Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Performance Study of an Eight-story Steel Building Equipped with Oil Dampers Damaged During the 2011 Great East Japan Earthquake Part 1: Structural Identification and Damage Reasoning

Performance Study of an Eight-story Steel Building Equipped with Oil Dampers Damaged During the... Oil dampers installed on the first floor of an eight-story steel building were completely destroyed during the 2011 Great East Japan Earthquake. It is believed to be the first time in the world that real oil dampers in service failed due to earthquakes. Before this failure event, the actual performance of buildings that use oil dampers during catastrophic earthquakes has never been verified. Investigating the cause of the damage of the oil dampers is thus necessary and urgent. In this paper, a comprehensive identification was conducted to rebuild the numerical model of this damped structure equipped with/without damaged oil dampers using the measurement data of the installed monitoring system. Furthermore, the damage process of the oil dampers was postulated based on the identification and simulation results. The limit states of the oil dampers were studied. Based on the damages of the dampers and connection, the oil dampers experienced the displacement limit state when the allowable displacement limit was surpassed and the central cylinder pushed against the nsufficient stroke limit abutment. The i is the main cause of the collision between the damper and the abutment on the floor, which finally led to the failure of the oil dampers. Keywords: oil damper; simulation; steel building; damage; the 2011 Great East Japan Earthquake 1. Introduction system was tested during nature-generated earthquakes Since the mid-1990s, passive energy dissipation (Chang and Lin, 2004) or using free and forced devices have been used more widely to enhance the vibration (Lai et al., 1995); some small-scale passively energy dissipation capability of a structure and reduce controlled buildings were tested in shaking tables the damage to the structural frame in which they are (Chang et al., 1995, Chang et al., 2001). However, the installed. Dampers and isolators, such as viscous fluid actual performance of these passive devices during dampers, viscoelastic dampers, metallic dampers, major and catastrophic earthquakes has never been and rubber bearings, are considered effective and tested. In 2007, Japan tested a full-scale five-story reliable devices to mitigate seismic hazards as well steel frame building with different types of passive as rehabilitate deteriorating or deficient structures dampers on the E-Defense, which is the largest three- (Symans et al., 2008, Soong and Spencer Jr., 2002). dimensional shaking table in the world, to validate the According to the Japan Society of Seismic Isolation performance of these passive dampers under extreme (Kasai et al., 2013), more than 7,000 buildings in circumstances (Ji et al., 2013, Kasai et al., 2010). The Japan are equipped with seismic isolation systems or yielding of the frame member cases and ultimate state supplemental damping systems. of dampers were never investigated experimentally Researchers have been experimenting with passively because the frame members remained mostly elastic controlled structures to examine and evaluate the real during the shaking table tests. performance of passive devices after deployment. The 3/11 Japan earthquake, which hit Sendai City, A full-scale building with a supplemental damping is an unprecedented case that provides an opportunity to observe the actual performance of an eight-story passively controlled steel building during a catastrophic *Contact Author: Songtao Xue, earthquake (Cao et al., 2012). All eight sets of oil Department of Architecture, Tohoku Institute of Technology, dampers on the first floor were completely destroyed Sendai, Japan with abutment breakage. This event is believed to be E-mail: xuest@tohtech.ac.jp the first reported case in which real oil dampers in ( Received April 7, 2014 ; accepted October 27, 2014 ) service were damaged during earthquakes. Journal of Asian Architecture and Building Engineering/January 2015/188 181 Generally speaking, passive energy dissipation devices were not yet installed. Forced vibration tests devices are considered sufficiently safe during major were conducted to identify the natural frequencies earthquakes. As suggested in current guidelines, passive of the structure with/without the oil dampers. Before devices have a displacement/velocity capacity with a the installation of oil dampers, the building was certain safety margin under the maximum considered vibrated by using an excitation machine that was earthquake level. However, this failure event may cast mounted on the roof, and the first natural frequencies doubt on the safety of these passive devices. Therefore, of the building without dampers were obtained. The investigating the cause of the damage as well as the frequencies were 1.050 Hz for the short side (nearly limit state of the oil dampers in this event is important. the NS direction) and 1.025 Hz for the long side This paper investigates the damage event of this (nearly the EW direction). After the installation of the passively controlled building during the 2011 Great dampers, the first natural frequencies of the building East Japan Earthquake and studies the cause of the with dampers were 1.125 Hz for the short side and 1.100 damage of the oil dampers during the earthquake. Hz for the long side. The natural frequencies in both The damage process of the oil dampers is postulated directions increased because oil dampers increase the through identification and simulation. The limit states stiffness of the building. of the dampers are discussed according to the damage description of the connection and dampers. 