|
|
|
|
|
(a) Typical story layout
|
(b) Vertical view
|
(c) Perspective view
|
Figure 1 OpenSees Models of a 606 m Mega-tall
Building
Two mesh schemes are considered, which results in two models: Model 1 and
Model 2. Model 1 is relatively coarse, which has 27220 nodes. The coupling
beams and shear walls of Model 1 is subdivided to generate Model 2, which
has 81572 nodes. The shear walls and coupling beams are simulated using ShellMITC4
in OpenSees.
There are some differences in defining the
elastic modulus of concrete in ETABS and OpenSees, which will affect the calculated
vibration periods. When we conduct the modal analysis, the elastic modulus
of concrete material in OpenSees should be adjusted following the equation
Ec=2fc/e0, where Ec is elastic
modulus of concrete in ETABS. Through the adjustment, the vibration periods
of OpenSees agree well with those of ETABS.
Note that the commercial software package
MSC.Marc has been widely used in nonlinear analyses of super-tall buildings
and well validated. The corresponding models are also established in MSC.Marc
(denoted as Model1M and Model2M, respectively). Through the comparison between
the time history analysis results of OpenSees and MSC.Marc, the reliability
of the OpenSees models is validated.
The El-Centro ground motion, which is scaled
to a value of PGA of 220 cm/s2 (i.e., the MCE level ground motion
of this building according to the Chinese code), is firstly used to conduct
the time history analysis of Model1 and Model1M. Note that both CPU and GPU
solvers (Tian et al. 2015) are used to perform the time history analysis to
validate the reliability and facilitate different researchers to use these
models. If you want to use the CPU version model, please use ˇ°dynamiccpu.tclˇ±,
else please use ˇ°dynamicgpu.tclˇ± in the zip files.
The time history curves of the roof and
the envelope of inter-story drift are presented in Figure 2. Good agreements
are observed. In addition, the time history results obtained from CPU and
GPU are also compared in Figure 3. The results are identical, thus validating
the reliability of the GPU solver.
|
|
|
|
(a)
Time
history curves of the roof
|
(b)
Envelope
of the inter-story drift
|
|
Figure 2 Comparison of the time history
analysis results of Model1 and Model1M
|
|
|
|
|
(a)
Time
history curves of the roof
|
(b)
Envelope
of the inter-story drift
|
|
Figure 3 Comparison of the time history
analysis results of Model1 using CPU and GPU
|
Then, the time history analysis of Model2
and Model2M are also performed. The time history curves of the roof and the
envelope of inter-story drift are presented in Figure 4. Good agreements are
also observed, thus validating the reliability of Model2.
|
|
|
|
(a)
Time
history curves of the roof
|
(b)
Envelope
of the inter-story drift
|
|
Figure 4 Comparison of the time history
analysis results of Model2 and Model2M
|
Note that we define several recorders. Among
these, dispx.txt, dispy.txt and dispz.txt record the translational displacements
of all nodes for post-process, which will consume considerable hardware memories,
so please use it with caution.
If you have any questions, please feel free
to contact with Prof. Xinzheng Lu via: luxz@tsinghua.edu.cn.
Reference
Lu XL, Jiang C, Jiang
HJ. A benchmark model of mega-tall Buildings and analysis of its seismic responses.
Structural Engineers, 2015a, 31(4): 100-107. (in Chinese)
Lu
XZ, Xie LL, Guan H, Huang YL, Lu X. A shear wall element for nonlinear
seismic analysis of super-tall buildings using OpenSees. Finite Elements
in Analysis and Design, 2015b, 98: 14-25.
Tian
Y, Xie LL, Xu Z, Lu XZ. GPU-powered high-performance computing for the
analysis of large-scale structures based on OpenSees. Computing in Civil
Engineering, 2015: 411-418.
|
Download
Model TCL File