09-10-2014, 04:21 PM
Earthquakes cause destruction to life and property in catastrophic proportions.
As a result, the constant aim of earthquake engineering has been to design
structures that would suffer minimum or no damage in the event of an
earthquake. In the design process, the most difficult and crucial problem is how
best to specify the design seismic load. This problem assumes great importance
particularly in the aseismic design of sensitive structures like nuclear power
plants and emergency structures like hospitals and high dams. Here the engineer
is confronted with a serious situation arising out of the randomness and
uncertainty of the earthquake motion yet to occur. The currently available strong
motion descriptors can be classified as either direct or indirect. For example,
magnitude, maximum ground acceleration, duration, etc. are direct descriptions.
Response spectra, Modified Mercalli Intensity (MMI) are indirect descriptors.
Neither of the approaches is adequate by itself as it cannot fully reflect all the
details associated with a shock. In either case, due to the inherent uncertainty,
the methodology has to be statistical. With this in background, the present thesis
aims at taking a fresh look at both the direct and indirect descriptors to evolve
alternate approaches for studying and interpreting strong motion earthquakes.
The presentation of the thesis is as follows: In Chapter I, a brief review of the
direct and indirect methods of specifying and describing earthquakes followed
by a discussion is presented. An attempt is made to bring out the limitations of
the existing descriptors so as to facilitate the development of new approaches
proposed in the succeeding chapters.
In Chapter II, the question of how to optimally combine the various descriptors
like magnitude, a maximum acceleration, frequency content, duration, etc. is
considered. For this purpose, a database of 92 strong motion earthquake records
is selected. The method of multivariate statistics is used to arrive at a
classification of the earthquakes. A new risk rating scale is also proposed. It is
demonstrated that the risk rating is closely related to MMI.
The most common indirect method of specifying earthquakes is the provision of
damped spectra in codes of practice. Whether these spectra are internally
consistent so that they could be derived from each other, is studied in Chapter
III. A linear regression analysis is carried out on paired spectra of 66 actual
earthquake records. The regression results obtained are sued to define internal
consistency in a statistical manner. With this, the internal consistency of the
spectra of United States Regulatory Commission, Canadian Standard Association
and Indian Standard 1893 (Fourth revision draft) is verified. It is shown that the
spectra of the draft revision of IS-1893 are not internally consistent.
A major limitation of the classical response spectra is that they are obtained as
responses of damped linear system. An alternate frequency domain description
is the Fourier Spectra. The sine and cosine Fourier Spectra completely define any
given earthquake record. In Chapter IV, the absolute values of 30 pairs of such
spectra are normalized and statistically processed to arrive at their means and
standard deviations. Thus a smoothened pair of these spectra could be sufficient
to specify an earthquake at a given site. The relation of the two proposed spectra
to the damped response spectra is also studied. It is further verified that time
histories generated to be compatible with the new spectra do indeed have
properties expected of actual earthquake records.
The time spent above given levels of ground acceleration is a characteristic
feature of the accelerograms. Also, an important information required many
times, particularly for nonlinear systems would be the time spent by actual or
equivalent linear systems above specified response levels. These informations
called here pulse characteristic are studied in Chapter V. The importance of the
time spent by the accelerograms above their 20% level is examined. The time
spent results for response acceleration of some linear single degree systems of
different dampings are also presented.
The thesis concludes in Chapter IV with a summary and a set of conclusions.
The thesis ends with Chapter IV wherein a brief summary of the work with a set
of conclusions is presented followed by a few suggestions for further research.
A paper published based on a part of the present work is enclosed in the
Appendix.