SOILS AND FOUNDATIONS Vol. 12, No. 3, Sept.

1972
Japanese Society of Soil Mechanics and Foundation Engineering

A METHOD TO ESTIMATE
MODULUS OF HORIZONTAL SUBGRADE REACTION
FOR A PILE

IWAO YOSHIDA* and RYUNOSHIN YOSHINAKA**

ABSTRACT
This paper proposes an experimental formula to estimate the modulus of horizontal sub-
grade reaction k to be used in design practice for analyzing the lateral deflection of a pile
at the top and the ground surface.
In this paper, (1) the relation between the values of deformation modulus measured by
different soil investigation methods was obtained from the results of a comparison study,
(2) the effect of loading width to the modulus of horizontal subgrade reaction was investi-
gated by the horizontal plate loading tests carried out in test pits, and (3) the influence of
deflection of a pile at the top or ground surface to the value k was investigated on the
basis of some experimental data of horizontal loading tests. According to those results,
the experimental formula to estimate the value k was obtained.

Key words: bearing capacity, deformation, in-situ test, modulus of horizontal

_re~_s.ti<?E_ 1_ pile

IGC: E2/D5

INTRODUCTION
Many studies have been conducted on estimation of the modulus of horizontal subgrade
reaction, k, for a pile (or piles) on the assumption that it may be considered as a beam on·
an elastic or inelastic foundation and various formulae are proposed for the value . of the
modulus k.
At the same time efforts have been made by many investigators to obtain numerical
solutions using those values of k 'for analyzing the lateral behavior of piles. Generally,
expressions fork which have been proposed may be included in Eq. (1) or in a similar form.
( 1)

where p is the horizontal subgrade reaction, y is the lateral deflection of the pile at its depth
x, and k, m and n are constants representing the subgrade condition. m (>O) shows the
tendency to change with depth in soil and n (l>n>O) shows the stress-strain characteristics
of soil, that is, n=l means elastic and n=Fl non-linear elastic.
The constants, m and n, in Eq. (1) are evaluated variously: for n=l by Chan (1937) (m

* Chief of Foundation Design Section, Honshu-Shikoku Bridge Authority, Shiroyama·cho, Shiba-Nishi-

kubo, Minato-ku, Tokyo.
** Assistant Professor in Foundation Engineering, University of Saitama, Urawa, Saitama.
Written discussions on this paper should be submitted before July 1, 1973.

This is an Open Access article under the CC-BY-NC-ND license.

NII-Electronic Library Service

2 YOSHIDA AND YOSHINAKA

=0), Row (1956) (m=1), Terzaghi (1955) (m=1), Reese-Matlock (1960) (m=1) and Palmer-
Tompson (1948) (m, arbitary), and for n:;t:1 by Rifaat (1935) (m=1), Shinohara-Kubo (1964)
(n= 1/2, m=1), Hayashi-Miyajima (1963) (n= 1/ 2 , m=O) and so on. Using these constants,
the solution was obtained by each investigator.
The studies to obtain a general expression of the modulus k such as Eq. (1) or others
are very important to analyze the lateral resistance of piles. However, dealing with prac-
tical design problems, we can not help thinking of the following difficulties:
( 1) Is it possible to determine each constant in Eq. (1) by means of the presently avail-
able techniques of soil invesigation methods ? Even if possible, can it be accomplished
with any degree of accuracy in comparison with efforts for the complicated and trou-
*
blesome computations required for the case of n /1 and m>O ?
( 2 ) Its it possible to adequately evaluate the influence of loading width on the modulus
of horizontal subgrade reaction ?
To deal with the problems described above, we present in this paper the following ap-
proach which have been proposed by Yoshida (1964) and Yoshinaka (1967).
On the assumption that the theoretical function F (x) to express the exact lateral deflec-
tion process of a pile is obtained, the constants required in F (x) are for instance a, b, c etc.
While point A in Fig. 1 can be determined by a straight line of the secant OA in place of the
function F (x), then the constant required to show the point A is only the inclination
angle a.
At the present time no satisfactory soil investigation method is available to determine the
constants a, b, c, etc., in order to express the function F (x) with sufficient accuracy, and
therefore, the point A may be shown by rough approximation within a range of some
tolerable error, using the only constant a.
When the state of actual deflection of a pile is shown in Fig. 2 (a), the same deflection
y 1 can be obtained from the unified modulus of subgrade reaction converted into the sim-
plified state as illustrated in Fig. 2 (b).
If the values required in the design are the lateral deflections of the pile at some special
points such as at the top or the ground surface, or the maximum bending moment as in
many cases, they could be obtained by determining the ground condition corresponding to
the state of Fig. 2 (b), even if the bending state over the whole length of the pile differs

ELASTI~
......
CLAY 1
·-)(

.Lt..

