Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 25

CHAPTER 1

1. INTRODUCTION
The occurrence of earthquakes & the damages related to such events are very hard to control.
Earthquakes cannot be prevented or predicted. If the damages can be determined well in
advance, precautionary measures can be taken at the time of designing a structure. For a structure
the stability of the foundation is as important as the stability of the structural elements. In the
geotechnical aspect, a designer is concerned about the level of ground shaking which
earthquakes induce at a particular site. The bedrock motion or shaking is measured at different
seismic stations in the form of the time history of the motion. An earthquake produces complex
type of motion covering a wider range of frequencies. Some of these frequencies may cause
excessive shaking due to the resonance condition. To know the frequency content of the
earthquake generated ground motion, one can generate Fourier spectra while the response spectra
will present the response corresponding to various frequencies. Whereas Fourier spectra describe
the motion in frequency domain, response spectra take into account the natural frequency of the
system. Based on the collaborative evidences during the past few decades it is now an
established fact that local soil characteristics plays major part in controlling the earthquake
induced damage. The 1989 Loma Prieta earthquake, 2001 Bhuj earthquake, 2011 Sendai
earthquake and the 2015 Nepal earthquake are classical examples where massive damages at
greater distance from the epicenter were recorded mainly due to the local site effects. When a
motion in the form of wave travels up from the bedrock to the surface level it amplifies. A
foundation designer is always keen to know this amplified ground motion at the surface. Suitable
site specific response analysis can be conducted to know these magnitudes along with the
frequency content. Site response analysis is done to evaluate the design response spectra for a
site with respect to the local geological conditions. The magnitude of input bedrock motion can
be quantified based on the detailed seismic hazard analysis. Finally the amplification factors &
the spectral acceleration are determined for each subsoil layer. As found by the earlier published
research work, it was found that the design spectral acceleration given according to Indian
standard codal provision are most of the times found lesser when it comes to regional micro &
macro level site response analysis studies. The response spectra found out from site response
analysis are not exactly smooth the way IS 1893: 2002 described it. Those consist of irregular
1

spikes and valleys. Thus there is a need to understand the smooth response spectrum from
regional studies for the Indian soils. This will help in mitigating the building damages by
considering the level of shaking for the future earthquake in the design values.

CHAPTER 2

2.

EARTHQUAKE DAMAGES:

Earthquake damages are catastrophic. Such damages can never be predicted in advance. From
ancient times mankind is suffering from the perils of earthquakes. The effects of an earthquake
are strongest in a broad zone surrounding the epicenter. Surface ground cracking associated with
faults that reach the surface often occurs, with horizontal and vertical displacements of several
yard[3]is common. The extent of earthquake vibration and subsequent damage to a region is partly
depends on characteristics of the ground. For example, earthquake vibrations last longer and are
of greater wave amplitudes in unconsolidated surface material, such as poorly compacted fill or
river deposits

[3]

bedrock areas receive fewer effects. The worst damage occurs in densely

populated urban areas where structures are not built to withstand intense shaking. In such areas,
waves can produce destructive vibrations in buildings and break water and gas lines, starting
uncontrollable fires. These are the induced effects of earthquakes.
Damage and loss of life sustained during an earthquake result from falling structures and flying
glass and objects. Flexible structures built on bedrock are generally more resistant to earthquake
damage than rigid structures built on loose soil. In certain areas, an earthquake can trigger
mudslides, which slip down mountain slopes and can bury habitations below. A submarine
earthquake can cause a tsunami, a series of damaging waves that ripple outward from the
earthquake epicenter and inundate coastal cities[3].
On average about 1,000 earthquakes with intensities of 5.0 or greater are recorded each year
across the globe

[1]

