O~ THE ASSESSMENT OF THE DYNAMIC RESPONSE OF SOILS
Alemayehu Teferra
Civil Engineering Department
Addis Ababa University
ABSTRACT
The paper discusses two approachu for determining
the dynamic ruponse s~ctra of structuru in which
the effect of site conditions are taken illlo
consideration. From the response s~ctra the
utimalion of maximwn base shear is indicated. In
light of the prue111ed discussion, the earthqlillJ/ce
loading as stipulaled in the Ethiopian code ofpractice
was examined. It has bun found 0111 that the loading
criteria given in the code for stnlcture.r founded on
soft soiLr need modijicollon and in consequence
thereof a re~ndalion is proposed.
INTRODUCTION
Sit.ea that are subjected to lbak:ina or vibration either
by man made meua (blutina, uaderiround oxj,losion
etc... ) or by natwa1 comequeoce1 (eartbquab.) may
suffer considerable damaae. The effect of a man
made vibration on a aiveo aite may be determined
since the extent and inteo.sity of the vibration is
controlled at will. But di.is is not the cue for
earthquakes.
The extent and intenaity of around lbak:ina caU8od
by ID earthquake are known only after the event.
The oooarw:e of euthquabe may lead to di.wtet
unle1a appropriate counter lm9IUJ1ll are taken. To
provide ID .dequate l&fety apimt di.wtet, it ia
necessary to have a hip dearee of expertile in
earthquake eoaineerina when in tum requires a deep
knowlodp in undentandina the effect of pound
motion on soil properties and that of ltnlcture8. The
study of the effect of around motion on soil
proeper:ity is aeaerally known u potocbenical
earthquake eoaineerina.
Geotecbeoical earthquake eupeerina ia a
relatively youq decipline. It ia in the early sixties
that systematic investiaation on the response of soils
as a telUlt of sbakina and influence of site effects on
the behavior of structures durina earthquakes were
undertaken (1). It will be the taak of this paper to
preeent an overview of the different techniques used
by acieotists and eoaineen to &88e88 the response of
soils when subjected to earthquake shaking.
Soils reepood differently to shaking according to
their characteristic properties, thickness of deposit,
peUem of stratification and site conditions. Loose
aranular deposits tend to compact due to around
vibration raultina in excessive lettlements. In cases
where the looao aranular deposits are saturated, the
around sbakina will induce the development of excess
pore wat preuure which would be sufficient to
C&U1e liquet.ction (complete lou of bearing caplcity)
of soil. Theee types of soil imtabilitiel would
undoubtedly C&U1e catastrophic damap to structures.
However throup 1ppropriato around investiaation
and desip, it is possible to avoid such danaer.
Apart from the above imtabilitiea or permaneat
deformations soil respond to around sbakina and
affect the structures that are founded on them. Tbe
am unent of thia respomo ia of equal importance to
the safe desip of structures. Coaaidenble effort bu
been expended to develop methods furUl I I Fiq Ibo
respomo of IOit. induced by _pound lbakiaa
Cumady there are two metbodl b (mtiltical)
rel.atiomhipi (in tbe form of curves) wbidl have been
deduced mtistically from establiabecl i-ttoma
oblervod in actual earthquakea. Tbe IOCQlld method
ia buod Oil theoretical analylia. Before delviq into
the discuuioo of the above two metbodl, it ia
appropriate at tbia .... to briefly dilcuu:
) respomo spectra
b) tbe behavior of soil chancteri.ac. wben
IUbjected to VlbratiOll
Journal of&EA, Vol~ 9, 1992
 Alemayehu Teferra

RESJ>0NsE-sPECTRA" Hence:

Seismic waves which are generated during an

earthquake travel in all directions away from the (x-)')
T
-K
fj(t)e
' -211
-(t-T)
r
source and be recorded by means of stron1 motion 2
recording instruments. These instruments usually
record the time history of accelerations. The data
2
obtained directly from the recording instruments .sin " (t-t}dt (2)
should be digitix.ed and the necessary corrections for T
instrument, noise, base-line, etc. errors be conducted. From Eq.2 , follaw:
A typical corrected accelerogram is shown in Fig.1
(3). From the accelerogram one would be able to see
the peak acceleration, the duration and the variation
(:J-j) fY<.t)e -r
t -2H(I
-T
)

of the acceleration during the period of vibration.

However, its effect on structures cannot be inspected
2
directly. In order to study the response of buildin1s .sin ; (t-t)dr (3)
when subjected to acceleration, the following
methodology has been introduced (4).

Consider a one-degree of freedom mechanical

system having a mass m, a linear spring with a spring
constant of k, and viscous damping c, acted upon by
an acceleration y(t) as shown in Fig.2. Let x denote 2Kn(t -t,....
.sm-- u. (4)
the absolute displacement of the mass m, any y
T
denote the absolute displacement of the ground. The
extension of the spring will thus be (x-y). The
respective absolute accelerations of the ground and The maximum values of x-y, :J-y and f are
the mass are y and x. The response of the mass m, important in engineering design work.
when acted upon by y(t) is given by the following
equation (4). Since the same integral expression appears in all
equations, the maximum value of the integral
expression will be donated bys. i.e.:
T -2"< - )
(x-y) J.5'(T)e---r-
I
. t 2A(t-T)
2nJt-n 2
o
, s,. [ f ;c-i>e r

2
sin ; b-n 2 (t-T')dr (1)
2
.ain (t-t)dr ] (')
Where T max

T = undamped natural period of oscillation

=2~~ Hence

(x-y) -
T s (6)
- 2n "
. ca1 damp"mg = -c
. of cnti
n = fraction ~
-
2
T = time parameter.
For small damping (n < 0.2) , Eq. 1 may be
simplified by equatin1 b-n2 1.

Journal of E:A.EA, Vol. 9, im.