2. Building Description and Damage Event 2.1 Overview of the Building The eight-story Administration Building on the main campus of the Tohoku Institute of Technology (in Sendai, Japan) was constructed in 2003 and has a length of 48 m, a width of 9.6 m, and a height of 34.2 m, as shown in Fig.1. The superstructure is a steel-frame structure with precast concrete slabs, and the one-story basement is reinforced concrete. This steel building was designed to satisfy the Japanese Earthquake Resistance Code for School Buildings without any control equipment. To verify the effectiveness of the newly developed oil damper at the Tohoku Institute of Technology and improve the structural earthquake resistance capability, 56 sets of dampers were installed, as shown in Fig.2. Each floor has eight sets of oil dampers, which are connected with the use of V-type braces between the adjacent floors, as illustrated in Fig.2. Allocation of Oil Dampers Fig.3. The first floor and the second floor were merged to form a large space with a height of 8 m. Fig.3. Brace of the Oil Damper 2.2 Oil Dampers Kawamata et al. (Funaki et al., 2001, Kawamata et al., 2000) proposed a new type of oil damper that is Fig.1. Administration Building of the Tohoku Institute of composed of a cylinder and a pair of pistons. The gaps Technology between the cylinder (oil container) and pistons are The main frame of the superstructure was completed sealed with viscoelastic polymer, as illustrated in Figs.4. in June 2002, but the partition walls and building and 5. The sealing scheme relaxes the manufacturing 182 JAABE vol.14 no.1 January 2015 Liyu Xie precision of matching the cylinders and pistons, and 2.3 Monitoring System makes the damper compact and cheap. The contained A monitoring system is installed in this building. oil flows through the narrow orifice and creates strong Two-direction accelerometers were located on fluid turbulence to dissipate energy when the piston the first, fourth, and eighth floors to record the moves reciprocally. Furthermore, the sealing polymer dynamic responses of the structure; the location of is subjected to dynamic shearing deformation, which accelerometers is shown in Fig.2. To examine the results in additional viscoelastic resisting force. On performance of oil dampers in real application, load the basis of the harmonic excitation tests, the resisting cells with a strain meter and displacement transducers force provided by oil flow acts nonlinearly with the were installed on the dampers on the first and eighth piston velocity and appears to be stronger under a floors in both directions to obtain the restoring force large amplitude and high excitation frequency, while and displacement of the dampers during earthquakes. the viscoelastic resisting force of the sealing polymer 2.4 Damage Event increases almost linearly with the piston velocity. During the 2011 Great East Japan Earthquake Two types of oil dampers are installed in the (also known as the 3/11 Earthquake), all eight sets of Administration Building to investigate the performance dampers on the first floor were destroyed, as shown of the oil dampers in real application. An oil damper is in Fig.6. The damper pistons on both sides were torn 424 mm wide and 328 mm high with different piston from the central cylinder (oil container). The U-type diameters and orifice specifications for the 1F floor abutments fixed on the first floor were opened wider and 3–8F floors. These two types of dampers also as the pistons ran out of the stroke limit and cushion have different stroke limits (the displacement extent limit, and pushed against the abutments. Sixteen sets of piston in one direction), which are 16 mm for 1F of dampers on the third and fourth floor had severe dampers and 8 mm for 3–8F dampers, respectively. oil leakages, because the sealing viscoelastic polymer The pistons are fixed with the U-type abutment on the had worn out. However, the mechanical parts of the oil floor and the central cylinder is attached to the V-type dampers remained undeformed. brace. The moving direction of pistons is horizontal. The acquisition devices failed to record the dynamic Therefore, the damper displacement is the same as the responses of the building because of the power failure interstory drift, without considering the deformation that occurred when the extremely intensive earthquake of the rigid V-type braces. The central cylinder moves struck the building on March 11, 2011. Fortunately, an back and forth along the axis of the pistons within the observation station of ground motion about 50 m from stroke limit when the interstory drift occurs because the building successfully captured the ground motion of ground motion. An extra cushion limit is prepared of the 3/11 Earthquake and recorded the peak ground (8 mm for 1F dampers, 5 mm for 3–8F dampers) to acceleration (PGA) of 354 gal in the EW direction and protect the central cylinder from colliding with the 280 gal in the NS direction. After the earthquake, a abutment. field observation was immediately conducted and no cracks were found on the surface of the beam-column joints. No other structural damage was found except the damaged oil dampers and oil leakage. This building was put into use without any retrofitting after a quick safety evaluation because it was designed to satisfy the Japanese Earthquake Resistance Code for School Buildings even without oil dampers. Fig.4. Intact Oil Damper Fig.6. Damaged Oil Damper and Abutment 3. Simulation Model of the Building 3.1 Numerical Model For the nonlinear dynamic analysis, the numerical model of this steel building was established by using finite element analysis software SAP2000, as shown in Fig.7. The superstructure was modeled and fixed on the Fig.5. Dimension of the Oil Damper JAABE vol.14 no.1 January 2015 Liyu Xie 183 ground without considering the basement and the soil- displacement and the restoring force of the oil dampers. structure interaction. All structural members, including The parameters of the damper model are determined by the steel columns, beams, and braces of the dampers, minimizing the difference between the captured data were modeled according to the described dimensions and the simulated response, which is used in the next in the design book. Moreover, the material parameters identification step. were set based on the Japanese Steel Design Code. The The second step is to find suitable values for the bilinear plastic-elastic model was used for the steel mass density of the slab, which are set as variables for material. The joints between columns and beams were the additional weight consideration. The optimization rigidly fixed. The concrete slabs were modeled using target is to minimize the differences of the resonant shell elements in plane stress. The partition walls, frequencies and the maximum acceleration responses equipment, and furniture on the floors were regarded as between the measured and simulated responses lumped mass that was attached to the floor. Therefore, by adjusting the variables using the optimization the mass density of the slab material is referred to as technique. During optimization, more weights are a changeable variable, because the lumped weight is assigned to the lower resonant frequencies because considered. l o we r m o d a l i n f o r m a t i o n i s r e l i a b l e a n d m o r e The actual restoring force of an oil damper consists important. of the resisting force generated by oil flows and the viscoelastic shear force from the sealing