. Great earthquakes (magnitude 8.0 or higher) occur once a year, major

earthquakes (magnitude 7.07.9) occur 18 times a year, strong earthquakes (magnitude 6.06.9)
10 times a month, and moderate earthquakes (magnitude 5.05.9) more than twice a day.
Because most of these occur under the ocean or in under populated areas, such earthquakes pass
unnoticed by all but seismologists. Moderate to strong earthquakes can cause more significant
destruction if they occur closer to the earth's surface. Notable earthquakes occurred include
Lisbon, Portugal (1755); New Madrid, Mo. (1811 and 1812); Charleston, S.C. (1886); Assam,
India (1897 and 1950); San Francisco (1906); Messina, Italy (1908); Gansu, China (1920);
Tokyo, Japan (1923); Chile (1960); Iran (1962); S Alaska (1964); Managua, Nicaragua (1972);
Guatemala (1976); Hebei, China (1976); Mexico (1985); Armenia (1988); Luzon, Philippines
3

(1990); N Japan (1993); Kobe, Japan (1995); Izmit, Turkey (1999); central Taiwan (1999);
Oaxaca state, Mexico (1999); Bam, Iran (2003); NW Sumatra, Indonesia (2004); Sichuan, China
(2008); S Haiti (2010); Chile (2010); South Island, New Zealand (2010, 2011); and NE Japan
(2011). The 1755 Lisbon, 1960 Chile, Alaska, 2004 Sumatra and 2011 NE Japan earthquakes
were accompanied by significant tsunamis (data collected according to infoplease.com [1])
Twelve of the twenty largest earthquakes in the United States had occurred in Alaska itself. Most
of the largest earthquakes in United States have occurred in California or elsewhere along the
Pacific Coast, but the three New Madrid earthquakes (181112) also were among the largest
continental events, as was the Charleston, S.C., earthquake (1886). On Good Friday 1964, one of
the most severe North American earthquakes ever recorded struck near Anchorage, Alaska,
measuring 8.4 to 8.6 in magnitude. Besides elevating some 70,000 sq mi (181,300 sq km) of land
and devastating several cities, it generated a tsunami that caused damage as far south as
California. Other recent earthquakes that have affected the United States include the Feb., 1971,
movement of the San Fernando fault near Los Angeles. It rocked the area for 10 sec, thrust parts
of mountains 8 ft (2.4 m) upward, killed 64 persons, and with an economic loss of approximately
$500 million[1]. In 1989, the Loma Prieta earthquake above Santa Cruz shook for 15 seconds at a
magnitude of 7.1, killed 67 people, and toppled buildings and bridges. In Jan., 1994, an
earthquake measuring 6.6 with its epicenter in northern part ofLos Angeles caused major damage
to the city's infrastructure and left thousands habitants homeless[1].
In India also there are numerous reported earthquakes which took massive casualties. These
include Indian Ocean EQ, 2004(Mw 9.3, 283,106 deaths); Kashmir EQ,2005(Mw 7.6, 130,000
deaths); Bihar EQ, 1934(Mw 8.7, 30000 deaths); Bhuj EQ, 2001(Mw 7.7, 20000 deaths); Assam
EQ,1950(Mw 8.6, 1526 deaths) are few of them. Damages during above events are shown in Fig
2.1 below.

Fig 2.1 (clockwise from top left) : Assam EQ(1897), 2001 Bhuj EQ, 2004 Indian Ocean EQ,1934
Bihar-Nepal EQ [2]
The magnitude of earthquake shaking is more in the epicentral region and decreases with the
distance away from the fault. In a single earthquake, however, the shaking at one site can easily
be 10 times stronger than at another site, even when their distance from the ruptured fault is the
same

[3]