 San Fernando Earthquake Feb. 9, 1971 - 0600 Pst
II-I 137 71.135.0 15910 Ventura Blvd., Basement. Los Angeles. Cal. Comp S81E
0 Peak Values: Accel = 140.2 cm/sec/sec Velocity= -16.1 cm/sec Displ = - 7. 1 cm
-250

250'--~~~~~~..L.....~~---'-~~~__..__~~--JL..-~~-1...~~~.l-..~~_J...~~~..L_~~__J~~--'~~~~

-20

20'---:-~~~~----'~-:--~__..___~~---L~~~..l-~~--L~~~J..._~~-1...~~__JL.,_~~-1-~~--1~~___J

-8

0 10 20 30 40 50 60
(From Caltech, Vol. II)

Fig. 1 .:fypical corrected accelerogram and Integrated velocity and displacement time histories [3].
4
.Alemayehu Tefe"a

Time t

a) Spring-mass system
b) Typical earthquake ground
motion

Fig. 2 Determination of response 5pectra

0. 4
O'
' 0 .2
i::
0
~
.j..) 0
rd
H
Q)
~
Q)
u
u
.i: I I I . I I I . I .

0 5 10 15 20 25 30
Time, s

a) Accelerogram, El Centro, California Earthquake,

May 18, 1940 (N-S component)

i::
'
-~ 1. 5 t--.,.--ttft---+---+---+-----1---+---l---~
.j..)
rd
H
Q)
~ 1. 0 ~----1----1--
u
u
~
~ 0 .5 1---,,11.,,C....:P.,.....~lo.l.-4.'-

~
~
~~=f:=:E=d
0 .....__....___.....___.....__ _,
c 0.5 1.0 1. 5 2.0 2. 5 3.0 3.5 4.0
Natural Period, s

b) Acceleration response spectra for El Centro

(l.940) earthquake.

Fig. 3 Accelerogram and acceleration response spectra for El Centro earthquake [1].

Journal of .&4.EA, Vol. 9, 1992


Dynamic Response of Soils
i 2x S
~ T"
(8)
T 2n
If one plots - S" versus T, or-S" versus T
2x T
for several values of the damping factor n; one
obtains the displacement, or the velocity, or the
acceleration response spectra. This operation is
normally done by engineering seismologists.
A response spectrum is therefore, the maximum
response induced by the ground motion in a single
degree of freedom oscillators of different fundamental
periods, but having the same degree of internal
damping.
The acceleration response spectra for El Centro
earthquake is reproduced in Fig.3. From the
acceleration response spectra one should be able to
determine the maximum acceleration of a building if
its natural period of oscillation and damping are
known (Fig.4).
BEHAVIOR OF son. CHARACTERISTICS
WHEN SUBJECTED TO VIBRATION
..
The response of soils to vibration is determined
mainly by their shear modulus and damping
characteristics. The stress-strain relationships of soils
under cyclic loading conditions are non-linear in form
(Fig.5) This means, the shear modulus, which is the
slope of the curve at given point, would vary. For
ease of computation, however; the shear modulus
may be expressed as the secant modulus determined
by the extreme points of the curve say for the two
conditions shown in the figure by G1 and G2 [5]
The damping factor is proportional to the area
inside the hysteresis loop (Fig.5). These soil
characteristics depend upon the magnitude of Qie
strain for which the hysteresis loop is determined.
Hence they are functions of the induced strain.
A wide variety of tests both in situ and in
laboratory have been conducted to determine shear
moduli and damping characteristics for a wide
spectrum of strain values [5]. Seed and Idriss made
a detailed parametric evaluation of a previous work
done by Hardin and Dmevich [6] and proposed
5
simplifiw equation for shear modulus of sand as
follows
(9)
Where
G Shear modulus (b!.ft2}
a = effective stress (lb/ff)
K2 a parameter given in Fig. 6
One may use Eq. 9 also to estimate the shear
modulus of gravelly soils with the appropriate value
of the parameter K2 which is given in Fig. 7. The
corresponding value of the shear modulus in kN/m2
may be obtained by multiplying the result of Eq. 9 by
the factor 0.479.
For saturated clays, Seed and Idriss [5] have
presented Fig. 8. In Figs . 9 and 10, the damping
ratios for sand and for saturated clays are given.
GROUND RESPONSE AS A RESULT OF
GROUND SHAKING
As stated earlier, the ground response may be studied
using the analytical or the empirical methods, all of
which have been developed at the University of
California, Berkeley. It should be stated that due to
the lack of sufficient earthquake records, the
empirical method was introduced at a later stage.
These two methods will be presented briefly.
ANALYTICAL METHOD
Closed-Form Solution for Homogeneous Soils
The vibration of soil layers induced by earthquakes
are mainly attributed to the upward propagation of
shear waves from the underlying rock of rocklike
layer. If the soil layers and the rock or rocklike
layer are essentially horiwntal, the lateral dimension
of the deposit does not affect the response and the
deposit may be considered as a semi-infinite layer
transmittin& one dimensional shear beam in the
vertical direction. It is using this approach that Idriss
and Seed [7] developed their method. In the
analysis, the materials are considered to be linear
elastic and their properties (shear modulus and
damping) may vary with depth following a simple
mathematical expression.
Journal of EAEA, Vol. 9, 1992
6
Alemayehu Teferra

-
-
-...

Natural Period T 0.3 s T o.s s T = 1.0 s

Damping Factor ). o.os ). = o.os ). = o.os

Maximum. Accn. umax o. 75 g umax 1.02 g umax 0.48 g

0.4
b>
..
c: 0.2-
.....0
.
0
M
"'
M
Q)

Q)
() 0.2
()
<
0.4
0 5 10 15 20 25 30
Time, s

Accelerogram, El Centro, California Earthquake, May 18, 1940

(N-S Component)
1.2

b>

c:
.. 1.0

-~
. a.a
"'
M
Q)
M
Q)
()
0.6
()
<
0.4
-~x
~ 0.2

0
0 o.s 1.0 1. 5 2.0
Undamped Natural Period, s
Acceleration Response Spectrum, El Centro Ground Motions

Fig. 4 Evaluation of acceleration response spectrum (1).