--'llo-X ACTUAL STATE CONVERTED STATE.

(a) (b)
Fig. 1. Approximate expressions of Fig. 2. Deflection of a pile and soil reaction
function (Fx)

NII-Electronic Library Service

 MODULUS OF HORIZONTAL SUBGRADE REACTION 3

between Fig. 2 (a) and Fig. 2 (b).

Our treatment is justified in view of the characteristics of the lateral deflection of a pile ;
that is, (1) the influence of the type of expressions on the modulus k decreases in the fol-
lowing order: the lateral deflection, the bending moment and the shearing force, and (2)
because the effective depth of ground with respect to deflection of the upper part of piles
is limited within relatively shallow depths as pointed out by many investigators, it is not
necessary to obtain the modulus k with the same accuracy along the whole length of piles.
Based on the consideration described above, we present a method to evaluate the modulus
of k for the practical design, using the simplest expression of Eq. (1) which assumes m=O
and n=l. Also a study on the method to measure the deformation properties of subgrade
is described. In addition, a study on the modulus k, fundamentally based on the proposal
by Yoshida (1964), conducted by Imai et al. (1969) using LLT (lateral load tester), new con-
cept on the specific, standard and general k-values is introduced.

RELATION BETWEEN THE MODULUS OF DEFORMATION OF

SOILS AND THE TESTING METHODS
There are many methods to measure the deformation properties of soil. But the values
measured do not always agree with other results due to the difference of the methods used
to determine the modulus of deformation. As well known, the difference between static
and dynamic methods is conspicuous, but even among the static methods the difference is
considerable.
Static methods include the uniaxial and triaxial compression tests employing undisturbed
samples, the plate loading test, the bore-hole deformation test, etc. In this paper an ex-
perimental study carried out by Yoshida and Yoshinaka (1967) and Yoshinaka (1967) on the
relation between the values of the deformation modulus obtained by these methods is shown.
The calculation of the modulus of deformation of soils or foundations is based on the elastic
theory, in which the Poisson's ratio of soils is assumed to be about 0. 3.
Fig. 3 shows the relation between the values of the modulus from the compression tests
on samples and from the borehole deformation tests (the pressure meter tests) conducted in
the same soil condition. The latter consists mainly of two types : the unicell type devised
by Kogler (1933) (60-80 em in length of probe, 50-80 mm in diameter), and the tricell
type by Menard (1962). It is already confirmed by Yoshida and Yoshinaka (1967) that the
difference in the mechanism of the both types does not affect the measured values of the
deformation modulus. Therefore, the values determined by these methods are not distin-
guished in this paper.
Though the stress condition caused by each test method differs, Fig. 3 indicates that the
results of the both methods agree very well, namely, Ec=Eb where Ec is the modulus of
deformation by compression tests on undisturbed soil samples and Eb by bore-hole defor-
mation tests.
Fig. 4 shows the relation between the results from the horizontal loading tests with the
circular rigid plate and the borehole deformation tests conducted at the same point in the
ground consisting of sand and clayey soil. It is shown that there is a good correlation
without the influence of soil type and the relation may be approximated as
( 2)

where Eso is the modulus of deformation of soil by a bearing plate, 30 em in diameter. The
deformation properties of soil from the loading plate will be described in the next chapter
in detail.
As the standard penetration tests are most commonly performed in the field, it is con-

NII-Electronic Library Service

4 YOSHIDA AND YOSHINAKA

~
5000
%
v
3000
v
2000
> v
v
~
1000

500
(.1

300
X y X
X
w
(/) 200 v
I~
::J <(,'Q
....J
::J ~ I
Cl 100
0
~

50
<QJ
1/
z
i= 30 ~~
.....

.... ... 7
<[
~ : SAND(TRIAXIAL CONFINED
2:
0::: 20 e: t9a= ~x depth)
~
CLAY (UNIAX)
0
IJ... ......
w A : CLAYEY SOl LS (UNIAXIAU
Cl 10
/. X : SOFT ROCK (CLAYEY SILTY)
5
3
/
v 0 :SOFT ROCK (TERTIARY)

~ : KANTO LOAM (TRIAXIAL)

2

I /
I 3 5 10 30 100 300
I
1000
I
3000
I

(Ycnf)
DEFORMATION MODULUS

Fig. 3. Relationship between Eb and Ec

(Et : modulus of deformation by bore-hole deformatian test, Ec : by compression
test on undisturbed sample)