. The ground shaking at a site is a function of magnitude the earthquake, distance from

the fault & local geological conditions. Once all these three parameters are known, the level of
shaking at a particular site during future earthquake can be predicted accurately. To be prepared
for the earthquake, the level of ground shaking & its frequency of occurrence should be
determined with accuracy. Most earthquake fatalities and damage occur when buildings and
other structures fail during violent shaking caused by seismic waves, which travel through the
ground from the rupturing fault like the ripples from a pebble dropped into a pond. The intensity
of shaking depends on the quake magnitude (size of the pebble) and distance from the fault
(ripples get smaller as they radiate outward [3]. However, the radiation of seismic waves is much
more complex than the steady progression of circular ripples in a pond. One reason is that the
Earth's crust is not homogeneous like water, but rather a complex mixture of rocks and sediments
of varying types that respond to shaking in different ways. Seismic waves travel faster in hard
rocks than soft rocks or sediments. As the wave travels from the harder stratum to softer stratum,
5

its amplitude increases to carry the same amount of energy. Thus shaking tends to be stronger at
sites with softer surface layers, where seismic waves move slower. Another reason for amplified
motion is that depth of sediment. As the sediment depth increases the amplitude also increases.

CHAPTER 3
3. EFFECTS OF LOCAL SITE CONDITIONS ON GROUND MOTIONS:
Local site conditions can profoundly influence all of the important characteristics- amplitude,
frequency content & duration-of earthquake generated ground motion. The extent of their
influence depends on the geometry, properties of the subsurface materials, on site topography
and on the characteristics of input motions. The nature of local site effects can be illustrated in
several ways: by simple, theoretical ground response analyses, by measurements of actual
surface & subsurface motions at the same site & by measurements of ground surface motions
from sites with different subsurface conditions [4].
Evidence of importance of local site conditions can be gained by comparing ground surface
motions measured at different sites. For example[6], recordings of ground motions at several
locations in San Francisco were made during a nearby M=5.3 earthquake in 1957. Variations in
ground motion, expressed in terms of peak horizontal acceleration & response spectra, are shown
with variation in soil conditions along a 4 mile section through the city in Fig 3.1. Ground
motions at rock outcrops (Market & Guerrero, Mason & Pine, Harrison & Main) were quite
similar but the amplitudes & frequency contents of the motions at sites underlain by thick soil
deposits were markedly different (according to Geotechnical Earthquake Engineering text
book[6]). Similar effects have been observed in many other earthquakes. From the stand point of
local site effects, two most significant earthquakes were 1985 Mexico earthquake& the Loma
Prieta earthquake . In the next paragraphs the case studies of those two earthquakes from
Geotechnical Earthquake Engineering text book [6] are taken and discussed.
3.1 MEXICO CITY, 1985
The September 19, 1985 Michocan (Ms 8.1) earthquake caused only moderate damage in the
vicinity of its epicenter (near Pacific coast of Mexico) but caused extensive damage some 350
km away in Mexico City. Studies of ground motions recorded at different sites in Mexico City
illustrated the significant relationship between local soil conditions & damaging ground motions
& led to important advances in understanding the cyclic response of plastic clays.
For seismic zonation purpose Mexico City was often divided into three zones with
different subsurface conditions[6]. Shallow, compact deposits of mostly granular soil, basalt, or
volcanic tuff are found in Foothill zone, located west of downtown. In the Lake zone, thick
7

deposits of very soft soils formed from the pluviation of airborne silt, clay and ash from nearby
volcanoes through the water of ancient lake Texcoco extend to considerable depths, as shown by
the contours of Fig 3.1.1. Ground water is generally found at a depth of about 2 m over most of
the Lake Zone. Between Foothill Zone & Lake Zone lies the Transition Zone, where the soft soil
deposits are thin & interspersed with alluvial deposits. Prior to 1985, a number of strong motion
instruments had been deployed in the Mexico City.Shown in Fig 3.1.1 are the locations of those
at the Universidad Nacional Autonoma de Mexico (UNAM) & the Secretary of Communications
& Transportation (SCT) site.The UNAM site was located in the Foot Hill zone on 3 to 5 m (1016 ft) of basaltic rock underlain by softer strata of unknown thickness. The SCT site was located
on the soft soils of Lake Zone.