Journal of .&t.EA., Vol. 9, 1992

 7
Dynamic Res~nse of Soils

Fig. 5 Hysteretic Stress strain relationships at different strain amplitues

1--.....;;.._ _ _ _ _-1-_ __ _ -------+------

Shear Strain - percent

Fig. 6 Shear moduli of sands at different relative densities [S].

Journal of EAEA, Vol. 9, 1992

8 .Alemayehu Tefe"a
220

200
I - -- +- DlnH Hnd and Hndy
9ravel (Southern California)

190
---..... --
160

10

120 --
100

80

60

20

0
10

Sh.Mr ltraln ... percent

Fia. 7 Moduli determination for aravelly aoila [S].

30,000
x Wilson and Dietrich (1960)
'nliers (1965)
6 Idriss (1966)
10,000 + Zeevaert (1967)
Shannon and Wilson (1967)
7llZ Shannon and Wilson (1967)
VJ 'nliers and Seed (1968)
3,000
8
Kovacs (1968)
Hardin and Black (1968)
t---1 Aiaiks and Tarshanaky (1968)
1IIIlI Seed and IdriH ( 1970)
1,000
~Tsai and Housner (1970)

....Q_
su ...... ......... e.
300
G>

100

..., x
301------+------+------'1------+-----i-----+-i-----+.-----+--~'----------l

10'-----'-----L-----'------'------''-----_.__ _ ___._ _ _ __.__ _ ___.___,_ _ _. j

10- 4 10
Shear Strain - percent

Fia. 8 ,In-situ ahear moduli fo~ saturated clays [SJ

Journal of EA.Et, Vol. 9, 1992

 Dynamic Response of Soils..

A lleissman and Hart (1961)

24
Hardin ( 1965)
0 Drnevich, Hall and Richart (1966)
r.'

r:J Matsushita, Kishido and Kyo (1967)

Silver and Seed (1969)
Donovan (1969)
20
,. Hardin and Drnevich (1970)
...c ~ Kishido and Takano (1970)
Cl>
...0
~ 16
/
/
.......0 /
,,, /
/

" 12 -
/

...."'a.c
0
~

8 --

0
10-4 10- 2 10- 1

Shear Strain-percent

Fis. 9 Dampins ratios for IUld [SJ

40

Taylor and Menzies (1963)

ID Taylor and Huqhes (1965) ----
35

IdriH (1966)
......
~ Krizek and Franklin (1967
30 Thiers and Seed (1968)
~

.
Kovacs (1968)
...
~ 25
A DonO,Yan (1969) ,
~
/
Taylor and Bacchus (1969)
I
i
.. Hardin and Drnevich (1970
-e /
A
/

.......~
~
20

...."'c 15
I
+
10

-------- -- ~--.,,.

Shear Strain - percent

Fis. 10 Dampin1 ratios for saturated clays [SJ

Journal of &Et, Vol. 9, 1992
10
Alemayehu Tefe"a
Consider a hom0aeneous soil layer indicated in
Pia. 11 underlain by a rock or rocklike material
which is subjected to a seismic motion u1(t). The
equation of m0tion is aiven. by the followina
.,_ = P. ~
OHi
cin:uJar _... &.queacy of

reW\ooship [7].
rl' mOde of vibtatfon

P<Y> th~> au __! [G()'>

ar2 at a, "]
a,
H = total thickness of the layer.

.-- p(y{".:; l (10)

x_ =. -c-
lpc.>,.
damp'mg rab'on.

Where
u(y,t) = relative displacement at a depth y at time
t.
r = gamma function.
b,8 = constant related top by :
ey) viscous damping coefficient at depth y.
p 1 -8+2b=O
G(y) = shear modulus at depth y.
p'+28+2=0
p(y) = mass density at depth y.

For a constant viscous damping p(y)=p and shear

It should be noted that the above solution is
modwtus varying with depth according to a function
obtained only when ps 0.5. Eq. 13 defines the
G(y)=K)', Eq. 10 becomes:
mode shapes of the system during the rl' mode of
vibration whose circular frequen_c y is c.i,. X,.(t), may
Cflu a.. a [ au ]-pu
p-+c---ky'- (11) be computed from Eq. 14 using iteration or numerical
ar2 at Oy Oy I procedures (7).

After having obtained the values of Yn (y) and X..

By letting
(t), the relative displacement u(y,t) of Eq. 12 is
calculated. The relative velocity, the relative
" (y.t) "' E Y,.(y)X.,,,(t)
rtl acceleration and strain at any depth y at any instant of
time can be determined by appropriate differentiation
of Eq. 12. A computer programme bas been
Eq. 11 may be written as follows (7):
developed by Idriss and Seed to solve problems that
are of the type discussed above (7).
Y11(y) = CP/2)6 r(l-b)(y/11)..,.
Lumped-~ Solution for Horizo Layered Soils
J_,, [P,.(y/11)~1 (13)
If the soil deposit is layered, the above equations do
not apply. An alternative method bas been worked
out by Idriss and Seed (7). Here the soil deposit is
represented by a series of layers as shown in Fig 12.
The mass of each layer is lumped at the .up and
-R,. a, (14) bottom of the layer and the masses are interconnected
by shear springs that resist lateral deformations. The
springs represent the stiffness properties of the
material between any two masses. The damping is
where
assumed linearly viscous.
J~ = Bessel function of the first kind of order -b.
For the soil mass lumped at n levels, there wil1 be
n equations of motion which may be represented in a
{J,. = roots of 1~(8,J = 0, n = 1,2,... matrix form as follows [7].

