Design of Earthquake Simulator: Study of Seismic Isolators & Classical Architectural Materials in Earthquakes.
Marco Eberth ‘14 & Claire Anderson ’14
Advisor: Dr. Joseph Palladino
Department of Engineering, Trinity College, Hartford, CT
Abstract
During the first semester the team [Marco Eberth & Claire Anderson] designed and
constructed a shaker table for earthquake simulation. Prior to construction, the
team consulted articles and Trinity’s own earthquake enthusiast, Professor Geiss,
and determined that horizontal ground motion was only necessary for testing the
failure of buildings. The team determined a jigsaw would be sufficient in the driving
mechanism of the shaker table, and at lowest speeds the finished shaker table was
able to create a 2g earthquake. This was determined sufficient in testing considering
a 1.24g earthquake typically results in “very heavy” damage. During the second
semester Marco worked independently and constructed a seismic isolator: a system
below the super structure of a building that creates a barrier between the ground
movement and the movement the building experiences. Through dimensional
analysis it was determined that the building and table were not dynamically similar;
meaning that the earthquake was unrealistically violent for a light-weight scaled
building. An accelerometer was utilized to determine the efficiency in performance
of the seismic isolator. Using bearings as isolators, the top floor acceleration of the
building was measured while the ground acceleration was measured
simultaneously. There was an average of a 75% decrease in acceleration
experienced by the building in comparison to the ground acceleration during the
same simulated earthquake. During the second semester, Claire designed and
constructed model columns, similar to columns used in the reconstruction of
Classical architecture. Three different methods of reinforcement were used to join
the column drums together, each type was tested on the shake table. The results
were inconclusive due to restrictions of the accelerometers.
Introduction
Earthquake Study: As mentioned in the Abstract, through references and professional
advice it was determined that horizontal ground acceleration was the main cause of
building failure. Table 1 displays the peak ground acceleration and the predicted
damage by the United States Geological Survey. This table was used in the design of
the shaker table to be able to create a ground acceleration greater than 1.24g. It is
also used during the study of seismic isolation to determine the theoretical decrease
in potential damage.
Table 1: Peak Ground Acceleration and Predicted Damage of Buildings
[http://en.wikipedia.org/wiki/Peak_ground_acceleration]
Seismic Isolators: The isolators create a “barrier” between the ground acceleration
during and earthquake and the acceleration the building experiences. There are two
commonly used isolators in commercial use: Elastomeric Isolator and Flat Sliding
Isolator. An Elastomeric isolator is one that is composed of layers of rubber and
stainless steel with a lead core. The elastomeric isolator acts like a vertebrae due to
its layer and elastic properties. The Flat Sliding Isolator uses tracks and bearing
designs that all the building to slide on the ground and the ground vibrates
horizontally. Through shear analysis, cost analysis, and the availability to material, a
flat sliding isolator was selected for testing
Classical Architecture: In conjunction with a thesis I’m writing on the conservation
and reconstruction of Classical Architecture, I chose to investigate the different
methods of reinforcing marble column drums. Classical Greek and Roman columns
are constructed by stacking a series of smaller marble drums atop one another and
reinforcing them with lead clamps. Lead, being a very soft metal, was replaced with
steel rebar reinforcements during the reconstruction of the columns. Where the
column fragments are missing, a new section is recast using a marble-aggregate
concrete reinforced with steel rebar throughout. These three methods of
reinforcement were modeled, tested on the shake table, and compared with each
other to determine the most effective method of reconstruction.
Seismic Isolator: Marco Eberth
Dimensional Analysis
Testing Methods: To test the ground acceleration a custom 3D printed iPhone case was made to hold the phone on the top
floor of the building while running the Accelerometer application and record the maximum horizontal acceleration
experienced while another iPhone was c-clamped to the ground and recorded the maximum ground acceleration
Flat Sliding Isolator: After the design of the Elastomeric Isolator was a failure the concept of the Flat Sliding Isolator was recreated and modeled using a 5” table drawer slide. It had the same components such as bearings and minor frictional forces.
See Below for Image 2 of Isolator Attached to Model Building.
Image 1: SolidWorks Design of 1:120 Scaled High Rise Testing Subject
Trial
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
AVG.
Max Acceleration (g)
with Isolators
0.541
0.494
0.532
0.551
0.499
0.545
0.540
0.588
0.546
0.578
0.480
0.495
0.492
0.490
0.501
0.505
0.497
0.496
0.506
0.510
0.519
Max Acceleration (g)
without Isolators
2.062
2.021
2.055
2.071
2.011
2.068
2.063
2.091
2.068
2.080
1.995
2.005
2.001
1.998
2.002
2.004
2.001
2.001
2.004
2.090
2.035
Building Design Process and Specifications: The materials used
for the design of the building were .5”x.5” square wood dowels
cut at 1” Length for the corner columns, .25”x.25” square wood
dowels cut at 1” length for the center columns, and 6”x6”
hardwood for the flooring [see Image 1]. The average floor to
floor height of a building is between 10’ and 13’, so in the model
1” represents 10’ in the prototype. Essentially the building is 10
stories, with 10ft between each floor
% Decrease of g Due to
Isolator
73.763
75.557
74.112
73.394
75.186
73.646
73.825
71.879
73.598
72.212
75.940
75.312
75.412
75.475
74.975
74.800
75.162
75.212
74.750
75.598
74.491
Image 2: Flat Sliding Isolator
Model [Drawer Slide]
Attached to Model Building
Results & Conclusion: When testing the model building with the
isolator attached the building experienced a horizontal
acceleration 75% less than that of the ground acceleration [See
Table 2 for Data]. After dimensional analysis was done, it was
concluded that acceleration ratio was 1:2437994. This ratio was
not realistic [later discussed in the Dimensional Analysis Section].