Fig 3.1Variation in the ground motion [6]

Although the Michoacan earthquake was quite large, its great distance from Mexico City
produced accelerations at the UNAM (rock) site of only 0.03g to 0.04g. In the Lake Zone
however, peak accelerations at the SCT sites were upto five times greater than that of the
UNAM. The frequency contents of the SCT motions were also much different than that of the
UNAM motions; the predominant period was about 2 sec at the SCT site [6]. Strong levels of
shaking persisted over a very long duration at the SCT site. Figure 3.1.2 illustrates the
pronounced effects of the Lake Zone soils. At periods of approximately 2 sec, spectral
acceleration at the SCT site were 10 times higher than UNAM site
damage was observed at the Lake Zone than the other two zones.

Fig 3.1.1 Geology of the Mexico City area [6]

[6]

.Consequently greater

Fig 3.1.2: Variation in ground motion [6]

3.2 SAN FRANCISCO BAY AREA, 1989


On October 19,1989, a Ms =7.1 earthquake occurred near Mt Loma Prieta located about 100 km
south of San Francisco & Oakland, California. The Loma Prieta earthquake produced MMI VIII
shaking in the epicentral region, but intensities were actually higher [6]MMI IX, in portions of San
Francisco & Oakland. The fact that earthquake caused extensive damage in certain areas, &
relatively little damage in others, suggested that local site effects were important.
The San Francisco Bay basin is largely filled with alluvial deposits of clays & silty to sandy
clays with some layers of sandy & gravelly soils. The deeper deposits are over consolidated by
historical glacial sea level drawdown, but the upper unit is deposited after the last drawdown
episode. This material, known as San Francisco Bay Mud, is normally consolidated silty clay. It
10

is highly compressible & its strength grades from soft near the surface to medium stiff at depth.
For purposes of seismic zonation, the Bay area can be divided into three zones shown in fig
3.2.1. The San Francisco Bay Mud is generally found at the margins of the Bay, where its
thickness varies to several tens of feet.
Both the epicentral region & the San Francisco Bay area were well instrumented with
seismographs & accelerometers. Peak horizontal accelerations were recorded at the locations
shown in fig 5. These accelerations were high near the epicenter but attenuated with distance
from source. The attenuation, however, occurs much more rapidly for sites in the rock/ shallow
residual soil zone than in the alluvium or Bay Mud zones (fig 3.2.1).
The response of two instruments, those located at Yerba Buena Island & Treasure Island in the
middle of San Francisco Bay, is particularly instructive. Yerba Buena Island is a rock outcrop &
Treasure Island is a 400 acre man made hydraulic fill placed partially on the Yerba Buena shoals,
a sandbar located immediately northwest of Yerba Buena Islands. Treasure Island is underlain by
a variable thickness of San Francisco Bay Mud; the Treasure Island seismograph was under lain
by 45 ft of loose sandy soil (hydraulic fill & natural soils) over 55 ft of San Francisco Bay Mud.
The Yerba Buena seismograph was directly located on rock. Though [6]the Yerba Buena Island &
Treasure Island instruments were located virtually the same distance from the source, they
recorded dramatically different ground surface motions (fig 3.2.2). Peak accelerations at Yerba
Buena Island were 0.06g in the E-W direction & 0.03g in the N-S direction; the corresponding
values of Treasure Island were 0.16g & 0.11g. Response spectra for the two sites are shown in
fig 3.2.2. Clearly, the presence of soft soils at the Treasure Island site caused significant
amplification of bed rock motion.
Amplification of ground motion by soft soil deposits in other areas contributed significantly to
the damage in other parts of San Francisco Bay area. The northern portion of the I-880 Cypress
Viaduct that collapsed in the earthquake was underlain by San Francisco Bay Mud; the southern
part that remained standing was not.