Journal of EAEA, Vol. 9, 1992 . I

 Dynamic Response of Soils 11

du
dy(o,t) =0 Surface

H
L~i - u(y,t)

u(B,t) =0

Base
i.1 (t)
g

Pia. 11 Semi-infinite aoil layer subjected to a horiz.ontal seismic motion at the base, according to Idriss and
Seed (7).

--------------- ----~ ------- ----- -

--------------- -------- ---------

2

H
-------------------------------- i-1 2hi-1

----
Damping factor - - -
c in each mode
1 1 i
- m + -m - ki = f (G) i
2 i-1 2 i
------------ - -- -- ------ -----

------------- - _______... __________

1 1
---- ,..-
-m - k f (G ) n 2h
2 n n n n

u
g

Fis. 12 Lumped-mass representation of layered system (7).

Journal of F.AF.A, Vol. 9, 1992

12 Alemayehu Teferra

[MJ (ii) + [qu + [K] {u} = {R(t)} (15) Wave - Equation Solution
[M] = mass matrix and is diagonal. Its terms are:
diag [M] = (m 1, m,., m3 , mJ The theory considers the response associated with the
vertical propagation of shear waves through linear
Where visco elastic system shown in Fig. 14. The sy'stem
m1 is the mass lumped at level i which is equal to consists of n horiwntal layers extending to infinity in
the horiwntal direction and a half-space ~ its
bottom. Each layer is homogeneous and isoiropic
and is characterized by thickness, h, mass dens'ity, p
Y1-1h1-1Y/a1 i l,3 shear modulus, G and damping factor, f3. '
m, ' , .11
g
The vertical propagation of the shear waves
'Yi total unit weight of the soil between
through the system of Fig. 14 causes only hori~ntal
level i and i + 1.
displacement " = u(x,t) which satisfies the wave
h, half thickness of the segment equation [8]. ..
between level i and level i + 1
g acceleration of gravity. p;p"- = GiP1' =
c' -
Tl - " (16)
0 t2 0 x2 ax2ar
[K] the stiffness matrix. It is tridiagonal
and symmetric. Its components are given
Where
by:
Ku = k1 71 = viscosity and is related to the critical damping
ratio, (3, by relationship w.71 = 2G{J.
Kv k,.1 + /G, for i=j
The solution of Eq. 16 for a harmonic motion with
= -IG, for ; = i - 1 frequency w is:

== -"',fori=j+l
u(x,t) = Eel(pt ~ + Fe -IO:r-..t> (17)
= 0 , otherwise
Where
spring constant between level i and E, Fare constants and
level i+ 1.
a, K = wlv, is the wave number.
=-
2h,
G; shear modulus in segment i The first term of Eq. 17 represents the incident
wave travelling in the negative x direction (upwards)
[ C] viscous damping. and the second term represents the reflected wave
travelling in the positivex direction (downwards).(Fig.
(R(t))= earthquake load vector = col (m., 14).
m,., ...m,J u,.
Eq. 17 is valid for each of the layers in Fig.14.
By solving the series of equations represented by Using a local coordinate system x for each layer, the
Eq. 15 one determines the time histories of displacement at the top and bottom of the lay,er m
accelerations, velocities, displacements, strains and are:
stresses related to each lumped mass. A computer
programme has also been developed for this case in (18)
which non-linearity of the soil is considered. -'

The variable shear moduls is replaced by a. and (19)

G2 (bilinear system) as well as by GED (equivalent
linear system) as indicated in Fig. 13 and the
calculation performed.

Journal of EAEA, Vol. 9, 1992

 Dynamic Response of Soils 13

Shear stress Shear stress

ey

Shear strain Shear strain

2&
y

a) Typical stress-strain b) Bilinear idealization

curve for soil

Shear stress

Shear strain

c) Equivalent linear idealization

Fig. 13 Idealization of stress-strain behavior of eoils

Layer Coordinate Propagation Properties

No. System Direction

ul

1--t-- Gl' 81 ' pl

:<"
x2

m
um+ 1
1-:-f-- Gm' f\, pl h
m

m + 1. xm + 1
-1-+- Gm+ 1, Sm+ 1 pm+ 1 hm + 1

um+ 2

xm+ 2

l u
In
n xn + 1 Gn, 8n 'pn
!"..
Incident wave
Fig. 14 .One dimeiisional system (8) Reflected wave Journal ofEA.Et, Vol. 9, 1992
......
........, . . . . . ...._ lttbltopMl~of
. . ..,. ... _ , be dlteriailll4 by . . . tbe
:

. ~ tbe free IUl'fllce tbe ........... lllUlt be

Min mo. llmce from Eq. 22, with Ta X1 0, ODO
obcliu E1 F1 1hil - . that tbe amplitudet of
er - 0 (1 +2i#) - compla . . . wave die incideat md reflected wav are alway equal at
a free IUl'fllce. BePmina with tbe IUr&co layer and
PNaBq.17. repelted '* of tbe recunioa formulu one would
detenailie tbe amplilude B .wl F of all tbe layen and
a (IUir - , ....,..,,., (21> ultimately _d1&1mbt ...e~ins ICCelen&ion
a~ and main from tbe followins equabOlll (8).

