are exceptionally small, the momentmaybe
eccentric ,load and the customary assumption of linear practically zero or evenbe reversed due to lateral
contact pressure distribution, the average tilt of a movementchanging the eccentricity on the base.The
foundation can be computed from elastic theory ; the relative column stiffness atthe head has generally
analytical results obtained by the author may be illus- comparatively little influenceon the resultsbutthe
trated by a square footing of width B (Fig. 8) when the beneficialeffect of .a tie between the feet is clearly
tilt 0 is sensibly directly proportional to eccentricity shown (Figs. ga and b). Knowing H and M the cor-
until the load approaches the footing edge responding stresses in thesuperstructure followfrom
standard analysis. It follows that foundationmove-
,

3M ment, especially horizontal and tilting, can govern the

and 0 = -, very nearly .
. . . . (18) maximum H , M and V which a columncan resist at
EfsB3 least at footing level in the absence of soil or structural
For an eccentric inclined load the above movements failure ; and such movements may therefore affect the
calculatedseparatelyare merelysuperimposed.The mode of failure and limit the collapse load of super-
movements observed on model footings under eccentric structures.
To studythe interaction between simple framed
structures and the soil, especially in relation to hori-
zontal movements and tilting of the foundation, some
loading tests weremadeon 4 in. high and 8 in.wide
model portal frames with I in. square footings resting
on clay and sand. To cover the limiting conditions,
the footings were either pads on the surface or blocks at
I in. depth and covered with sand-paper to ensure full
mobilisation of the shearing strength of the soil ; in
l some tests a stiff tie wasprovided at baselevel.The
portals weremade of annealed mild steel (lower yield
stress = 16.2 t/in.2 from tensile tests) of either in.
square (stiffframe) or in.wide by Q in. deep (slender
frame) section throughout. The soil was soft remoulded
London clay (cohesion = 2.4 lb./in.2) or compact Ham
River sand (angle of internal friction = 45"). Control
tests weremadeon centrally loaded footings (average
results in Table 2) and simply supported beams, which
gave a mean lower yield stress = 16.4 t/in.2 andmodulus '
of elasticity = 12,ooo t/in.2.
The frames were loaded with a central point load on
4 in. width of the beam.Theloadwas applied by a
proving ring at a slow and fairly constant rate of de-
formation until structural collapse or failure of the soil
occurred. The vertical movement of the beam centre
Fig. 8.-Tilting of square footing under eccentric and ends (= footing settlement) andthe horizontal
load movement of the column feet (= footing displacement)
were recordedby dial gauges ; the rotation of the footings
and inclined load on clay* agreed reasonably well with wasobserved by mirrors (Fig. Ioa). In some tests a
the above analysis, but no field observations are avail- vertical load was applied together with a horizontal load
able as a check. on the beam andboth were increased to failure, the
Although horizontal and tilting movement can affect horizontal beamdisplacementbeingmeasured byan
the stresses and stability of load bearingwalls, it is additional gauge. Typical portalsafter failure by
especially important in arched and framed structures vertical load and by combined vertical and horizontal
of single storey height. In multi-storey building frames loads are illustrated in Figs. IO b and c. The main test
the effect is greatest in the bottom storey and decreases results are summarised in Table 2, and typical relation-
rapidlywithdistance from the base. The bending ships between applied load and observedmovements
moments and shearing forces induced by the movement are given in Figs. 11and n a and b for a stiff frame with
JWM,1953 161

soil failure and slender frames with structural collapse of these loads and was accompanied by some spread of
.under vertical and combined loads. the feet unless they were tied.
In general, the load-settlement relation of the footings The rotation of the footings of the stiff frames without
was sensibly independent of the stiffness of the portal tie under vertical load or combined vertical and small
and presence of a tie and was therefore similar to the horizontal loads was generally outwards(negative) at
relationship of the corresponding single footing control small loads and then inwards (positive) at greater loads
test. In frames under combined vertical and horizontal due to lateral movement reversing the original footing
loads the leeward footing settled more than the wind- bending moment ;the same c o n d i t h held f& the slendei
ward footing, roughly in proportion to their respective frames on the surface. When footing spread was