~ 1.8'n J 2.4'nl J
I • I 71 I 8.8 IAVE.VAUJEt •: SAND
I x I 32 I 4. 9 I ., ., I )(: KANTO LOAM

-
~7
.0

w
IJ..
0
0
1- 3
<(
0:::2
6
5
4
~
~
~
-- ---- -~

Es : MODULUS OF DEFORMATION BY PLATE TEST I

Eb : MODULUS OF DEFORMATION BY BORE-HOLE TEST

30 60 90 120

01 AM ETER OF LOADING PLATE {em)

Fig. 4. Comparison of Es and Et

NII-Electronic Library Service

 MODULUS OF HORIZONTAL SUBGRADE REACTION 5

(~~'
1000
-~( ~~
0)~
o\

//
v
~~~
~
IV

~f ~b
r-
.CI vo~
w 500

z 300
r-
...
v Y. ~~
0
~ !!>
(JI ~ q,.
~4)) v Q•

~ ~

~
.:E
0::: ~ ~tO
0 10or }(' r~

x_A ~·
IJ.. ....
r-
w r-
l
IJ..
0
0
50

30
r-

r-
)('
....
A~
I
~ . ~'Q

(f)
'Y, ~ ~0
3 20
/j vx
:::>
0
0
10 // / ..
'l
::E
• SAND

5 0 : CLAY

0 : GRAVELY SOIL

X :( SILTY SAND

LOAM
I I I L I I I I I I
I
1 2 3 5 10 20 30 50 100
STANDARD PENETRATION RESISTANCE ( N-VALUE)

Fig. 5. Relationship between Eb and N-value

venient if the .Mvalue can be converted into the modulus of deformation. The relation of
N-value and the modulus of deformation is shown in Fig. 5. From the figure it is clear
that there is some correlation between these measured values statistically and the relation
may be expressed as,

( 3)

where Eb is the modulus of deformation defermined by the borehole tests and N is the stand-
ard penetration resistance.

INFLUENCE OF LOADING WIDTH ON THE MODULUS OF

HORIZONTAL SUBGRADE REACTION
The elastic theory indicates that the settlement of a loading plate is simply proportional to
the width of loading, but it is doubtful whether the same relation holds for soil. It is well
known that the settlement of a foundation on sandy ground decreases with the width of
the foundation, as shown by Terzaghi (1955), that is, the modulus of vertical subgrade re-
action in sand increases with the loading width, relatively.
We carried out a study on the similar relation for the horizontal case in test pits as shown
in Fig. 6. The soil conditions are as follows: in case of sand, the test pits were filled with

NII-Electronic Library Service

6 YOSHIDA AND YOSHINAKA

DIAL GAUGE
LOADING
PLATE
I

(ALL SIZES IN m)

I...;> 5 :oRDER OF TESTING

LOADING
PLATE

,..___ _ _ _ _ _ 5.5

Fig. 6. Arrangement of horizontal loading tests

compacted soils, Gs=2. 71, rd=l. 40-1.47, and w=8-13%, and in case of clayey soil, the
ground consisted of a natural deposit of volcanic ash, the so called Kanto-loam, Gs=2. 73,
rd=O. 52-0.65, w=ll0-135%.
The test plates used for loading were rigid and circular, 30, 60, 90 and 120 em in diameter.
The sequence for loading is arranged from smaller to larger plates, the rate of loading was
slow, and the loading pattern was cyclic.
The results of experiments are summarized in Fig. 7. The figure shows the ratio of the
k value obtained by plates of various diameters to that of 30 em in diameter versus the
diameter of plates. The value of k is defined as,
Load intensity (kg/cm2 )
k ( 4)
Settlement (em)
and is calculated from the secant of the linear part below the yield load of a stress-strain
curve. From Fig. 7 it is clear that the influence of loading width on the horizontal sub-
grade reaction k is expressed by
k
( 5)
ko
where l is constant being about 3/4 and B 0 is the standard diameter of 30 em.
Eq. (5) can be expressed by the modulus of deformation Es given to diameter B, that is,
the value Es obtained from the result of a loading test using the relation between k and Es
based on the elastic theory ;

NII-Electronic Library Service

 MODULUS OF HORIZONTAL SUBGRADE REACTION 7

X SAND
• KANT5-L.OAM
7.0

~+ ,,~~~~
', ,~,,
',',... '',
',:.,,
',, ' ...' ........ (
LL ' .................. :: ·~ .. .R"'
... ,
0
.... ........ .... ..::-· ..-?~3
.... ................. _...
0 0.5 ....... ... ....- .............
............ .e .. -!(:..--.. . . _____
__ ------- .... ::::-----~
1- .3---
Y.q
..........................
<!
0::: --._e"··----- ------
I ----------

30 60 90 720

DIAMETER OF LOADING PLATE (em)

Fig. 7. Effect of loading width on k-value

E.
k (6)

where IP is the shape factor and p. is Poisson ratio of soil.

By substituting Eq. (6) at B and Bo in Eq. ( 5 ), then
B ) 1-z