11

Fig 3.2.1 Geology of the San Francisco bay area [6]

12

Fig 3.2.2 Variation in ground motion [6]

CHAPTER 4
SITE RESPONSE ANALYSIS & GUIDELINES
One of the most important & most commonly encountered problems in geotechnical earthquake
engineering is the evaluation of site response. Site response analyses are used to predict ground
surface motions for development of design response spectra to evaluate dynamic stress & strains
for evaluation liquefaction hazards and to determine the earthquake induced forces that can lead
to instability of earth & earth retaining structures. Site response can be conducted by three
methods- linear approach, equivalent linear approach & nonlinear approach. As preceding
articles have described the importance of local subsoil, site response analysis should be done to
quantify the local site effect. In linear approach first a suitable bed rock motion is selected
(acceleration time history). Then Fast Fourier Transformation (FFT) is done to get the motion in
frequency domain. A transfer function corresponding to each frequency is computed and
multiplied with the bed rock motion to get the amplified motion at the surface level in the form
of Fourier series. Then an Inverse Fourier Transformation is done to get the surface motion in
time domain. In Equivalent Linear approach first some values of shear modulus (G) and damping
ratio () are assumed for each layer (for constant strain level). Using these G and values site
13

response is conducted, and at each level (stratification) the corresponding shear strain values are
calculated. Suitable G & values chosen from the modulus reduction curve & damping curve.
For next iteration these new found G & values are used. This way the iteration is continued
until the difference in strain level falls below some acceptable percentage. Though the equivalent
linear model represents non-linear behavior of soil in some extent, still it is an approximation of
the exact situation (during earthquake G, values constantly change with strain in Equivalents
linear method those values are assumed to remain constant throughout the earthquake duration).
In non-linear method the equation of motion is numerically integrated using finite difference
method in time domain. Any constitutive stress strain models can be used. At the beginning of
each time step G & are considered from the model and those values are used in that time step.
In this way the non-linear behavior can be followed.
The concept of response spectra was first incorporated into the United States building codes in
the late 1950s by the Structural Engineers Association of California. Over the decades, response
spectra have been playing an increasing role in the development of earthquake design criteria.
Response spectra provide a very handy tool for engineers to quantify the demands of earthquake
ground motion on the capacity of buildings to resist earthquakes. To know how to approach a site
specific study here a work by Abhishek et al [4]is discussed.
In construction industry, many times the client wants the designer to conduct project specific
studies. Further, the outputs should be used for foundation design purposes rather considering
those given in Standard codes. Site specific studies are particularly necessary for important
projects such as nuclear power plant or dam.
The study area belonged to seismic zone IV as per IS 1893:2002. Past studies suggested repeated
moderate to severe damages in the study area due to distant earthquakes either in the Himalayan
seismic belt or due to regional earthquake from other nearby sources. In order to understand the
subsoil lithology, 41 boreholes were drilled upto 30 m depth. All the boreholes were drilled with
a diameter of 150 mm and N-SPT values were measured regularly at 1.5 m interval. A typical
borelog obtained from the field studies is shown in fig 4.1. In this work, the site response of a
typical construction site was attempted to understand. Equivalent linear approach was considered
in the work using SHAKE 2000.
14

Recorded ground motions are available in India only after 1986 & till then no significant
earthquake has occurred. In the absence of recorded data global database from PEER (Pacific
Earthquake Engineering Research) were collected. Total 30 ground motions were selected.
Though those ground motions were from different areas (different from the actual site), the
region specific ground motion characteristics merely affected the outcome as 30 different
motions were considered and the amplification factor obtained for each of them were normalized
with corresponding bed rock motion. Outputs in the form of acceleration time history, stressstrain time histories at selected layers were obtained. Also variation in PGA with respect to depth
was obtained. Further the surface PGA obtained from output files were used for determining
amplification factors at different levels. Figure 4.2 shows the variations in amplification factors
obtained from the outputs.