{b) RENDING MOMENT

Fig.9.-Effect of Eelative stiffness on horizontal thrust and bending moment at

base of column on square footihg

vertical loads. The horizontal and tilting movement considerably reduced as in the slender frames at depth,
of the footings of the stiff frame under a given load was or was prevented by a tie, the footings tilted outwards
less than that of a similar flexible frame, but the differ- due to moment reduction. The footing rotation under
encewas not nearly as great as that of the flexural combined vertical and larger horizontal loads followed
rigidities ;the movements of the stiff frame corresponded the overturning moment induced by the horizontal load.
therefore to a much smaller horizontal thrustand Thefoundationmovementsin clay and sandwere of
bendingmoment at the footing comparedwith the the same type and decreased with greater footing depth
flexible frame. The horizontal footing movement of as would be expected. These experimental results give
frames subjected to vertical load or combined vertical some qualitativesupport for the above theoretical
and small horizontal loads was outwards (positive) analysis, whichcanbechecked quantitatively for the
unless prevented by a tie ; undergreater horizontal frameson clay, as illustrated in Fig. Iza. The move-
loads the footing movementoccurred in the direction ments,whichwere estimatedfromequations (38) to
 162 The Strmturnl Engineev

(40) (Appendix 3 ) using an effective modulus of the clay effect of axial load ; no plastic hingeswereformed at
from compression tests a s for immodiatc: settlement the base of the columns.According to plastic theory2’
(Section 2 ) , are of the same order as those observed. the centre of the beam is espected to reach full plasticity
:Is was to i w expected, ; d l t l w tcsts with stiff portals at ;L vertical load of .p lb. for :F frame with a tised base,
c,shilited soilfailurc I)c.causc t h o cstimatcd collapse while at 36 lb. structural collapsc shorlld occur 1)rovidd

(r n-

Fig. IO (a)..Arrangement of loadingtestonmodel Fig. IO (c).-Failure of slender portal frame without

portal frame under combined vertical and horizontal tie under combined vertical and horizontal loads on
loads clay(structural
cpllapse)

load of the frame was 1.14lb. compared with;m estimated

bearing capacity ofthc: soil from 1 0 to 70 lb., depending
on the soil type and footing depth. ‘The load a t failure
(Table 2) was practically the same as that of the cor-
responding control footingssince the inclination and
eccentricity of the load on the soil (i.e., horizontal thrust
and bending moment) werc very small on account of the
large relative stiffness of frame to soil..4ccordingly,
thc~provision
,. of a tie did not affectthe failing load.
1heslender portals failedgenerally by structural
collapsc.Two
framcs
withoutties on thc surface,
however,failed a t ;L reduccdbearing capacity o f thc
soilbecause the small rdativc. stiffncss of the framc.
dlowed a substantial inclination ; ~ n dccccntricity of thc:
loadon thc soil t o h: dwclopxl. To induce ;L soil

Fig. 1l.-Typical results of loading tests onstiff portal

Fig. 10 (b).-- Failure of slender portal frame without framesunder vertical
load onsand(soilfailure)
tieunder vertical load onclay (structuralcollapse)

failurct at clc:ptll, ont‘ of the framcs ( N o . 17) on sand was the horizontal load is less than one-half of the verticd
notannealed. Failure of theothcrportals subjcctcd lo:1d.
to vertical load or combined vertical and small horizontal The observedfailing loads ofslender portals under
loads beganwith a plastic hinge at thecentre of the vertical load(Table 2) arc close totheestimates and
beam and was,with ;F few exceptions (seebelow), somewhat greater due to some strain hardening at the
followedby plastic hinges at the beam-column joints, plastic hinges for all frames with tie and for the frame
usually at the head of the columnsonaccount of thc ondeepfootings without tieonclay. For the frames
 TABLE 2.-Results of Loading Tests on Portal Frames

(a) Footings onClay

-
-
Frame Footing 1 Failing Load I Bearing
-
NO. Section Depth Restraint 1 p---

Beam 1 Column Soil

(total)
Mode
of
Failure
Capacity
of Control
Footing
(mean)
.Remarks

(lb.)
(i) VerticalLoad
21 - - - 34.5
22 Tie - ,*
23
- I - 64 .o
- -
29
24
25
I Ti, -