E.=Eso ( Bo ( 7)

Thus, Eq. (7) becomes an experimental formula for correcting the modulus of deformation
of soils for the effect of loading width.

LATERAL DEFLECTION AT PILE TOP

A pile subjected to a horizontal force at its top resists with the lateral subgrade reaction
and the rigidity of the pile itself. As mentioned in the preceding section the subgrade
reaction on piles is so complicated it is difficult at the present time to express the modulus
k accurately.
However, considering the deflection from the practical point of view, the following two
points are to be noted:
( 1 ) The effective depth of soil relating lateral deflection of piles is limited only to the
upper part of piles, even if the pile is driven in the foundation consisting of several strata
with different engineering properties.
( 2) When piles are used for the foundation, the deflection which may actually occur is
usually less than a few centimeter at the ground surface, considering the relation between
the upper structure and the soils.
We investigated the modulus of subgrade reaction k to be used in practical design of piles
from the preceding point of view, on the bases of the following assumption ; piles are sup-

NII-Electronic Library Service

8 YOSHIDA AND YOSHINAKA

ported in the lateral direction by the simplest condition, i. e., by independent springs with
same constant (the so called Winkler's hypothesis). In this case, the behavior of a pile is
expressed by

(8)

From Eq. (8), the deflection of a free head pile at the ground surface, y 1 , is given by

( 9)

where

f3=f kB
4Eplp

The characteristics of the modulus k in Eq. (8) and Eq. (9) could be obtained by applying
the relation of each corresponding value of the loading tests.
The piles tested consisted of steel pipe piles, prestressed concrete piles and cast-in-place
concrete piles of 30 em to 150 em in diameter, and were installed in various test sites in
Japan.
Fig. 8 in the logarithmic expression shows
CLAY the relations between the ratio of k to kt (at
STEEL PIPE PILE y 1 1 em) and the deflection y 1 of piles at the
¢ =150 em ground surface.
4.0 6.0 Yf It is concluded clearly from Fig. 8 that the
modulus k has a functional relation with the
deflection y 1 within its limited range.
Thus, an experimental relation is estab-
30
lished as follows :

12D SAND em
J--+----+--+---=.li:~---STEEL PIPE PILE f/61 log( : ) =-mlogy 1 (10)
~ - PECAST CONCRETE 1
"'="::- PILE ¢50cm
..-::-:::~lt.o 2.0 4.0 6.0 Yf and in a general form,
OIZ Oft o.~ oa ~~---.. cm
b.6 ' ..... (11)

10:4 where, a and m are constants.

It seems that the constant m represents
SILTY CLAY the intensity of plastic and non-elastic pro-
~..P...-.---+--1--~ CAST IN PLACE perties of subgrade within the effective depth,
CONCRETE PILE ¢100cm and the adhesive force between the pile and
4.0 o Yt the surrounding soil.
em The values of m obtained from the lateral
loading tests are summarized in Table 1.
It should be noted that the modulus k dis-
cussed in the above, strictly speaking, can
Fig. 8. Relation between (k/k 1) and Yf only be applied for an estimation of the de-
flection at the pile top.

NII-Electronic Library Service

 MODULUS OF HORIZONTAL SUBGRADE REACTION 9

Table 1. Experimental values of m

m 0"""0.3 0.3"""0.4 0.4"""0.5 0.5-0. 7 0. 7-1.0

driven diluvial
pile (elastic) sediments, or sand and (plastic)
soil condi- clay and sandy soil
clayey soil
tions
cast-in place generally in case of
pile Y!>20mm

ESTIMATION OF k-VALUE

Derivation oj k
The modulus of horizontal subgrade reaction k to be used for the practical design of
piles can be obtained from the facts described above.
When the load intensities acting on soil by plates, B 0 and B in diameter (average load
on the rigid plates) are both p, and the settlement caused by this load in case of Bo is Yo,
then the settlement of the plate with the diameter of B is, from Eq. (5),

B
Y=Yo ( Bo
)-z (12)

If the relation concerning the settlement of loading plates is applied to piles, the following
derivation on k may be possible.
The modulus k for B 0 is, from Eq (11),
(13)