polymer. In 4. Data Analysis and Damage Reasoning experiments, the restoring force is both amplitude 4.1 Earthquake Records and frequency dependent. However, in a numerical The structural responses induced by earthquakes simulation, an oil damper model is simplified, which is before and after the 3/11 Earthquake were collected by represented by link elements with the type of damper using the monitoring system; the collected data provide in SAP2000. The mathematical model for describing valuable information about the state of the building. the behavior of oil dampers is given through the The dataset can be divided into two categories; one following nonlinear force-velocity relation: is the building equipped with dampers and the other without dampers. The foreshock record of March 11, 2011 and the aftershock record of April 7, 2011, which represents each category, respectively, are used where F(t) is the restoring force developed by the for identification and analysis. No measurement was damper; u(t) is the relative displacement between the recorded during the 3/11 Earthquake. Thus, the ground oil piston and the abutment; u(t) is relative velocity motion captured by an observation station is used as between the oil piston and the abutment; C is the the ground excitation for the simulation of the main damping coefficient; is the exponent whose value is shock on March 11, 2011. determined using the piston head orifice design; and 4.2 Foreshock on March 9, 2011 sgn[·] is the signum function. The building equipped with dampers experienced an earthquake two days before the main shock. The ground motion of March 9, 2011 had a PGA of 32 gal in the EW direction and a PGA of 26 gal in the NS direction. The two-step identification procedure was performed to build the numerical model for nonlinear dynamic analysis. Fig.8.(a) shows the acceleration responses on the fourth floor of observation and the simulation in both directions. Fig.8.(b) demonstrates the transfer functions between the fourth and the first floors in the frequency domain. Fig.8.(c) compares the maximum acceleration on each floor of observation and simulation. The acceptable agreement between the measured and simulated responses verified the identified numerical model. Fig.7. FEM Model of Building with Dampers 4.3 Aftershock on April 7, 2011 After the 2011 Great East Japan Earthquake, a large 3.2 Identification Results aftershock struck the building in the middle of the For t h e e st a bl i she d n um e r i c a l m o de l , som e night on April 7, 2011, which was almost one month parameters remain undetermined. Based on the after the main shock. The seismic intensity of this measured response data of the building, two steps aftershock was the same as that of the main shock in are needed to identify the unknown parameters. The Sendai City. This aftershock damaged some buildings first step is to build the mathematical model of the oil in Sendai City, which were safe during the main shock. dampers. The monitoring system collects the relative Fortunately, this aftershock caused no further damage 184 JAABE vol.14 no.1 January 2015 Liyu Xie (a) Time History of Acceleration (a) Time History of Acceleration (b) Transfer Function (b) Transfer Function (c) Maximum Acceleration on Each Floor (c) Maximum Acceleration on Each Floor Fig.9. Comparison of the Observation and Simulation Fig.8. Comparison of the Observation and Simulation (April 7, 2011) (March 9, 2011) JAABE vol.14 no.1 January 2015 Liyu Xie 185 to the administration building. The ground motion of April 7, 2011 had a PGA of 176 gal in the EW direction and a PGA of 289 gal in the NS direction. The main shock on March 11, 2011 devastated the dampers on the first floor and disabled the dampers on the third and fourth floors because of oil leakage, whereas the main frame of the building had no structural damage. No retrofitting measurement was conducted before April 7, 2011. The numerical model is altered to consider the change of damper configuration. The oil dampers of the first, third, and fourth floors were removed from the model in both directions. The mass distribution of this building in the vertical direction was also changed because most of the bookshelves that fell over during the 3/11 Earthquake remained on the floor on April 7. The identification was conducted to rebuild the numerical model for the case of April 7, 2011. Fig.9. shows the acceleration responses that were recorded and simulated in the time domain of the fourth floor as well as the transfer functions in the frequency (a) Time History of Acceleration domain and the maximum acceleration of each floor. 4.4 Simulation of Main Shock The acquisition system did not work during the 3/11 Earthquake because of the power blackout. Thus, the numerical model of the steel building could not be built through direct identification by using response measurement. However, the numerical model for the main shock can be established by combining the identification results of the foreshock and aftershock. The mass distribution of the building during the main shock is the same as that of the undamaged building before the main shock. It is assumed that the dampers on the first, third, and fourth floors were not in operation from the beginning on March 11. The identified mathematical models of the oil dampers in the case of aftershock are used to simulate the velocity- force behavior of dampers during the 3/11 Earthquake. In the simulation of the main shock, the ground motion captured by an observation station was used as the ground excitation. This step aimed to simulate the dynamic responses of the building and predict the maximum displacement of the dampers during the main shock. (b) Transfer Function The acceleration responses on the fourth floor simulated for the main shock are plotted in Fig.10.(a). Fig.10. Simulated Results of Earthquake (March 11, 2011) The transfer functions between the fourth floor and the ground excitation are plotted in Fig.10.