11-=e
~ (rJ) llQ e&- ,,..,.... (22)

Hmce tbe lbelt ..,_ at the bottom of layer m y.!!ii C&-'>-F,-.....,,J (28)
Ar
119 mipecUvely:

~.(;l-o) llQ ~.-F,) ,,.,
......
Since bitweea .ch layer there 1DU1t be a
dilpllcemeDt wl . . continuity' tbe ditpllCelDellt
wl..._.. It tbe bottom layer of m lbould be equal to
tbe dilpl9cemeDt wl IMlr of tbe tap layer of m + 1.
Prom dlia comideratioa, one oblainJ tbe followins
recunioa formulu for tbe amplitudel. E and F+I
of tbe iDcidea.t wl reflected wave in layer m + 1
exprmld in tenDI of tbe amplitudea in layer m (8).
(23) , Tho equationa developed above are va1id for
.-.dy ~ 'banooQic motiobs'.. The theory baa been
extended to ~eiit motion throuah the U8e of
Pou.tier tQmformation. A computer propamme
known u SHAKE baa been developed by a team at
(24) ...._ univermty of California, Berkeley, which is hued
on the conhiiuoua solution of the wave equation
ldapted (or uae with transient motions throush the
Fut Fourier Tranafotmation alaorithm (8). The
nonlinearity of the .8'ieu' modulus and dampina is
ICCOunted for by the uae of equivalent linear soil
properties Which are strain dependent. In order to
emure that the material properties used in the
aaalyai1 are compatible with the atraiaa induced at
different levels in the aoil deposit by the base
excitation, several iteratiooa are conducted uaina
iacreasiaaly improved valuea of material properties

... . !2 (1 +,)I..'. until a strain-compatible solution baa been obtained.

Baaed on studies in the field and in the laboratory,
&eoeral relationships between shear modulus and
t F.(l+)e-. .6 (25)
dampina factors and strain have been developed (Fia
6 to 10) and incorporated in the proaramme.

In the analy~is an elastic half space baa been

"I 1l ILL-
B'.(1 - &,)I -.- usumed formina the bottom layer of the system. In:
reality however, the bottom layer consists of riaid
+ !, (1 +.,..
. 2
.\. ..,.,..,.. (26) rock and this riaid rock may outcrop. This situation
will undoubtedly creat a disparity between the theory
*
and reality. The factors which play role in cre.atina

Journal of ...t&, Vol. 9, 1992


Dynamic Response of Soils
this disparity are discUS8Cd by the authors and their
effects 'is taken into consideration (8).
The pro&raJillllC SHAKE is the one that is widely
used to assess ground respo08e8 induced by
W'thquake exitations.
The pro~ performs the following set of
operations:
1. Read the input motion, find the maximum
acceleration, scale the values up or down, and
compute the predominant period.
2. Read data for the soil deposit and compute the
fundamental period of the deposit.
3. compute the maximum stresses and strains in the
middle of each sublayer and obtain new values
for modillus and damping compatible with a
specified percentage of the maximum strain.
4. Compute new motions at the top of any sublayer
inside the system or outcropping from the
system.
5. Print, plot and punch the motions developed at
the top of any sublayer.
6. Plot Fourier Spectra for the motions.
7. Compute, print and plot responae spectra for
motions.
8. Compute print and plot the amplification function
between any two sublayen.
9. Increase or decrease the time interval without
c:hibigiog the predominant pCriodor duration of
the record.
10. Set a computed motion as a new object motion.
Change the acceleratioo level and predominant
period of the object motion.
11. Cotopute, print and plot the streu or strain time-
history in the middl~ of any sublayer.
Theee operations are performed by exerciaiq the
varioua available optiooa in the proaramme.
15
EMPERICAL METHOD
As indicated earlier, the response spectrum is the
most significant characteristics of an earthquake
motion and plays an impottant role in the assessment
of the influence of local soil conditions. The shape
of the response spectrum is highly influenced by the
soil conditions.
The design spectral shape of a site is obtained by
first determining the normaliz.ed acceleration response
spectrum. The normaliz.ed acceleration response
spectrum is obtained by expressing the ordinate as
aprop<>rtion of the maximum ground acceleration for
the motion for which the spectrum was derived or the
zero period spectral values as shown in Fig. 15.
As an alternative to the analytical approach of
assessine the response of soils to earthquake shakine,
a systematic study was conducted on different
spectral response shapes of various site conditions for
which complete records of actual earthquake were
available, with the objective of establishing a
relationship between spectral shape and local site
conditions. A total of 104 horiwntal eround motion
records out of which: 28 for rock sites, 31 for stiff
soil sites, 30 for deep cohesionless sites and 15 for
sites with soft to medium clay and sand were
statistically analiz.ed (9). The study was limited to
s
spectra determined for 5' dampine.
The analysis revealed that the spectral shapes Were
different for the 4 categories of site conditions. The
mean spectra of different site conditions are
compared in Fig. 16 and the 84-percentile spectra are
compared in Fig. 17. It is desirable that additional
records be included to all categories especially to
"soft to medium clay and sand sites" in order to
increase the releability of the statistical analysis. The
validity of the ~stically deduced curves was,
checked with other spectral curves proposed by
reaearcliers for the four cateaories of site conditions:
A reasonable dearee of a,reemmt was observed for
all case8 (9).
For the ~ cohesionless site, it was felt that if
expoeed to hip intesity of sbakine, the spectra may
be influenced. To check this, the authors conducted
a lep&l'&le analysis by comideriq only tbOse records
(6 records) that had ma.wimnm llCCeleration &reatet
than 0.3g. The memi spectn&m shape for theee aix
reOOrds WU compared with meml spectn&m shape for
all tbe records (f'.ia. l~) . From the fipre it appears
Journal of EA.&, Vol. 9, 1991
16
Alemayehu Teferra

Acceleration Response
0 . ?1--_:::S:.t::.;::e~c~t~rum:=i---r----r----+--~

'
i;::
0
-~
.v 0 . 51--.----1~-'-"--+---l----+----+----i
~
Q)
M
~ 0 . 4 1----'-U-........--~ 5% Damping
o
I(

M
~ 0.3
.v
0
Q)
bl' o . 2 J----J-~--+---"-~--+----+----i

Maximum
0.1 Ground
Acceleration

0.5 1.0 1.5 2.0 2 .5 3.0

Period, s
4 I I
Normalized Acceleration
...._ Response Spectrum

i;::
0 II.
-~ 3

\
i;:: .v
0
-~
11' ~
.v ""
Q)
11' M
V'
\,
""
Q) 0Q)