15

Fig 4.1 Typical borehole data[4]

Table 4.1 Properties of the ground


motion [4]

It was observed that higher values of amplification factors were corresponding to lower bed rock
PGA values. Similarly no or minimal amplifications were observed in case of bed rock PGA
exceeding 0.52g. Another useful observation was that in the case of sandy soil having SPT-N
value greater than 50 beyond 25 m depth, amplification factor was low. Similarly for low to
medium compressibility clays having SPT-N value greater than 70 also showed minimal
amplification. These observations are useful for restricting the borehole depth to be drilled and
minimizing the cost. National Disaster Management Authority (NDMA), Govt of India
developed the probabilistic seismic hazard maps for entire country. According to NDMA bed
rock PGA for the site considering 10% probability of exceedence in 50 years is 0.08g. Three post
site response analysis observations were made. For
small bed rock PGAs the amplification values were
higher but the surface level PGAs obtained from
those PGAs were unable to trigger liquefaction. The
low amplification values corresponding to the
higher level bed rock PGAs were also unable to
trigger liquefaction. From designers point of view
the level at which foundation is to be seated was
considered for CRR calculations. Assuming factor
of safety 1 to liquefaction the corresponding CSR
value and PGA were calculated. So any surface
PGA value which is less than the calculated one
was eliminated.

16

Fig 4.2 Amplification factors [4]

Fig 4.3 Modified amplification factors[4]

17

Fig 4.4 Frequency distribution of amplification [4]


The modified amplification chart were prepared (fig 4.3). So the previously calculated
amplification factors range of 1-7.4 was reduced to 1.8-4.1. Further frequency distribution of the
amplification factors were done considering all bore hole data and a frequency distribution plot
was prepared (fig 4.4). It has been seen that frequency corresponding to amplification factor 2.5
was much higher than the other amplification factors. Keeping economy considerations in mind a
value of amplification factor 2.5 was proposed which rendered surface PGA as 0.20g which is
slightly less than the value proposed by IS 1893:2002 for zone IV (PGA 0.24).

CHAPTER 5
SITE SPECIFIC DESIGN RESPONSE SPECTRA
Indian standard code of provisions IS 1893:2002 has given design response spectra for three
types of soils Rock or Hard soil, Medium soil, Soft soil depending on the SPT-N value. Response
spectra obtained from site response study comprises of lots of irregular sharp spikes and valleys,
and those values of spectral accelerations are higher than the codal provisions. This is due to the
fact that in Indian standard code the local site effects were not considered. If we can smoothen
the response spectra according to some guidelines followed in the other codes which consider the
local site effects, a site specific design response spectra can be formed and the same can be used
18

for similar site conditions later on. In this report a procedure followed by Sigmund A. Freeman in
RESPONSE SPECTRA AS A USEFUL DESIGN AND ANALYSIS TOOL FOR
PRACTICING STRUCTURAL ENGINEERS has been discussed.
Generally response spectra are plot of spectral acceleration vs time period but spectral velocity
and spectral displacements also can be plotted. Those spectral parameters are interrelated to each
other by Duhamel Integral.
Sv=Sd
Sa=2Sd
Where Sd, Sv, Sa are pseudo spectral displacement, velocity & acceleration respectively.
These relationships hold good as we are concerned about the maximum values of spectral
parameters in Response spectra. Response spectra can be plotted in variety of formats. A format
commonly used earlier was Tripartite logarithmic plot, where spectral velocity is plotted in
vertical axis and time period in horizontal axis. Spectral displacements and spectral accelerations
can be plotted on two diagonal axes (fig 5.1). Some other formats also have been used.

Fig 5.1 Response spectrum in Tripartite format[5]

19

Fig5.3 Response spectrum in ADRS


Format [5]
Fig 5.2 Response Spectra commonly used

[6]

20

Fig 5.2 represents the format we get from site response analysis. Fig 5.3 shows the ADRS
(Acceleration-Displacement Response Spectrum) format. ADRS format is a tripartite format in a
rotated linear coordinate system. The response spectra found from site response consist of sharp
peaks & valleys, and sensitive to slight change of frequency. This fact leads to the construction
of smooth response spectra which can be used for future ground motion prediction.