and Eq. (13) holds for other piles of B in the same supporting condition, where ao is a
constant and y 10 is the lateral deflection of pile of Bo at the ground surface.
By substituting Eq. (5) in Eq. (13),

B
k=aoY!o-m ( Bo
)-1 (14)

and similarly from Eq. (12) to Eq. (14), then

(15)

Thus, the modulus k can be obtained by Eq. (15), using the deflection properties of a
standard pile of Bo driven into the ground under the same condition.
On the other hand, when we estimate the modulus k by Eq. (15), applying the results of
soil investigation, it is necessary to determine the constant ao in Eq. (15).
If the modulus k for piles is given in the form of Eq. (16) as for an infinite beam on
elastic foundation defined by Vesic (1961) as described later, that is,
kB=a'Es (16)

By substituting Eq. (7) in Eq. (16), the following equation is obtained;

(17)

NII-Electronic Library Service

10 YOSHIDA AND YOSHINAKA

Thus, it may be said that Eq. (17) is the relation taking into account the properties of
soil as compared with k for the elastic foundation by Vesic. However the k value by Eq.
(17) corresponds to the strain condition defining Eso.
Since Eq. (15) and Eq. (17) become equivalent respectively under the conditions of B=Bo
and Yt=l, ao in Eq. (15) is;
ao=a'Bo- 1Eso (18)
By substituting Eq. (18) in Eq. (15), the general equation for k (in kg per cubic em) is
obtained,

(19)

(19')

where a' is the constant determined from the bending characteristics of piles.
The details about a' will be discussed later. Fig. 9 shows the effects of the deflection
y 1 and the diameter of piles on the modulus k with parameters l and m.
The value of m obtained from experiments are summarized in Table 1.

7.5 B =CONSTANT
ko (Bo=30cm, Yt=lcm)
X
J.o
LL
0
0
f= 05
<(
a:
0
0 7 2 3 4 5
DEFLECTION OF PILE ( Yt em)

Yt = CONST

0 0~--~-----2~.-----3~--~4-----5~--~6~

RAITIO OF ( B/s0 ) .

Fig. 9. Effect of Yf and B on k-value

NII-Electronic Library Service

 MODULUS OF HORIZONTAL SUBGRADE REACTION 11

The Constant a' of Eq. (19)

1) On the theoretical values
The spring constant for a beam with a concentrated load, placed on the infinite-elastic
foundation, has been studied by Vesic (1961), (1963).
According to his results, when a beam satisfies the requirement for infinite length, ..1L>
2. 25, the modulus for the beam becomes constant, and is determined by

(20)

Broms (1964), Francis (1964) and Bolows (1968) et al., suggested that the relation expres-
sed by Eq. (20) can be applied to piles, and they have used Eq. (20) for estimating the
modulus k for the lateral resistance of piles.
However, Francis considered that the pile condition differs from the beam mentioned
above, that is, piles are completely surrounded by foundation materials, and therefore the
modulus k for piles should be twice as much as the k value obtained by Eq. (20), i. e.,

kB=1 3012/ EsB4 . ~

• v Ep/p 1-p 2 8
(21)

For the practical use, Eq. (21) may be simplified since the effect of 12
-V Es is relatively
small, then,

kB=l. 3012/ B' ·E -12.~ (21')

V !p P 1-p 2 8

The range of 1 V-B!p4

is limited due to high order of the function, and Francis (1964)
showed that this value was about 1. 2-1. 3.
It has recently been pointed out, however, that this value appears to range from 1. 4 to
1. 6 in many cases of concrete and steel pipe piles.
The modulus of elasticity of piles is 2. 1 x 106 kg/cm2 (steel pile) and 2. 0-3. 5 x 10 5 kg/cm 2
(concrete pile), so the term of 12-VE; becomes about 3. 4 (the former), and 2. 3 (the latter).
Now, assuming Poisson's ratio of soils to be 0. 3, then Eq. (21) is further simplified as
follows:
kB=a'Es (22)
in which
: 1/1. 6 for steel pipe piles
{ : 1/1. 4 for concrete piles

Hence, the modulus k by Eq. (22) can be applied if the foundation is elastic. Studies
such as by Bergfelt (1957) on a beam contacted on the both sides by elastic half spaces
has been conducted by Gramholm (1929), Biot (1937) and Jampel (1947) et al.
Biot considered a sinusoidal loading condition for a beam, and assumed that half of the
load acted on the beam was supported by each side, giving the following relationship:
8rrG//1.1(1- ,Us)
kB (23)
Eplp k
ln--:84+7.85-ln B

NII-Electronic Library Service

12 YOSHIDA AND YOSHINAKA

Applying the general dimension of piles and the following relation held in an elastic
body,

(24)

then Eq. (23) is simplified as

kB= (1/1. 6--1/2. O)Es (25)

By Jam pel (1947), the following equation was obtained :

BrcEs/B (26)
kB
Eplp k
ln-B 4-+7. 96-ln B

In the same manner, we obtain

kB= (1/1. 03--1/1. 88)Es (27)

From the consideration mentioned above, it seems that when a beam is considered as a
pile, the spring constant (kB) can approximately be expressed in the form of Eq. (22), and
the value of a' is about 1. 0--1/2. 0 depending upon the assumption for calculations.
2) On the experimental values
It is possible to examine the constant a' by applying the experimental deflection curves
of lateral loading tests for a pile to Eq. (19).
When Eq. (19) gives the modulus k for piles, the value a' of the reverse calculation men-
tioned above must have a tendency converging to a special value.
The deformation modulus Es to be used in Eq. (19) is obtained by averaging the values
between the ground surface and the depth near the point of maximum bending moment of
piles. This range is considered to be the effective depth on lateral deflection.
As the first step to examine, the constant a' was calculated under the condition in which
no influence of deflection y 1 on the modulus k was considered and the deflection was as-
sumed approximately elastic being in the range y 1 <10 mm.
Strictly, the amount of deflection y 1 to be in equivalent strain to each pile due to the
difference of diameter of pile is not equal as shown by Eq. (12), but it may be permitted
to examine the applicability of Eq. (19), even though by this rough condition.
The results of calculation (Yoshinaka
(1967)) are shown in Fig. 10. In the figure,
8
the solid line shows all the values using
7
- TOTAL Es from test results by many different
>- 6 methods, and the dotted line shows only
--- DATA BASED
~ 5 the values by the bore hole deformation
w ON Eb
0 4 tests which are considered relatively reli-
w
ff 3 able.
2 It seems from Fig. 10 that the measured
values have a predominant one although
scattering in the manner of the normal
74 distribution, and the value a' is about 1/
VALUE OF 1. 3--1/1. 6.
Fig. 10. Distribution of experimental value of a' Next, we examined the value a' by

NII-Electronic Library Service

 MODULUS OF HORIZONTAL SUBGRADE REACTION 13

considering the effect on the modulus k, according to Eq. (15) and Eq. (18), on the experi-
mental data based on the bore-hole deformation modulus Eb and N-value.
In the calculation process, m was determined by using the k-y 1 curve of each pile. The
results are shown in Table 2 and Fig. 11.
Fig. 11 shows the relationship between the experimental values from Table 2 and the
theoretical value by Eq. (22).
Consideration
In the preceding paragraphs the value of a' to be used in Eq. (19) was considered.
From those considerations, it is clear that the value a' can be taken constant judging
from the experimental results, and that this value is almost equal to the value calculated
by Eq. (22).
That is, a' obtained from thirty five examples converged to the value between 1/1. 3,...., 1(

Table 2. Summary of the calculated results

PILE B Eso a a' m Ip soil a'** test

81.28 33 2.14 1/1.54 0.50 3/4 2-3x 105 sand 1/1.65

101.6 33 2.95 1/1.30 0.50 3/4 4-6x 105 sand 1/l. 53
steel pipe 101.6 9 0.48 0.56 5x10 5 clay Ministry of
1/1.85 3/4 1/1.53 Construction
pile 9 clay
152.4 0.63 1/1.46 0.56 3/4 1. 7x 106 1/1.74
60.96 25 1. 36 1/1.86 0.28 3/4 1.1 X 105 {fine 1/1.54
sand
precast concrete {fine
pile 50.0 25 2.70 1/0.93 0.43 3/4 1. 3x 105 sand 1/1.45

* Eso=3Eb the average over the effective depth

** theoretical value

2.5

2.0
':J 1.\t"
\;.
0'
c:.,-<-· • ta'"
• -<._'!
••
0
...J 7.5
<( 0 ~
~~
(.)
"& 0~ ,oi'
r- 0\:: -<..\.
w
0::: 7.0
0
~
"'
w RC: RC. PILE
I
r- ST: STEELPIPE
0.5 PILE

0.5 1.0 7.5 2.5

EXPERIMENTAL
Fig. 11. Relationship for a' between experimental and theoretical value