26.7
1
l
-

-
l*

,*
34.5
26 30.8 e 36.7 Dl

27
28 i Tie
64 .o
,*
(ii) Vertical and Horizontal Loads
Soil 17.5 Vertical Load
30 6.8 Horizontal ,,
31 -
l - i ,l 30.6
20.0
V
H
V
l -
Tie ,
33 H
!. , IO. I V
34 5.2 H
- 32 .o V
35 0.8 H
32 e5 V
36 11.4 H
Tie 33 V
39 11 .o H
-- 11.5 V
37 30.0 H '
l -- -- -~ ~

For observed modes of structural collapse, .see Note below Table 2b.

(b) Footings on Sand

--

Frame I
-I__-.-

Footing I
1
Failing Load
I
t
Mode
haring
Capacity
I .
NO. Section Depth Restraint I Beam 1 Column Soil
(total)
-----__I_-
of
Failure
of Control
Footing
(mean)
--___-I--__-.______
Remarks

(in.)

5 0
I2
13
6
.,
l*

I
IO .,
9 l, Tie
I1 ,,
20 0 - ,l

- Not -4nnealed
1
1 - I -
I
17
18
19
,#

,,
-
Tie
31 e4
33.8 1 35.6
-

(ii) Vertical and Horizontal Loads

0
Vertical Load .
7 Horizontal ,,
16 I
v
H
I

Note : Observed modes of structural collapse denoted thus :-

Frame ( I ) : Plastic hinge a t centre of beam only.
Frame ( 2 ) : Plastic hinges at beam centre and beam-column joints.
Frame (3) : Plastic hinges a t beam centre and leeward beam-column joint.
Frame (4) : Plastic hinge at leeward beam-column joint only.
without tie on the surface of clay and at depth in sand, decreases rapidly as the relative stiffness of structure to
however, the soil resistance tolateraland rotational soil increases, while correspondingly greater bending
foundation movement was insufficient to permit plastic moments and shearing forces are induced inthe structure.
hinges to develop at the beam-column joints and thus The results obtained can be applied directly tothe
reduced the collapse load by up to about 40 per cent. design of uniformlyloaded rafts and centrally loaded
Similarly, for all slender frames at depth in clay under strip footings.
combined vertical and horizontal loads, when failure Recent observations on the behaviour of brick walls
began with a plastic hinge at the leeward beam-column and cladding to frames have been analysed to suggest
joint and wassometimesfollowed by a hinge at the approximate safe limits for the deformation and stresses
beam centre, no plastic hinges weredeveloped at the from which the required factor of safety on the ultimate
windward joint nor at the column feet even when they bearing capacity of foundationscan beassessed. A
were tied ; the corresponding collapse loads were there- study of the movements of a large cladded framedbuilding
fore much less than estimated on the basis of no footing on a raft on London clay illustrates the effect of struc-
movement. It may be noted that for the type of struc- t u r d stiffness on the settlement distribution. While the

'7. a

Fig. 12 (a).-Typical results of loading test on

slenderportalframewithout tie undervertical
load on clay (structural collapse)

ture investigated the maximum theoretical reduction in

the collapse loaddue to foundation movements is
50 per cent., which is however unlikely to arise in
practice. It is hoped that field observations are made
onfull-sized structurestodeterminethelateraland
rotational movement of the foundation as well as the
settlement and to study their effecton the behaviour
of different types of superstructures.

Conclusion 0 ?O I 1 I I J
0 IO to 30 40 50 LO ?O

Rationalfoundation design requires an assessment APP.//o YIRtlCAL LOdO V/LCr)