'~,5% Damping
M 0
GI I(
0
0 "O
I( i;::
2 -
M 0
11' ""
""
.vt!>
::s
\ ..
0 ~
Ul ~
>< 1
) \ ~~ '
...$
~

0
--
0 0. 5 1.0 1.5 2.0 2. 5 3.0
Period, s

.....
Pia. lS Determination of normaliz.ed acceleration re1pOD10 spectnun (1)

Journal of.EC.Et, Vol. 9, 1992

 Dynamic Response of Soils 17

4 I I
I I I I I I I I I l
T l I I I I I I I I I
.._ Total number pf records analyseo: rn~ ~~

Spectra for 5% damping

I I l I I I I I : I

,_
Soft to medium clay and sand-15 records
- J

I /~ I/ J.. Deep cohesionless soils (>250 ft) -30

I"( ~ I I/ -I records
r1 \ ~ hi; I ...~
v \' - - 1"-
/
I/ I
Stiff soil conditions(<150 ft) -31
records ,
\ \ '\ I
...._
...._ ~ -
1
I\ ......,._ i,..-""- v ' IJ\ J

1
' Rock - 28 records
. -- --- -- -- 1--

I\ "\.
"' .._ /
I i-...... I
" I/

J
' 'r-..
" ........
r-.... ....""'
""' ......... /

_,__
- ~ ;;,;:,,
~
r"--. ....._
. J.
/
--
-i-- .....__
r-......_-
~ "'--
,_
...... . . . . . . .__-.-
~ -- ...........

- --
-
0
Q.5 1.0 1. 5 2.0 2.5 3.0
Period-seconds

F~a .. . 16 Averaae ~leration spectra for different site conditions [9]

'
---
4
I I I I

Total number of recorc;ls analysed:

I I I I I I I I I

1( 4
I I I I I I
I I
Spectra for 5% dampin

I""\
/ ,rr l\ I I
--....._
I \ \\ I \ i I
Soft to medium clay and sand -15 -
\~ I \ I
i
records -
r-.. I
/'" l.. \\

"' , 1
~

'_/ :r- Deep cohesionless soils (>250 ft) -

J \ :\'. y I -30 records
I
..
-~ 7
I ~Stiff soil conditions (<150 ft)
-
I '\ \ \
. !\.
~
!
"'r--..._
1'
7
...... 7 ~I
\ 17 , _ -31 records
Rock -28 records

/ ' -,;..
1
" '

V' l ~
-
f \. "' .,,.r-....
~

; ; !./
' '"" "r--...
..... ! 7
...... K.
/"""" ........
......,.,,
r-.... r--...~

. "
r--.... r""o.
"'--
-- 1--.
i--.
;---

....... r-..... -r-

--
. . . r-....
. - r---.. i-.-~
-- ""- --- ~

0
Q o.s 1.0 1.5 2.0 2.5 3.0
Period-seconds

Pia. 17 Perceatilo acceleration-spoetra for different lite coaditioaa (9)

JOIU7IQJ of&&t, Vol..9,' 199')

18.
Alemayehu Tefemi
that the increased lhiakina intemity lead:a to marked
~ in spectral amplification factors for periods
between 1 and 3.S sec. 'Moro data should be
incoproNted into the existina ones to ultimately
propose a modified spectral shape for the
cohesiooless soils.
Once the normalized spectra arc available, it is
possible to draw the de&ian response spectra for
different site conditions provided the maximum
ground accelerations arc known.
Suppoee it is desired to detetmine the acceleration
response spectra for different site conditions at a
distance of 5 miles (8 km) from the source of enerar
release of an earthquake with a mapitude of 6.S.
The maximum iround acceleration for the different
site conditions is found to be as follows:
Rock = 0.44g; stiff soil = 0.4g; deep
cohMioDless soil = 0.3g; and soft to medium clay
and sand = 0.26g. By multiplying the ordinate of
the mean normali:t.ed spectra of Fig.16. by the above
values of maximum acceleration, the design response
spectra arc obtained (Fia.19).
ESTIMATION OF PEAK HORIZONTAL
ACCELEl\ATION
In order to determine the desian response spectra of
a site from the normaliud response spectra, it is
necessary to know the peak horizontal ground
acceleration of the site in question. This may be
achieved through the use of Fig. 20 and 21. The
predominant period for maximum accelerations in
rock for a site located at a certain distance from the
cave fault may be determined from Fig. 22.
Suppose it is desired to estimate the peak
horizontal ifOund acceleration of a site consisting of
stiff soil located at a distance of 5 miles (8 km) from
the source of energy release of an earthquake with a
magnitude of 6.S. From Fig. 20 the peak horizontal
ground acceleration in rock is found to be about
0.44g with a predominant period of about 0.3 sec
'cfrom Fig. 22). From Fig.21, the correspopding
maximum acceleration for the stiff soil Would be
about 0.40g.
Journal of .&f.EA, Vol. 9, 1992
ESTIMATION OF THE MAXIMUM BASE
SHEAR OF BUILDINGS
After determinina the reaponae spectrum for a given
soil condition either by usina the analytical appro.ch
or the empirical approach, one may easily determine
the maximum base shear of a buildina usina the
following approximate formula [1).
SA
Y. W- (29)
- g
Where
W = weiaht of the structure.
s. = spectral acceleration corresponding
to the natural period of the
stiucture.
g = acceleration of gravity.
v_ = maximum base shear.
EARTHQUAKE LOADING ACCORDING TO
ETHIOPIAN SfANDARD CODE OF PRACTICE
After bavina disc)iSsed the different methods that are
available for assessing the dynamic responses of soils
and the influence of site conditions on the character
of the response, it is considered pertinent at this sta1e
to examine the recommendation given by the
Ethiopian Standard Code of Practice regrading
earthquake loading.
The Ethiopian Standard Code of Practice
ESCP1:1983 [101), stipulates that "every structure
shall be designed and constructed to resist minimum
total lateral seismic forces assumed to act non
concurrently in the direction of each of the main axes
of the structures in accordance with the equation.
F,. c. a.. (30)
Where
C, = seismic base shear coefficient.
G,. = equivalent permanent load
I
20
Alernayehu Te.ferra