NII-Electronic Library Service

14 YOSHIDA AND YOSHINAKA

1. 6. But in this case the data were not classified with respect to the materials of piles:
steel pipe pile, precast concrete pile, cast-in-place pile, etc., considering the accuracy of
the modulus of deformation to be used for analysis. The theoretical values of a' are ap-
proximately 1/1. 6 in case of steel piles, and 1/1. 4 in concrete piles, whicn are generally used.
It may be concluded from the above described results that the value a' are;
1/1. 6 for steel pipe piles
1/1. 4 for precast concrete piles
or the calculated value by

(Kg} Finally, we will show an example in

1500 which the results obtained in the above
were applied to the deflection behavior
of a pile tested by R. Miyamoto, M.
Sawaguchi (1968) et al., in Fig. 12. The
w driven piles tested were 126. 3 mm in
(.)
Ct:: diameter, 15. 5 m in length, and the mo-
OIOOO~--~~~~--~~L4----+---~
tJ.. ment of inertia of pile section was 2, 910
cm4 •
_j
<( The foundation soil was soft alluvial
1- clay extending to the depth of 30 m. The
z
2 effective depth (0.-.1/3--1/4lmt) was about

I
g; 500 1--1!--1--:H=--l----+-
0--1.8 m, and the modulus of deforma-
tion E. (secant modulus corresponding to
half the ultimate compression strength),
though gradually increasing with depth,
5. 7 kg/cm 2 at the depth of 1. 8 m, obtained
by unconfined compression tests on undis-
turbed samples.
0 1.0 2.0 3.0 6.0
Yt (em) CONCLUSIONS
Fig. 12. Comparison of estimated and experi- From this experimental study on the
mental pile deflection Yr effect of testing methods and the loading
width to the measured value of the deformation properties, and on the lateral deflection
behavior of a pile under static loading, the following results are obtained;
( 1 ) The measured values of the modulus of deformation of soils vary depending upon
the methods used. It is found that the following relations hold :
Ec ~Eb
E.o~3Eb
Eb~7N

where Ec is the modulus of deformation measured by uniaxial or triaxial compression tests

on undisturbed samples, Eb by the borehole deformation test, E. by the plate loading test
and N by Standard Penetration Test.
( 2) The modulus of subgrade reaction kb in the horizontal direction by plate loading

NII-Electronic Library Service

 MODULUS OF HORIZONTAL SUBGRADE REACTION 15

tests is influenced by the width of loading.

From the results of horizontal loading tests with various diameter, the following equation
is obtained

k=ko ( Bo
B )-1
where 14:3/4.
( 3 ) Analyzing many data on lateral loading tests for a pile, the following result on the
modulus of subgrade reaction for a pile is obtained. This is valid when the lateral deflec-
tion of the pile at the ground surface is relatively small (approximately less than a few
centimeter), and it assumes that the lateral behavior of the pile follows the Winkler's hy·
pothesis.
The modulus k by reverse calculation (from Eq. 8) used the stress-strain curves of loading
tests at each deflection y 1 , can be ~expressed by the equation on an experimental basis,

where a 0 and m are the constants determined from the properties of soils and the pile.
( 4) By using the results presented in (1) to (3) the equation for estimation of the mod-
ulus of subgrade reaction k for the practical design can be obtained by

1 )( B
k=3a'Eb ( !3-; Bo
)(3/4)(m-1)
·y 1 -m (kg/cm3 )

where B 0 is the standard width of a pile (30 em), the constant a' equals about 1/1. 3,..., 1/1.6
from experimental results. The value a' almost agrees with the theoretical value given by
the following equation proposed by Vesic (1961), Francis (1964), Broms (1964), et al.,

1 3
.
lv B4 .
Eplp 1-p$ 2
1

Eb is the average deformation modulus over the effective depth (approximately, from the
ground surface to the depth of the maximum bending moment of piles).
ACKNOWLEDGEMENTS
The work reported herein was mainly carried out in the Foundation Section of Public
Works Research Institute, Ministry of Construction. The authors are grateful to the Insti-
tute for their invaluable assistance.

NOTATION
B=diameter of loading plate or pile