l
of the stability of the soil and the movement of the
foundation and its effect on the behaviour of the super- Fig. 12 (b).-Typical results of loading test on slender
structure. Theultimate bearing capacity of shallow frame with tie under combined vertical and horizontal
loads on clay (structural collapse)
and deep foundations under vertical load can be esti-
mated from soil and foundation characteristicsunder
most conditions in practice. Eccentricity and inclination estimated and observed totalsettlements compared
of the load reduce the bearing capacity of the soil and reasonably well, the estimated stresses in the steelwork
may limit the maximum bending moment and shearing assumingbare and rigidly connectedmemberswere
force which can be resisted by a column at base level. considerably less than those observed on account of the
Thesettlement of foundations is moredifficult to cladding of the structure.
estimate although approximate relationships have been Although settlement is generally the most important
suggested for preliminary estimates of the total settle- type of foundationmovement, horizontal movement
ment of shallow foundations ;a more accurate prediction and tilting of foundations has a considerable influence
for shallow and deepfoundationscan onlybemade on the behaviour of walls, single-storey structures and
from a detailed site investigation. The differential the lower parts of multi-storey buildings. An analysis
settlementdepends on the layout and rigidity of the of this problem shows that as therelative stiffness of the
whole structure in addition to the soil and foundation lowest columns to the soil increases, the horizontal and
characteristics. Approximate methodshave beende- tilting movements increase, while the corresponding
veloped to estimate the stiffness of structures and to horizontal thrust and bendingmoment at the column
relate it to the deformation characteristics of the soil. base decrease rapidly. Some supportforthis analysis
This analysis shows thatthe differential settlement has been obtained from the results of loading tests on
JMW, 7953

model portal frames onclay and sand, which alsoindicate frame, the component framehas abeam stiffness K b
the beneficial effect of a tie between the feet. and columns stiffnesses K I / Z and K&.
where Kb = K1 = I l / h ~and K , = Iu/huarethe
Acknowledgement stiffnesses of the members of the main frame (or equiv-
.The author is indebted to his colleagues for assistance alent stiffnesses in a semi-rigid frame).
in the settlement observations at Whitehall Gardens and Then for a differential settlement 6 of the supports of
in carrying out the model portal tests. He is par- the component frame, it can be shown that the bending
ticularly grateful to Dr. L. F. Cooling forhis interest and moment at the end of the beam is
encouragement and to Dr. R. H. Wood for stimulating
discussions ; Dr. Wood was also responsiblefor the
measurement of the stresses at Whitehall Gardens.
Mb = 6 E’Kb ( +

Kb+KI+Ku
) l
. . . (20)

The workwas carried out as part of the research

programme of the BuildingResearch Board of the where E’ = effective modulus of elasticity to allow for
Department of Scientific and Industrial Research, and elastic and creep deformations.
the paper is published by permission of the Director of Consider a beam line of length L and flexural rigidity
Building Research. E ’ l b of the main frame subjected to opposite external
bending moments M O (Fig. 13a). Then the central
References deflection
ITerzaghi, K. “ Theoretical Soil
New York, 1943,pp. 118and 265.
Mechanics.” J. Wiley,
n?9taX= n o - A b . . . . . . . (21)
‘Meyerhof,G. G. TheUltimate
‘I Bearing Capacity of
Foundations.” GEOTECHNIQUE, 1951,vol. 2, p. 301.
. ‘Skempton, A. W. The Bearing Capacity of Clays.” Proc.
‘I
MO
Bldg.Res.Congr., 1951, vol. I , p. 180.
4Meyerhof, G. G. The Bearing Capacity of Foundations
‘I

under Eccentric and Inclined Loads.” Proc. Third Int. Conf.

SoilMech., 1953, vol. I, p. 440 b
SMeyerhof, G . G. “ A Study of the Bearing Capacity of Piles.’’
Ann. Bat. et Trav. Publ., 1953 (in press).
6Terzaghi, K., and Peck, R. B. SoilMechanics in Engineer-
I
ing Practice.” J. Wiley, New York, 1948,p. 407.
‘Timoshenko, S.,and Goodier, J. N. Theory of Elasticity.”
I’

McGraw-Hill,
New York, 1951,p. 362. (a) OPEN FRAME
OMeyerhof, G. G. The Tilting of a Large Tank on Soft
‘I

Clay.” Proc. S. Wales Inst. Engrs., 1951.vol. 67, p. 53.