~ Based on ovailable data

- - - - CltimlltiO:"I

0. 2 f.-------1-----

s :o 20
Horizontal dis(ance from 1ource of nr9Y re!ea1e - miles
1.6 l.2 8 16 32 81.l 160

Horizontal distance froa source of enr9Y release laa

Fig. 20 Average values of maximum accelerations in ~kill

Note: Relationships shown aboye 0.39

are based on extraP<>l~tion of data baH

O. Sl-----+-----+-----+-------+--------,t-::-Ro-c-.k--t-------4

Stitt soil
'
I 0 .414-----4--'-~---l------'--+-----~-1'o---71f'-"--t-----t-----:----1
].,
~
....ti
~ 0 . 3 1 - - - - 4 - - - - - 1 - - - - --;7"---+-_,,,,......__t-----+-------1
u
.c;

]
i 0. 21--~~+-~~--~-7!'--+-~~-1-__.__.--t--__,~-;-~---.,=-1

0.1 0. 2 0.3 0 .4 o.s o..6

Maximum acceleration in rock - 9

Fig. 21 Approximate relationships between maximum accelerationa on rock and other local site conditions [l]

\
Journal o.f EAEA, Vol. 9, 1992

I
\
 19
Dynamic Responst of Soils

4 .
Spectra for S\ damping
Mean . for 6 records where
(a ) k > O. 35 g approx.
max roe

Mean for 30 records where

average value of (a ) k = _O. 19 g
max roe

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Period - second
Fia. 18 Effect of sbakin1 intensity on normaliaed acceleration ~ for deep cobeei9nlea t0il lites (9)

1.6
Spectra for 5% damping
1.4 Stiff s~te condition~--,..--,..,..--~-----~-t--------

~
1.2 ------+--------
I
c:
0
~
+' 1.0
11'
M
.....111 ,,,..---- Soft to medium clays and sand
111 0.8
()
()
~
.....
cU
M
0.6
+'
()
111
0.
Cl) 0.4

0 .2

00 0.5 1.0 1.5 2.0 2.5 3.0

Period - seconds
Fia. 19 Anticipated mean spectra for mapitude 6 1h earthquake at distance of Smiles [9)

Journal o/EAEA., Vol. 9, 1992

 Dynamic Response of Soils 21

M=8

M=7.5

~ 0. 6
c:
.....
~
~
~

0
0 40 80 120 160 200 240 280 320 360
Distance from causative fault-km
0 25 50 75 100 125 150 175 200 225
Distance from causative fault- miles

Fig. 12 Predominant periods for maximum accelerations in rock [1]

2.50 1 - - . . - - -

m 2.00 ~--+--~-...:.:~+-~~~~~~~~-1-~~~~~~~~-+-~~~~~~~~~~--l
M
s
0
Soft to medium clays
t! and sands
.,II B l . ST-l/2 < 2.0 (proposed)
c

:
.&
c
......"'
8 1.00+-~~~~-1-~~-+~~-=~~~~~J.-==:;p...""""":--~~~-==~----==-~~~~~~~

soil conditions

B - 1.2 T- 112 s 2 . 5
@
---------------
Deep cohesionleas or
stiff clay soils

B. 1.5T- 112 s 2. 5

Period, T (sec)

Fig. 23 Desip responses factor accordina to ESCP 1:83

Journal ofEA.EA, Vol. 9, 1992

~ .
s:: IZ I I

~
-
~
.
.i::
....
Q)

....0
1.0
............
'\.
,/
-.........
<::oil Typef.i:\

--
,,L
..........
~
~

Soil Type@
1

-
c
s
R . Z.C
t o'
for moderate
earthquake

..~
IM
Q) ")<:.Soil Type(!). cs F
t
.z.c0 .Ds .Fcs '.for severe
Q) I
I '\. '('
Earthquake
0 ..... ,('
u I
'i...
M
0.5 -
""'"'-....../ .......... ............
~
ctS
~ .......... r-- ......

---- -
!""" .
0
Q)
~ 0.
C/l
...... i::
~ ...."'
(/)
Q)
0
Cl 0 1 2
Period T, (Sec)
a) Design spectral coefficient Rt (BSLJ)

1.2
.

'O
....iii
~
Soil Type 3C\----+-I
_ _ _ _ _ _-..,_ _ _soil Type@
1.0

2 0.8
QJ IM Soil Type 11'
~ 8 \.!,/
k
MU k 0.6
max
~~1-t-+-----+-~-+-~~-==-+-=--rl
~ ~ .
0.4

Q)
(/)
0.2
ctS
~O+-t---+----+----+------1--~+-1 0
0 1 2 0 o.s 1.0 1.5 2.0 2.5 3.0
Period T, (sec) Period (sec)
b) Normalized base shear coefficient to
account for spectral content (NEHRP)