Bo= B of 30 em in diameter
Ec=modulus of deformation of soil by a compression test on undisturbed sample
Eb=modulus of deformation of soil by a bore-hole deformation test
E$=modulus of deformation of soil by a plate loading test
Eso = E, for 30 em in diameter
Ep=modulus of elasticity of pile
G.s=specific gravity of soil
H=lateral load
]p=moment of inertia of pile section

NII-Electronic Library Service

16 YOSHIDA AND YOSHINAKA

k=modulus of horizontal subgrade reaction

L =length of beam
N=number of blows by Standard Penetration Test
w=water content
x=coordinate distance along pile length
y =lateral deflection of pile
y 1 = y at ground surface
l, m, n, a, and a'=constants
(3 = r-k_f3--
4Eplp
ra=dry density of soil
). = (3 of beam
Ps =Poisson's ratio of soil
lmt=depth of first zero moment

REFERENCES
( 1) . Bergfelt, A. (1957) : "The axial and lateral load bearing capacity, and the failure by buckling of
piles in soft clay," Proc. 4th ICSMFE, Vol. 2, pp. 8-13.
( 2) Bolows, J. E. (1968) : Foundation Analysis and Design, pp. 503-507, McGraw-Hill.
( 3) Broms, B. B. (1964) : "Lateral resistance of piles in cohesive soils," Proc. ASCE, Vol. 90, No. SM
2, pp. 27-63.
( 4) Chan, Y. L. (1937) : Discussion on "Lateral pile-loading tests" by Feagin, Trans., Vol. 102, ASCE.
pp. 272-278.
( 5) Francis, A. J. (1964) : "Analysis of pile groups with flexual resistance," Proc. ASCE, Vol. 90, SM
3, pp. 20-23.
( 6) Hayashi and Miyajima (1963) : "Study on the horizonal resistance of H-piles," Port and Harbour
Res. Inst., and Yawata Steel Ltd., (in Japanese).
( 7) Imai, T. (1969) : "Study on modulus of horizontal subgrade reaction ( 3) ," Japanese Society of
SMFE, Vol. 17, No. 11, pp. 13-18, (in Japanese).
( 8) Kogler (1933) : Bauingenieuer, 14, pp. 266, cited by E. Schult2e and H. Muhs (1967).
( 9) Kubo, K. (1964) : "A new method for the estimation of later 1l resistance of pile," Rept. Port and
Harbour Res. Inst., Vol. 2, No. 3, pp. 1-37, (in Japanese).
(10) Matlock, H. and Reese, L. (1960) : " Generalized solutions for laterally loaded piles," Proc. ASCE.
Vol. 86, SM 5, pp. 63-91.
(11) Menard, L. (1962) : "L'evaluation des tassements, tendances nouvells," Sols=Soils, Vol. 1, No. 1,
pp. 13-28.
(12) Miyamoto, M., Sawaguchi, M., et al. (1968) : "Field experiments on lateral pile resistance in soft
clay ground," Technical Note of Port and Harbour Res. Inst. No. 47, pp. 3'-34, (in Japanese).
(13) Palmer, L.A. and Thompson, J. B. (1949) : "The earth pressure and deflection along the embeded
length of piles subjected to lateral thrust," Proc. 2nd ICSMFE, pp. 156-161.
(14) Rifaat, I. (1935) : Die Spundwand als Erddruck Problem, A. G. Gebr. Leeman and Zurich, Co.
(15) Row, P. W. (1956) : "The single pile subjected to horizontal force," Geotechnique, June.
(16) Sawaguchi, M. (1968) : "Soil constants for piles," Rept. Port and Harbour Res. Inst., Vol. 7, No.
2, pp. 65-94, (in Japanese).
(17) Terzaghi, K. (1955) : "Evaluation of coefficient of subgrade reaction," Geotechnique, Vol. 5, No.4.
(18) Vesic, A. B. (1961) : "Bending beams resting on isotropic elastic solid," Proc. ASCE, Vol. 87, No.
EM 2, pp. 35-53.
(19) Vesic, A. B. (1961) : Discussion on ''Generalized solutions for laterally loaded piles" by Matlock
and Reese (1960), Proc. ASCE, No. SM 3, pp. 123-125.
(20) Vesic, A. B. (1963) : "Model studies of beams resting on a silt subgrade," Proc. ASCE, Vol. 89,
No. SM 1, pp. 1-31.
(21) Yoshida, I. (1964) : "On the coefficient of horizontal subgrade reaction for design of piles (II), ..

NII-Electronic Library Service

 MODULUS OF HORIZONTAL SUBGRADE REACTION 17

Civil Engineering Journal, Vol. 6, No. 11, pp. 21-29, (in Japanese).
(22) Yoshida, I. and Yoshinaka, R. (1967) : " Comparison study of the bore-hole deformation test
methods," Civil Engineering Jour., Vol. 9, No. 6, pp. 3-8, (in Japanese).
(23) Yoshinaka, R. (1967) : " The modulus of horizontal subgrade reaction and its correction on the
loading width," Technical Note of Public Works Res. Inst. No. 299, pp. 1-52, (in Japanese).

(Received November 9, 1971)

NII-Electronic Library Service