| Summary: | 碩士 === 國立成功大學 === 土木工程研究所 === 86 === In recent years, as we recognize a very pronounceed increase
about earthquakes, shear walls are often used in the tall
building to resist the lateral force from the earthquake.
However, to obtain an accurate analysis under the rigid-floor
assumption, the error caused by this method should be first
evaluated. Therefore, the main purpose of this study is to
quantitative investigate the differencebetween the flexible-
floor and rigid-floor analyses of building with shear walls.
Thus, using a number of numerical analyses to obtain an
approximate error formula statistically is alternative. In this
study, the finite element methodwas used to analyze both one-
axis-symmetric-shape and L-shape buildings with unsymmetric-
thickness shear walls and floors. In order to generate more
accurateerror formula between rigid-floor and flexible-floor of
analyses for buildings with shear walls, 1500 pairs of buildings
with both one-axis-symmetric-shape and L-shape buildings with
unsymmetric-thickness shear walls and floors were used
inresponse-spectrum analyses. A dynamic finite element program
Micro-Sap was usedin this study in order to generate an error
formula. For the building analysis under the rigid-floor
assumption, the master and slave nodes were used in this study.
Each rigid floor contains a master nodewith three degrees of
freedom at the mass center of the floor to control the twoin-
plane translations and one out-of-plane rotation of all the
other nodes calledslaved nodes in this rigid floor. The slaved
nodes include three additional degrees of freedom, two in-plane
rotations and one out=of-plane translation. Thenumerical model
of the master-slave-node algorithm can be performed by using
aconstraint matrix. For building analyses under the flexible-
floor assumption, each node containssix degrees of freedom,
three translations and three rotations. In addition tousing the
3-D beam elements to model the members in the building,
4-nodemembrane elements with the incompatible mode was used to
model the floor slabs.The membrane element was also used to
model the shear walls of buildings under the rigid- or flexible-
floor assumptions. For the equivalent static lateral force
method, the horizontal forces are often applied to the master
nodes of a rigid-floor analysis. However, it isdifficult to add
those horizontal forces to the nodes of a building using the
flexible-floor assumption. To compare the results of rigid- and
flexible-floor analyses, dynamic analysis is probably a better
choice since the earthquake loading can be applied to the
building base without difference for the rigid- and flexible-
floor analyses. Forced dynamic analyses include time-history
andresponse-spectrum analyses. For time-history analysis, it is
not easy to compare the complex analysis results between rigid-
and flexible-floor analyses.For example, the two results can be
a significant shift with a period of time, so the comparison of
them at a certain time will cause error. But this problem
doesnot exist in the response-spectrum analysis, since only the
maximum responsesare calculated in this method. A number of
different formulas have been proposedto obtain a reasonable
estimate of the maximum response from the spectral values.The
simplest and most popular of these is the square root of the sum
of the squares (SRSS) of the maximum modal response. Thus, this
method was used to calculated the displacement and the member
force in this study.
|
|---|