Fig 5.4 Smoothening of Response spectrum in Tripartite format UBC 1997 [5]

Fig 5.5 Smoothening of Response spectrum in ADRS format UBC 1997 [5]

21

From fig 5.2 it is clear that spectral acceleration is constant when the time period is small,
spectral velocity is constant in mid time period range & spectral displacement is constant at
higher time period. If we assume some constant acceleration at higher frequency, constant
velocity at mid frequency range and constant displacement at lower frequency and plot the same
in Tripartite plot or ADRS plot. Fig 5.4& fig 5.5 show the respective Tripartite & ADRS format
of design response spectra for zone 4 soil type C (1997 Uniform Building Code). The PGA is
0.4g. The constant Sa is 2.5 times PGA (1.0g), constant velocity is based on Sa at 1 sec i.e. 1.4
times PGA (0.56g) this translates constant velocity 87cm/sec. assuming cutoff period of 4 sec the
constant displacement becomes 56 cm.
If somebody desires to construct a smooth response spectrum from irregular one, Fig 5.6
illustrate a simple procedure. The peak acceleration is first identified & a horizontal line is drawn
(constant acceleration). Similarly a vertical line can be drawn for constant displacement
identifying the maximum displacement. Then moving out along the radial lines from the origin
we locate the maximum velocity value. Connecting the lines forms a maximum smooth
spectrum. Similar procedures can be adopted for constructing minimum smooth response
spectrum. Now an average of these two spectra can give a reasonable design response spectrum.

Fig 5.6 Smoothening of response spectrum in ADRS format [5]


22

Fig 5.7 Smoothening of response spectrum in Tripartite format [5]

23

CHAPTER 6
CONCLUSIONS
Indian Standard Code of provision has divided the country into 4 different seismic zones.
According to the seismic zonation different seismic coefficients have been assigned to each zone.
Within a seismic zone there can be large variation in ground motion parameters for different
locations. For many years the shapes of all response spectra were for a give class of soil assumed
to be identical. Later it was understood that site characteristics and magnitude have important
role to play for determining maximum site response. So it is important to find out site specific
design response spectrum which can give the local amplification values. After getting response
spectrum for a site the next important step is to account for the frequency dependent shape
variations of the response spectrum. Those can be done by assuming some constant values of
spectral parameters. Another method is to use regression analysis to develop predictive
relationships for spectral ordinates at various oscillator periods. This method gives smooth
response with respect to the soil conditions they were developed for.

24

References
1. 2010-2015,
Pearson
Education,
Major
earthquakes
<http://www.infoplease.com/encyclopedia/science/earthquake-majorearthquakes.html> (seen on 15.04.2015)
2. 2015 HubPages Inc. and respective owners, Biggest earthquakes in India,
<http://kannanwrites.hubpages.com/hub/Biggest-Earthquakes-in-India> (last seen
on 15.04.2015)
3. 2015 Southern California Earthquake center, Earthquake shaking accounting for
local site effects <http://www.scec.org/phase3/overview.html> (last seen on
15.04.2015)
4. Abhishek kumar, Kumaran M., Vetriselvan, A. Global data based Site response
analysis and output filtering for liquefactionassessment of shallow region in
India, IGC-2014, December 18-20, 2014, Kakinada, India
5. Sigmund A. freeman. response Spectra as useful design and analysis tool for
practicing structural engineers, ISET Journal of Earthquake Technology, paper
No. 475,vol. 44, No. 1, March 2007,pp. 25-37
6. Steven L. Kramer, Geotechnical Earthquake Engineering, Local site effects and
design ground motions, (pg 312-317), Dorling Kindersley (India) Pvt ltd

25

You might also like