In theory, if this system was to be recreated at full scale and 1.24g
acceleration occurred during a Magnitude 10 earthquake, the
building would experience .31g. The system would reduce the
potential damage from “very heavy” to “moderate”, and a
Magnitude 10 earthquake to a magnitude 7.
Dimensional Analysis: Thanks to Professor Mertens, the dimensional
analysis was done based off of a previous study [See References: Study on
shaking table tests of building models]. Equation 1 displays a number
known as the mass density ratio.
• mm=mass of structural members in the model
• ma= artificial mass in the model
• mom= mass to simulate live loads and nonstructural members
• lr= length ratio
• mp= mass of structural members in prototype
• mop= mass of live loads and nonstructural members in prototype
Seismic Isolator Dimensional
Analysis Results & Conclusion: As
discussed in the Results of the
Seismic Isolator, the acceleration
ratio was unrealistically high. If this
Classical Column Dimensional
project were to be done again the
Analysis Results: For a length
dimensional analysis would have
ratio of 1:35, the shaker table as
been done first based on a scale
designed was capable of
model design. The model would
simulating the dynamically
have to be significantly heavier and
similar event.
the shaker table would be more
Prototype
Model
powerful to support the load and Weight (lb)
45,000
11.5
have a slower horizontal
Time (s)
15
2.4
acceleration.
Frequency (Hz)
1.27
7.94
Acceleration
Model Design and Construction:
Dimensions – The prototype columns I used for modeling are found at a site in Sardis, Turkey, and they are 27 ft.
high with a diameter of 3.4 ft. With a length ratio of 1:35, the resulting model columns are 9.3 in. high with a diameter
of 1.14 in. They rest atop a base measuring ¾ in. in height, for ease of attachment to the table.
Materials – Marble was not a feasible material to use for the columns, due to time and money constraints, so a
concrete formula similar to one used in replacement of column fragments at Sardis was used. The marble chip
aggregate was too large to use for the model scale, so a marble powder To simulate the column drums, the model was
constructed using four separately cast cylinders. For the first reinforcement method, steel rod (1/8th in.) was used to
“pin” the drums together, as seen in Figure 4a. For second method, the steel rod was inserted through the entire
column and into the base, as seen in Figures 3 and 4c. The last method, to simulate the original construction method,
lead solder was used to join the column sections together, as seen in Figure 4b.
Figure 3: SolidWorks representation of the full
column construction.
Testing Procedure: The columns were
mounted on the shaker table, and an
additional ten pounds were added using a
wooden superstructure. A seismic event that
was determined to be dynamically similar to
an earthquake with the 1:35 scale was
created, and the acceleration difference
between the top of the column and the table
was recorded. After each test, the columns
were surveyed to identify any structural
damage incurred.
4.53g
Shaker Table Specifications
Table 2:Chart Displaying Percent Decrease of Ground Acceleration Experienced due to Isolator
Classical Architecture
1.24g
Components:
• Ryobi ZRJS481L 4.8 Amp Variable Speed Jigsaw with SpeedMatch
• Liberty 22 in. Ball Bearing Full Extension Drawer Slide (2-Pack)
• MCEC Materials (Special thanks to Andrew Musselin)
Discussion: As discussed in the fall portion of the project, the table used
a DAQ board and computer program to determine the duration of the
simulation. As the building testing was done it was noticed that
sometimes the trigger needed to be pumped a few times to overcome
the moment of inertia. From that the team tried using an external
potentiometer to determine the amount of voltage supplied to the
Jigsaw. In the end the trials had to be run by hand at the slowest speed
settings resulting in earthquakes around 2.0g. The iPhone application
maxed out at just over 2.1g, which allowed this system to work.
References & Acknowledgements
References
1.
2.
3.
Figure 4 (a, b, c): Solidworks models of the
reinforcement methods. (a) steel rod pins. (b) lead
solder “clamps”. (c) steel rod throughout
Results and Conclusion: Although the shaker table was capable of creating a 4.5g earthquake, the accelerometers used
were only able to record an event up to 2g. For shake table tests below this limit, the behavior of all three column
constructions were the same. When the acceleration was increased beyond the 2g limit, the construction with the steel
rod displayed the most desirable behavior, but the accelerations were not able to be recorded.
4.
5.
Base Isolation: Origins and Development: http://nisee.berkeley.edu/lessons/kelly.html
Seismic Isolation and Energy Dissipation Systems:
http://webshaker.ucsd.edu/homework/Base_Isolation_Web.pdf
Study on shaking table tests of building models:
http://sstl.cee.illinois.edu/Beijing_Symposium/p-60/IV-8.Zhang.pdf
Yegul. 1986. The Bath-Gymnasium Complex at Sardis. Harvard University Publishing
Hanfmann. Excavation Reports from Sardis.
Acknowledgements:
Andrew Musulin for his help in constructing the shaker table
Professor Mertens for his help in dimensional analysis
Professor Geiss for his help in Earthquake Engineering