(b). The transfer functions from the first floor to the eighth floor in the NS direction are shown in Fig.11. for a comparison between the dynamic properties of the foreshock, the main shock, and the aftershock. The first period of aftershock is larger than that of foreshock because of the absence of the dampers on the first, third, and fourth floors. The maximum magnification factor in the first period of the aftershock is higher than in the other two cases. These facts reveal the effectiveness of the dampers in adding damping as well as suppressing building vibration. Fig.11. Comparison of the Transfer Function 186 JAABE vol.14 no.1 January 2015 Liyu Xie 4.5 Limit State of Oil Dampers Buildings equipped with fluid oil dampers have not been subjected to intensive earthquakes. In addition, no data or failure events of oil dampers during catastrophic events were available until the oil dampers of this steel building were devastated during the 3/11 Earthquake. Even in experiments, the limit states of the dampers were seldom tested. To evaluate the near collapse state of passively controlled structures, Miyamoto et al. (Miyamoto et al., 2010) conducted a comprehensive experimental investigation to determine the limit states of a viscous damper and established a detailed mathematical model of a viscous damper that considers the limit states of dampers. Miyamoto Fig.12. Diagrammatic View of the Damper's Limits examined several limited states in the laboratory. The two most common types are the force limit state and the displacement limit state. The force limit state can occur when a damper is subjected to a large-velocity (force) pulse within the stroke limit and when one of the mechanical parts of the dampers reaches the corresponding material limit. The displacement limit state can occur when the stroke limit in extension and retraction is reached. During the 3/11 Earthquake, the dampers on the first floor were damaged severely and the abutments were wrecked. The damage description indicates that both the stroke and cushion limits of the pistons seemed to have been reached, and the central cylinders were pounded against the abutment repeatedly until the abutment connection and oil dampers were broken. Therefore, the displacement limit state is assumed to be the damage scenario of the oil dampers. In this study, the damage process of oil dampers was replayed based on identification and simulation. According to the design book, the designed stroke limit displacement of the damper for the first floor is 16 mm (pink line in Fig.13.) and 8 mm for the third to eighth floors. In addition to damper protection, a cushion limit was also designed with 8 mm (red line) for the first floor and 5 mm for the third to eighth floors. The damper reaches the displacement limit state and the cylinder pushes against the abutment when the stroke and cushion limits run out. If the cylinder does not retract and continues to push against the abutment, the relative piston-cylinder velocity drops to zero. Thus, the restoring force of oil dampers becomes zero, and the V-type steel brace and the abutment on the floor carry the resistant force caused by restraining the interstory drift beyond the stroke and cushion Fig.13. Displacement of Dampers During the 3/11 Earthquake limits. Once the dampers reach the destructive limit displacement, as depicted in Fig.12., the accumulated out and the oil leaked during the 3/11 Earthquake. A deformation energy is abruptly released to break the reasonable assumption is that the oil dampers on the abutment connection and oil dampers. However, the fourth floor were in critical condition and were about difficulty lies in determining the destructive limit to reach the destructive limit displacement. Based on displacement of the abutment. the simulation results of the fourth floor oil dampers, Considering the situation of dampers on the fourth the ratio of the maximum interstory drift to the sum of floor, the oil container, piston, and abutment were not stroke and cushion limits can be calculated and used damaged. However, the sealing material had worn as a normalized index to determine the deformation JAABE vol.14 no.1 January 2015 Liyu Xie 187 limit of other floors (2.6 for the fourth floor). With this displacement/velocity capacity under the considered normalized index applied to the first floor, the destructive earthquake level. Improper design of dampers limit displacement was determined to be 59.8 mm, which should be avoided. Dampers are not made to survive is shown as black dashed lines in Fig.13. The simulated catastrophic earthquakes. The limit states of dampers interstory drift of the damper on the first floor in the EW should be considered during design, especially when direction on March 11, 2011 is plotted in Fig.13. The earthquakes beyond the considered level are likely to dampers of the first floor were most likely destroyed in shake the passively controlled structure. about 90 seconds during the 3/11 Earthquake. The failure event has prompted researchers and Of course, there are some flaws in this damage engineers to review the design philosophy of passively simulation. Continuous impacts of the central cylinder controlled buildings. Considering the limit state of against the abutment caused the U-type abutment to the dampers during the design is necessary at this open. The number of collisions between the central point because future catastrophic events may damage cylinder and the abutment before the breakage of the the dampers. This study will benefit the earthquake abutment and oil dampers cannot be determined. In engineering community. this study, the breakage is assumed to be caused by the impact of maximum displacement. Furthermore, the References 1) Cao, M., Tang, H., Funaki, N. and Xue, S. (2012) Study on a real mathematic model of the oil damper does not incorporate 8F steel building with oil damper damaged during the 2011 Great the limit state of dampers. Thus, the maximum interstory East Japan Earthquake. 15th World Conference on Earthquake drift in the simulation is not the same as in the real case. Engineering. Lisbon, Portugal. 2) Chang, K. C., Soong, T. T., Oh, S. and Lai, M. L. (1995) Seismic 5. Conclusion and Discussion behavior of steel frame with added viscoelastic dampers. Journal of structural engineering, 121(10), pp.1418-1426. This paper reports the first failure event of oil 3) Chang, K. C., Tsai, M. H. and Lai, M. L. (2001) Shaking table dampers during the 2011 Great East Japan Earthquake. study of a 2/5 scale steel frame with new viscoelastic dampers. Rebuilding the numerical model of oil dampers and the Structural Engineering and Mechanics, 11(3), pp.273-286. steel structure is possible through a monitoring system 4) Chang, K. and Lin, Y. (2004) Seismic response of full-scale and two-step identification. Based on the identification structure with added viscoelastic dampers. Journal of Structural Engineering, 130(4), pp.600-608. and simulation, the damage process of the oil dampers 5) Funaki, N., Kang, J. and Kawamata, S. (2001) Vibration response was investigated. of a three-storied full-scale test building passively controlled by The main cause of the damper failure was the liquid dampers sealed by viscoelastic material. 