*Meye;hof, G . G , TheSettlement
‘I Analysis of Building
Frames, THE STRUCT. ENGR., 1947,vol. 25, p. 369.
‘ORorowicka, H. Influence of Rigidity of a Circular Founda-
tion Slak on the Distribution of Pressures over the Contact
Surface. Roc, First Int. Conf.SoilMech., 1936.vol. 2, p. 144.
lWorowicka, H. The Distribution of Pressure under a
Uniformly Loaded Elastic Strip Resting on Elastic-Isotropic
Ground.” Final Rep. SecondCongr. Int. Assn. Bridge and
Struct. Eng., 1938, p. 840.
**Ohde,J. “ D’:, Berechnung der Sohldruckverteilung unter
Grundungskorpern. BAUTECHNIK, 1942, Vol. 23, p. 99.
“Pippard, A. J. S.,,,and Baker, J. F. “ The Analysis of
Engineering Structures. E. Arnold, London, 1943, p. 208. CLADDEDF R A M E w m WALL PANEL
14W00d, R. H. Studies in Composite Construction, Part I.”
‘I

Nat. Bldg. Studies Res. Paper No. 13. H.M.Stationery Office,

London, 1952.
lSRigby, C.A., and Dekema, C. J . ” Crack-Resistant Hous-
ing.” Publ. Works of S. Africa, 1952, vol. 11, No. 95.
IeTErzaghi, K. The Actual Factor, of Safety in Founda-
‘I

tions. THE STRUCT. ENGR., 1935,vol. 13,p. 126.

“Thomas, F. G . The Strength of Brickwork.” THE
STRUCT. ENGR., 1953, Vol. 31, p. 35.
l*Short, A., and Simms, L.G. Survey of Loading Tests on,
%mePOSt-War House Prototypes.” THE STRUCT.ENGR., 1949,
vol. 27, p. 67.
‘*Pike, C. F. New Government Offices, Whitehall Gardens.”
‘I

THE STRUCT. ENGR., 1948,vol. 26, p. 217.

*@Cooling,L.F. Settlement Analysis of Waterloo Bridge.”
I‘

Roc. Second Int. Conf. Soil Mech., 1948, vol. 3, p. 130.

*‘Baker, J . F. ‘IThe Design of Steel Frames,” THE STRUCT.
ENGR., 1949, Vol. 27, p. 397.

Appendix Fig. 13.-Component frames of interiorbays of

multi-storey buildingframes
I . STIFFNESS OF MULTI-STOREY BUILDING FRAMES
(a) Ope% (bare or errcased) Frame where A, = downward deflection due to M,
A building frame of length L with maximum (central) MOL2
differential settlement Amax and central points of
contraflexure in thecolumns (Section 2) can be split into
=- . . . . . . . . . . (22)
8E’Ib
single bay component frames of beam length 2 and
+
average storey height h = (h1 htp)/2with hinged ends and Ae = upwarddeflection due to z Mb at every
(Fig. 13a). To represent an interior bay of the main (internal and external) joint
166

when 6 = 21 Amax/L . . . . . . . . . . (24) :.

Transverse bendingmoment at distance a from
centre is
Substituting eqs. ( 2 0 ) and ( 2 2 ) to ( 2 4 ) into (21)
B12
(X - a ) dl~.-
2

I
= p B2 [
-
,
!
(
I - (za/B)2) +
2r:

where E'I'b = effective flexural rigidity of the beam B

sin-' (2a/B)
.- - h} - p ( I -- "1
B
(33)
line or storey.
Hence from eqs. ( 2 5 ) and (26) PB
with a maximum M,,,ax = - at the centre.
29.2
E
Further, the transverse shearing force at a is

(b) Cladded Frame with Panels

An -interior bay of a framework transmitting only
shearing stresses to solid infilling panels can be repre-
sented by a component frame with average beam and
columnstiffnesses of K b / 2 and K c / 2 , respectively,
where K , = Ic/h (Fig. 13b). For a . differential settle-
ment 6 , the moment at end of the beam is M b / 2 , where
M b isgiven by eq. (20) ; further,thetotal bending
moment from the shearing stresses on the panel of
thickness b, shear modulus G and Poisson's ratio 1l.m is with = 0.105 f& at a = o.386B.
M P = G bk 6
The longitudinal bending moments and shearing forces
are zero.