Fig. 24 Influerice of Spectral Content


Dynamic Response of Soils
The seismic base shear coefficient is given by
c. - p y (31)
Where
a = a)
a0 = bedrt>ck acceleration ratio for the site
and depends on the seismic zoning.
I = importance factor.
/3 =0 elastic design response spectrum
factor for bedrock A foundation and
standard damping of S % and
determined from:
/3 =0
(1.2) T', where Tis the period.
S = soil classification and site condition
factor.
'Y = structural system type factor.
Within the framework of the present discussion,
attention is given exclusively to the parameter f3 even
though the seismic zoning map that is currently being
used in design leaves much to be desired. The design
response factor may be expressed by the following
equations for the given three soil types.
(i) For rock and stiff conditions (soil Type 1)
(ii) For deep cohesioniess or stiff clay soils (soil
Type 2) 1
p 1.S(7)-112 ~ 2.S
(iii). For soft to medium clays and sands (soil Type 3)
p = 1.8(7)-112 ~ 2.S
A plot of /1 versus T (Fig. 23) for the three soil
types indicatee a general similarity with the Building
Standard Law of Japan (BSU) and the National
23
Earthquake Hazard Reduction Project (NEHRP) of
the United States of America (11], who have had
many years of experience in asesmic design of
structures.
According to BSU, the coefficient R, is assigned a
constant value of 1.0 for a period up to 0.4 sec and
0.8 sec for~~ Types 1, 2 and 3 respectively (Fig.
24 a). NE~f' on the other band assigns a constant
normali:zed t>;se shear coefficient of 2.0 for a period
up to 0.9 F for soil Type 3 and a constant
normalized b'ase shear coefficient of 2.5. for a period
up to 0.6 sec for soil Type 2 an ~
soil'!XPC 1 (Fi .
The spectral curves of Fig.24 are derived from the
statistically deduced average spectra of different soils
presented in Fig. 16. From F.ig. 16 it is apparent
that the variation of soil Type 3 is different from the
other soil types. Its normalized pP.alc acceleration is
lower than the other soil types. The normalized peak
accelerations of the two soil lypes drop quickly in
comparison with soil Type 3. In fact the normalized
peak acceleration of soil Type 3 remains constant till
about 1 sec.
Hence the proposal of ESCP1 :83 of assigning an
equal maxi.m um /3-value for all the soil types does not
conform with the real situation. It would therefore
be appropriate to assign a different f3- value for soil
Type 3 in conformity with NRHRP. It is
recommended to use f3 = 2 up to a period of 0.9 sec
for this soil as indicated in Fig. 23. With this
modification the ESCP1 :83 recoQUJJendation
regarding earthquake loading is comparable with the
other codes that have been discussed above.
(32)
CONCLUDING REMARKS
From the preceding discussions, it has been possible
to present two approaches (one analytical and one
empirical) for determining the response spectra of a
(34) given site.
Even though the analytical method simplifies the
complex wave patterns that develop during an
earthquake to one-degree of freedom, experience with
(34) this method has demonstrated its usefulness. In fact
the dynamic response analysis highlights details of
ground response characteristics of a given site more
than the empirical method, where these characteru\ics
are defused by the statistical averaging techniques.
Journal of EAEA, Vol. 9, 1992
24 Ale11l{l'Jehu Te.ferra

Due to its simplicity, there is a tendency to use
exclusively the empirical curves rather than
performing the dynamic response analysis. It should
be stated however that before the analytical method
is completely replaced by the empirical curves,
additional earthquake records should be incorporated
in the existing data and new evaluation should be
made. The new spectral curves should of course be
incorporated in the codes. As things stand at present,
there is still a need to use both methods in order to
arrive at reliable design spectra.
Levington, Kentucky, July 1970.
7. Idriss, I.M, Seed, H.B, Seismic Response of
Horiz.ontal Soil Layers. Journal of the Soil
Mechanics and Foundation Division. ASCE,
SM4, 1968.
8. Schnabel, P.B, Lysmer, J, Seed, H.B, SHAKE:
A Computer for Earthquake Response Analysis
of Horiz.ontally Layered Sites. University of
Califo~a, Berkeley, 1972.

In order to estimate the base shear force of a
building, one may either use the spectral acceleration
derived from the analytical method.(e.g SHAKE) and
apply Eq. 29, or directly use the appropriate Curves
of Fig. 23 or the equivalent equations described in
ESCP1:83. The result thus obtained should be
compared.
REFERENCES
1. Seed, H.B , Idriss, l.M, Ground Motions and
Liquifaction during Earthquakes.
Earthquake Engineering Research Institute,
Berkeley, California, 1982.
2. Seed, H.B, Site Effects in Earthquake Resistant
Design. Proceedings, First Venezuelan
Conference on Seismology and Earthquake
Engineering, October, 1974.
9. Seed, H.B, Ugas, C, Lysmer, J, Site-Dependent
Spectra for Earthquake Resistant Design.
Report No. EERC 74-12, Earthquake
Engineering Research Center, University of
California, Berkeley, November, 1974.
10. ESCP1:1983, Ethiopia Standard tode of
Practice for Loading.
Issued by Ministry of Construction, Addis Ababa,
Ethiopia. 1983.
11. Ishiyamo, Y, Hans Rainer, J, Comparison of
Seismic Provisions of 1985 NBC of the
USA. S1h Canadian conference.
Earthquake Engineering, Ottawa,
1987.

3. Idriss, I.M, Characteristics of Earthquake

Ground Motion. Woodword-Clyde Consultants.
San Francisco, California, 1974.

4. Hudson, D.E, Response Spectrum Techniques

in Engineering Seismology. Proceedings, First
World Conference on Earthquake Engineering.
Berkeley, California, 1956.

5. Seed, H.B, Idriss, I.M, Soil Moduli and

Damping Factors for Dynamic Response-
Analysis. Report No.EERC 70-10, Earthquake
Engineering Research Center. University of
California, Berkeley, December, 1970.

6. Hardin, B.O, Dmevich, V.P, ShearModulusand

Damping in Soils.
I. Measurement and Parameter Effects.
II. Design Equations and Curves.
Technical Reports UKY 27-20-CE 2 and 3
College of Engineering, University of Kentucky,

Journal o.f EAEA, Vol. 9, 1992