16th International insufficient stroke range of pistons. The total allowable Conference on Structural Mechanics in Reactor Technology. m o v i n g r a n g e o f o i l d a m p e r s wa s 2 4 m m f o r t h e Washington DC, USA. 6) Ji, X., Hikino, T., Kasai, K. and Nakashima, M. (2013) Damping dampers on the first floor in one direction (the sum identification of a full-scale passively controlled five-story steel of the 16 mm stroke limit and 8 mm cushion limit). building structure. Earthquake Engineering and Structural Thus, the interstory drift angle of the first floor was Dynamics, 42(2), pp.277-295. 0.3% when the damper reached the displacement 7) Kasai, K., Ito, H., Ooki, Y., Hikino, T., Kajiwara, K., Motoyui, limit. Furthermore, the elastic limit of the interstory S., Ozaki, H. and Ishii, M. (2010) Full scale shake table tests of 5-story steel building with various dampers. Proceedings of the drift angle in the Japan Steel Building Code is 0.5% 9th U.S. National and 10th Canadian Conference on Earthquake for minor earthquakes. The authors can infer that the Engineering. Toronto. original design of this administration building intended 8) Kasai, K., Mita, A., Kitamura, H., Matsuda, K., Morgan, T. A. to use oil dampers to control structural responses and Taylor, A. W. (2013) Performance of seismic protection to minor earthquakes while the structure responds technologies during the 2011 Tohoku-Oki Earthquake. Earthquake Spectra, 29(s1), S265-S293. elastically. However, the unprecedentedly intense 9) Kawamata, S., Funaki, N. and Itoh, Y. (2000) Passive control of earthquake caused large interstory drift, which the oil building frames by means of liquid dampers sealed by viscoelastic dampers could not withstand and consequently broke material. 12th World Conference on Earthquake Engineering. the dampers. Auckland, New Zealand. The limit states of oil dampers were studied based on 10) Lai, M. L., Chang, K. C., Soong, T. T., Hao, D. S. and Yeh, Y. C. (1995) Full-scale viscoelastically damped steel frame. Journal of the damage scenario. The oil dampers experienced the Structural Engineering, 121(10), pp.1443-1447. displacement limit state when the allowable limit was 11) Miyamoto, H., Gilani, A. S., Wada, A. and Ariyaratana, C. (2010) surpassed and the central cylinder pushed against the Limit states and failure mechanisms of viscous dampers and the abutment. In the future, more experiments should be implications for large earthquakes. Earthquake Engineering and conducted to determine all the possible limit states of Structural Dynamics, 39(11), pp.1279-1297. 12) Soong, T. T. and Spencer Jr, B. F. (2002) Supplemental energy oil dampers, and a numerical model that incorporates dissipation: state-of-the-art and state-of-the-practice. Engineering the limit states can be established to test the collapse Structures, 24(3), pp.243-259. state under intensive earthquakes. 13) Symans, M. D., Charney, F. A., Whittaker, A. S., Constantinou, Passive energy dissipation devices are deployed M. C., Kircher, C. A., Johnson, M. W. and McNamara, R. J. to reduce the inelastic energy dissipation demand (2008) Energy dissipation systems for seismic applications: current practice and recent developments. Journal of Structural induced on a structure, especially requested by a Engineering, 134(1), pp.3-21. major earthquake. The passive devices shall have a 188 JAABE vol.14 no.1 January 2015 Liyu Xie http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Asian Architecture and Building Engineering Taylor & Francis

Performance Study of an Eight-story Steel Building Equipped with Oil Dampers Damaged During the 2011 Great East Japan Earthquake Part 1: Structural Identification and Damage Reasoning

Download PDF
8 pages

Loading next page...
 
/lp/taylor-francis/performance-study-of-an-eight-story-steel-building-equipped-with-oil-JfZyQE7Vwm

References (12)

Publisher
Taylor & Francis
Copyright
© 2018 Architectural Institute of Japan
ISSN
1347-2852
eISSN
1346-7581
DOI
10.3130/jaabe.14.181
Publisher site
See Article on Publisher Site

Abstract

Oil dampers installed on the first floor of an eight-story steel building were completely destroyed during the 2011 Great East Japan Earthquake. It is believed to be the first time in the world that real oil dampers in service failed due to earthquakes. Before this failure event, the actual performance of buildings that use oil dampers during catastrophic earthquakes has never been verified. Investigating the cause of the damage of the oil dampers is thus necessary and urgent. In this paper, a comprehensive identification was conducted to rebuild the numerical model of this damped structure equipped with/without damaged oil dampers using the measurement data of the installed monitoring system. Furthermore, the damage process of the oil dampers was postulated based on the identification and simulation results. The limit states of the oil dampers were studied. Based on the damages of the dampers and connection, the oil dampers experienced the displacement limit state when the allowable displacement limit was surpassed and the central cylinder pushed against the nsufficient stroke limit abutment. The i is the main cause of the collision between the damper and the abutment on the floor, which finally led to the failure of the oil dampers. Keywords: oil damper; simulation; steel building; damage; the 2011 Great East Japan Earthquake 1. Introduction system was tested during nature-generated earthquakes Since the mid-1990s, passive energy dissipation (Chang and Lin, 2004) or using free and forced devices have been used more widely to enhance the vibration (Lai et al., 1995); some small-scale passively energy dissipation capability of a structure and reduce controlled buildings were tested in shaking tables the damage to the structural frame in which they are (Chang et al., 1995, Chang et al., 2001). However, the installed. Dampers and isolators, such as viscous fluid actual performance of these passive devices during dampers, viscoelastic dampers, metallic dampers, major and catastrophic earthquakes has never been and rubber bearings, are considered effective and tested. In 2007, Japan tested a full-scale five-story reliable devices to mitigate seismic hazards as well steel frame building with different types of passive as rehabilitate deteriorating or deficient structures dampers on the E-Defense, which is the largest three- (Symans et al., 2008, Soong and Spencer Jr., 2002). dimensional shaking table in the world, to validate the According to the Japan Society of Seismic Isolation performance of these passive dampers under extreme (Kasai et al., 2013), more than 7,000 buildings in circumstances (Ji et al., 2013, Kasai et al., 2010). The Japan are equipped with seismic isolation systems or yielding of the frame member cases