(b) Central Line Load

where E'Jp = flexural rigidity of panel For a footing as before but with a central line load P
= E',bh'/12. per unit length
For a beam line with M , as before
2P
Amax = A, - A b - A, . . . (29) . . . p, =
TB \'(I - (2x/B)2)
-- from eq. (32)
where A,,= upward deflection due to M , at every
panel
HI2
MpL3
==. --- . . . . . . . * (30)
24 E' IJ
Substituting eqs. ( 2 8 ) , (30) andthe previous results
into ( 2 9 ) and putting I / m == 0,the totaleffective flexural
rigidity of the storey is .
= P B - V/ ( I -(z~/B)~) +
L:.
EtpIpL2
E'I'b + -- . . . . . . . . . (31)
zh2
where the first term represents the effect of the frame-
work (eq. 2 7 ) and the lastterm the effect of the panels. PB
with Mmax= -at centre.
2. Bending Moments and Shearing Forces in 21F
Rigid Strip Footing
(a) Uniformly Distributed Load
Thecontact pressure p , at a distance x from the
centre of a uniformly loaded (pressure = p ) rigid strip
Similarly, V. = Ip.
footing of width B on an elastic solid is1' a
June, 7953 167

extension of tie under force Ht in tie for symmetrical

loading and layout, is
Htl
P 28 = -
with V,,, = - at centre. At Et
2
2 ( H -Ht)
The longitudinal bending moments and shearing forces - for footings
from
eq. (38).
are again zero. The transverse bending moments and E', B
shearing forces for uniformly distributed and line loads
are shown in Fig. 5. Eliminating Ht,
3. LateralMovementand Tilting of Footings of
Building Frames
(a) (*Fixed Colwnn Feet
"

Let Fig. r3c represent an internal beam-column joint

of the bottom storey of height h1 of a multi-storey frame
with the (actual or equivalent) stiffnesses of the members Substituting eqs. (39) and (41) into (37), the term
as shown and a horizontal thrust Ho, bending moment B2H/hl in eq. (40) is now replaced by
M Oand vertical load V , on a horizontal and fixed column.
base. If now the endsremote from the joint of all B2H
members are fixed except the base, which is displaced
horizontally by 6 and rotated by e as indicated, it can t (42)
be shown by the slope-deflection method (taking clock-
wise moments and rotations as positive) that the result-
(I+ -)
2AtEt
BZE',
ing horizontal thrust H and bending moment M at the
base of the component frame (Fig. 13c) can be expressed
by
. For an inextensible tie (Et = 00 ) ,B2H/h1= 0.

2E'Kl (b) Hinged Column Feet

"

H =H 0 - P e+ (e+3 S/hl)(2 - If in the above frame (Fig. 13c) hinges are provided
h fill2 W at the columnfeet (Mo = 0) so that after horizontal
displacement 6 the columnbaseremains horizontal
( M = o), it can be shown that the resulting horizontal
thrust
where = Kbl+Kbr+Kl+Ku
and other symbols as before.
The resulting vertical load V affects the settlement of
the footing only and need not be considered now.
For a square footing of width B on a soil with constant where m' = Kbl+Kbr+3Kl/4+Ku
effective modulus E', it was indicated (Section 4 ) that and other symbols as before.
approximately Substituting eq. (38) into (43)
H
a=- . . . . . . . . . . . . (38) H =
Ho
. . e .
(44)
E'&

3M
and ' e = - . . . . . . . . . . . . (39)
Eta3 with symbols as before.
For a tie between the feet as before, from eqs. (41)and
Substituting these relationships into eq. (37)
(43)

H = How-
h1
6K'r { M+ (M+B2H/h1)
(2 - 3K1/2

and M = MO - 6K'r M+ (M+ B2H/hl)

{(I - K & m)
where K'r = relative stiffness of column to soil
, For a rigid soil or inextensible tie, H = H, as in cm-
tomary analysis. The horizontal thrustsand bending
E'I1 moments for "fixed and "hinged column feet are
- shown in Figs. ga and b.
" "

E'sBShl 4

from whichH and M can be determinedin any given case. Discussion

For a rigid soil with E', = 00 , from eq. (40) The Literature Committee would be glad to consider
' H = H , and M = M Oas in customary analysis. the publication of correspondence in connection with the
If between the column feet a flexible tie is provided above paper. Communications on this subject intended
of length I , area At and modulus of elasticity Et, then for publication should be receivedby SeptemberIst, 1953.