and ultimate state supplemental damping systems. of dampers were never investigated experimentally Researchers have been experimenting with passively because the frame members remained mostly elastic controlled structures to examine and evaluate the real during the shaking table tests. performance of passive devices after deployment. The 3/11 Japan earthquake, which hit Sendai City, A full-scale building with a supplemental damping is an unprecedented case that provides an opportunity to observe the actual performance of an eight-story passively controlled steel building during a catastrophic *Contact Author: Songtao Xue, earthquake (Cao et al., 2012). All eight sets of oil Department of Architecture, Tohoku Institute of Technology, dampers on the first floor were completely destroyed Sendai, Japan with abutment breakage. This event is believed to be E-mail: xuest@tohtech.ac.jp the first reported case in which real oil dampers in ( Received April 7, 2014 ; accepted October 27, 2014 ) service were damaged during earthquakes. Journal of Asian Architecture and Building Engineering/January 2015/188 181 Generally speaking, passive energy dissipation devices were not yet installed. Forced vibration tests devices are considered sufficiently safe during major were conducted to identify the natural frequencies earthquakes. As suggested in current guidelines, passive of the structure with/without the oil dampers. Before devices have a displacement/velocity capacity with a the installation of oil dampers, the building was certain safety margin under the maximum considered vibrated by using an excitation machine that was earthquake level. However, this failure event may cast mounted on the roof, and the first natural frequencies doubt on the safety of these passive devices. Therefore, of the building without dampers were obtained. The investigating the cause of the damage as well as the frequencies were 1.050 Hz for the short side (nearly limit state of the oil dampers in this event is important. the NS direction) and 1.025 Hz for the long side This paper investigates the damage event of this (nearly the EW direction). After the installation of the passively controlled building during the 2011 Great dampers, the first natural frequencies of the building East Japan Earthquake and studies the cause of the with dampers were 1.125 Hz for the short side and 1.100 damage of the oil dampers during the earthquake. Hz for the long side. The natural frequencies in both The damage process of the oil dampers is postulated directions increased because oil dampers increase the through identification and simulation. The limit states stiffness of the building. of the dampers are discussed according to the damage description of the connection and dampers. 2. Building Description and Damage Event 2.1 Overview of the Building The eight-story Administration Building on the main campus of the Tohoku Institute of Technology (in Sendai, Japan) was constructed in 2003 and has a length of 48 m, a width of 9.6 m, and a height of 34.2 m, as shown in Fig.1. The superstructure is a steel-frame structure with precast concrete slabs, and the one-story basement is reinforced concrete. This steel building was designed to satisfy the Japanese Earthquake Resistance Code for School Buildings without any control equipment. To verify the effectiveness of the newly developed oil damper at the Tohoku Institute of Technology and improve the structural earthquake resistance capability, 56 sets of dampers were installed, as shown in Fig.2. Each floor has eight sets of oil dampers, which are connected with the use of V-type braces between the adjacent floors, as illustrated in Fig.2. Allocation of Oil Dampers Fig.3. The first floor and the second floor were merged to form a large space with a height of 8 m. Fig.3. Brace of the Oil Damper 2.2 Oil Dampers Kawamata et al. (Funaki et al., 2001, Kawamata et al., 2000) proposed a new type of oil damper that is Fig.1. Administration Building of the Tohoku Institute of composed of a cylinder and a pair of pistons. The gaps Technology between the cylinder (oil container) and pistons are The main frame of the superstructure was completed sealed with viscoelastic polymer, as illustrated in Figs.4. in June 2002, but the partition walls and building and 5. The sealing scheme relaxes the manufacturing 182 JAABE vol.14 no.1 January 2015 Liyu Xie precision of matching the cylinders and pistons, and 2.3 Monitoring System makes the damper compact and cheap. The contained A monitoring system is installed in this building. oil flows through the narrow orifice and creates strong Two-direction accelerometers were located on fluid turbulence to dissipate energy when the piston the first, fourth, and eighth floors to record the moves reciprocally. Furthermore, the sealing polymer dynamic responses of the structure; the location of is subjected to dynamic shearing deformation, which accelerometers is shown in Fig.2. To examine the results in additional viscoelastic resisting force. On performance of oil dampers in real application, load the basis of the harmonic excitation tests, the resisting cells with a strain meter and displacement transducers force provided by oil flow acts nonlinearly with the were installed on the dampers on the first and eighth piston velocity and appears to be stronger under a floors in both directions to obtain the restoring force large amplitude and high excitation frequency, while and displacement of the dampers during earthquakes. the viscoelastic resisting force of the sealing polymer 2.4 Damage Event increases almost linearly with the piston velocity. During the 2011 Great East Japan Earthquake Two types of oil dampers are installed in the (also known as the 3/11 Earthquake), all eight sets of Administration Building to investigate the performance dampers on the first floor were destroyed, as shown of the oil dampers in real application. An oil damper is in Fig.6. The damper pistons on both sides were torn 424 mm wide and 328 mm high with different piston from the central cylinder (oil container). The U-type diameters and orifice specifications for the 1F floor abutments fixed on the first floor were opened wider and 3–8F floors. These two types of dampers also as the pistons ran out of the stroke limit and cushion have different stroke limits (the displacement extent limit, and pushed against the abutments. Sixteen sets of piston in one direction), which are 16 mm for 1F of dampers on the third and fourth floor had severe dampers and 8 mm for 3–8F dampers, respectively. oil leakages, because the sealing viscoelastic polymer The pistons are fixed with the U-type abutment on the had worn out. However, the mechanical parts of the oil floor and the central cylinder is attached to the V-type dampers remained undeformed. brace. The moving direction of pistons is horizontal. The acquisition devices failed to record the dynamic Therefore, the damper displacement is the same as the responses of the building because of the power failure interstory drift, without considering the deformation that occurred when the extremely intensive earthquake of the rigid V-type braces. The central cylinder moves struck the building on March 11, 2011. Fortunately, an back and forth along the axis of the pistons within the observation station of ground motion about 50 m from stroke limit when the interstory drift occurs because the building successfully captured the ground motion of ground motion. An extra cushion limit is prepared of the 3/11 Earthquake and recorded the peak ground (8 mm for 1F dampers, 5 mm for 3–8F dampers) to acceleration (PGA) of 354 gal in the EW direction and protect the central cylinder from colliding with the 280 gal in the NS direction. After the earthquake, a abutment. field observation was immediately conducted and no cracks were found on the surface of the beam-column joints. No other structural damage was found except the damaged oil dampers and oil leakage. This building was put into use without any retrofitting after a quick safety evaluation because it was designed to satisfy the Japanese Earthquake Resistance Code for School Buildings even without oil dampers. Fig.4. Intact Oil Damper Fig.6. Damaged Oil Damper and Abutment 3. Simulation Model of the Building 3.1 Numerical Model For the nonlinear dynamic analysis, the numerical model of this steel building was established by using finite element analysis software SAP2000, as shown in Fig.7. The superstructure was modeled and fixed on the Fig.5. Dimension of the Oil Damper JAABE vol.14 no.1 January 2015 Liyu Xie 183 ground without considering the basement and the soil- displacement and the restoring force of the oil dampers. structure interaction. All structural members, including The parameters of the damper model are determined by the steel columns, beams, and braces of the dampers, minimizing the difference between the captured data were modeled according to the described dimensions and the simulated response, which is used in the next in the design book. Moreover, the material parameters identification step. were set based on the Japanese Steel Design Code. The The second step is to find suitable values for the bilinear plastic-elastic model was used for the steel mass density of the slab, which are set as variables for material. The joints between columns and beams were the additional weight consideration. The optimization rigidly fixed. The concrete slabs were modeled using target is to minimize the differences of the resonant shell elements in plane stress. The partition walls, frequencies and the maximum acceleration responses equipment, and furniture on the floors were regarded as between the measured and simulated responses lumped mass that was attached to the floor. Therefore, by adjusting the variables using the optimization the mass density of the slab material is referred to as technique. During optimization, more weights are a changeable variable, because the lumped weight is assigned to the lower resonant frequencies because considered. l o we r m o d a l i n f o r m a t i o n i s r e l i a b l e a n d m o r e The actual restoring force of an oil damper consists important. of the resisting force generated by oil flows and the viscoelastic shear force from the sealing polymer. In 4. Data Analysis and Damage Reasoning experiments, the restoring force is both amplitude 4.1 Earthquake Records and frequency dependent. However, in a numerical The structural responses induced by earthquakes simulation, an oil damper model is simplified, which is before and after the 3/11 Earthquake were collected by represented by link elements with the type of damper using the monitoring system; the collected data provide in SAP2000. The mathematical model for describing valuable information about the state of the building. the behavior of oil dampers is given through the The dataset can be divided into two categories; one following nonlinear force-velocity relation: is the building equipped with dampers and the other without dampers. The foreshock record of March 11, 2011 and the aftershock record of April 7, 2011, which represents each category, respectively, are used where F(t) is the restoring force developed by the for identification and analysis. No measurement was damper; u(t) is the relative displacement between the recorded during the 3/11 Earthquake. Thus, the ground oil piston and the abutment; u(t) is relative velocity motion captured by an observation station is used as between the oil piston and the abutment; C is the the ground excitation for the simulation of the main damping coefficient; is the exponent whose value is shock on March 11, 2011. determined using the piston head orifice design; and 4.2 Foreshock on March 9, 2011 sgn[·] is the signum function. The building equipped with dampers experienced an earthquake two days before the main shock. The ground motion of March 9, 2011 had a PGA of 32 gal in the EW direction and a PGA of 26 gal in the NS direction. The two-step identification procedure was performed to build the numerical model for nonlinear dynamic analysis. Fig.8.(a) shows the acceleration responses on the fourth floor of observation and the simulation in both directions. Fig.8.(b) demonstrates the transfer functions between the fourth and the first floors in the frequency domain. Fig.8.(c) compares the maximum acceleration on each floor of observation and simulation. The acceptable agreement between the measured and simulated responses verified the identified numerical model. Fig.7. FEM Model of Building with Dampers 4.3 Aftershock on April 7, 2011 After the 2011 Great East Japan Earthquake, a large 3.2 Identification Results aftershock struck the building in the middle of the For t h e e st a bl i she d n um e r i c a l m o de l , som e night on April 7, 2011, which was almost one month parameters remain undetermined. Based on the after the main shock. The seismic intensity of this measured response data of the building, two steps aftershock was the same as that of the main shock in are needed to identify the unknown parameters. The Sendai City. This aftershock damaged some buildings first step is to build the mathematical model of the oil in Sendai City, which were safe during the main shock. dampers. The monitoring system collects the relative Fortunately, this aftershock caused no further damage 184 JAABE vol.14 no.1 January 2015 Liyu Xie (a) Time History of Acceleration (a) Time History of Acceleration (b) Transfer Function (b) Transfer Function (c) Maximum Acceleration on Each Floor (c) Maximum Acceleration on Each Floor Fig.9. Comparison of the Observation and Simulation Fig.8. Comparison of the Observation and Simulation (April 7, 2011) (March 9, 2011) JAABE vol.14 no.1 January 2015 Liyu Xie 185 to the administration building. The ground motion of April 7, 2011 had a PGA of 176 gal in the EW direction and a PGA of 289 gal in the NS direction. The main shock on March 11, 2011 devastated the dampers on the first floor and disabled the dampers on the third and fourth floors because of oil leakage, whereas the main frame of the building had no structural damage. No retrofitting measurement was conducted before April 7, 2011. The numerical model is altered to consider the change of damper configuration. The oil dampers of the first, third, and fourth floors were removed from the model in both directions. The mass distribution of this building in the vertical direction was also changed because most of the bookshelves that fell over during the 3/11 Earthquake remained on the floor on April 7. The identification was conducted to rebuild the numerical model for the case of April 7, 2011. Fig.9. shows the acceleration responses that were recorded and simulated in the time domain of the fourth floor as well as the transfer functions in the frequency (a) Time History of Acceleration domain and the maximum acceleration of each floor. 4.4 Simulation of Main Shock The acquisition system did not work during the 3/11 Earthquake because of the power blackout. Thus, the numerical model of the steel building could not be built through direct identification by using response measurement. However, the numerical model for the main shock can be established by combining the identification results of the foreshock and aftershock. The mass distribution of the building during the main shock is the same as that of the undamaged building before the main shock. It is assumed that the dampers on the first, third, and fourth floors were not in operation from the beginning on March 11. The identified mathematical models of the oil dampers in the case of aftershock are used to simulate the velocity- force behavior of dampers during the 3/11 Earthquake. In the simulation of the main shock, the ground motion captured by an observation station was used as the ground excitation. This step aimed to simulate the dynamic responses of the building and predict the maximum displacement of the dampers during the main shock. (b) Transfer Function The acceleration responses on the fourth floor simulated for the main shock are plotted in Fig.10.(a). Fig.10. Simulated Results of Earthquake (March 11, 2011) The transfer functions between the fourth floor and the ground excitation are plotted in Fig.10.(b). The transfer functions from the first floor to the eighth floor in the NS direction are shown in Fig.11. for a comparison between the dynamic properties of the foreshock, the main shock, and the aftershock. The first period of aftershock is larger than that of foreshock because of the absence of the dampers on the first, third, and fourth floors. The maximum magnification factor in the first period of the aftershock is higher than in the other two cases. These facts reveal the effectiveness of the dampers in adding damping as well as suppressing building vibration. Fig.11. Comparison of the Transfer Function 186 JAABE vol.14 no.1 January 2015 Liyu Xie 4.5 Limit State of Oil Dampers Buildings equipped with fluid oil dampers have not been subjected to intensive earthquakes. In addition, no data or failure events of oil dampers during catastrophic events were available until the oil dampers of this steel building were devastated during the 3/11 Earthquake. Even in experiments, the limit states of the dampers were seldom tested. To evaluate the near collapse state of passively controlled structures, Miyamoto et al. (Miyamoto et al., 2010) conducted a comprehensive experimental investigation to determine the limit states of a viscous damper and established a detailed mathematical model of a viscous damper that considers the limit states of dampers. Miyamoto Fig.12. Diagrammatic View of the Damper's Limits examined several limited states in the laboratory. The two most common types are the force limit state and the displacement limit state. The force limit state can occur when a damper is subjected to a large-velocity (force) pulse within the stroke limit and when one of the mechanical parts of the dampers reaches the corresponding material limit. The displacement limit state can occur when the stroke limit in extension and retraction is reached. During the 3/11 Earthquake, the dampers on the first floor were damaged severely and the abutments were wrecked. The damage description indicates that both the stroke and cushion limits of the pistons seemed to have been reached, and the central cylinders were pounded against the abutment repeatedly until the abutment connection and oil dampers were broken. Therefore, the displacement limit state is assumed to be the damage scenario of the oil dampers. In this study, the damage process of oil dampers was replayed based on identification and simulation. According to the design book, the designed stroke limit displacement of the damper for the first floor is 16 mm (pink line in Fig.13.) and 8 mm for the third to eighth floors. In addition to damper protection, a cushion limit was also designed with 8 mm (red line) for the first floor and 5 mm for the third to eighth floors. The damper reaches the displacement limit state and the cylinder pushes against the abutment when the stroke and cushion limits run out. If the cylinder does not retract and continues to push against the abutment, the relative piston-cylinder velocity drops to zero. Thus, the restoring force of oil dampers becomes zero, and the V-type steel brace and the abutment on the floor carry the resistant force caused by restraining the interstory drift beyond the stroke and cushion Fig.13. Displacement of Dampers During the 3/11 Earthquake limits. Once the dampers reach the destructive limit displacement, as depicted in Fig.12., the accumulated out and the oil leaked during the 3/11 Earthquake. A deformation energy is abruptly released to break the reasonable assumption is that the oil dampers on the abutment connection and oil dampers. However, the fourth floor were in critical condition and were about difficulty lies in determining the destructive limit to reach the destructive limit displacement. Based on displacement of the abutment. the simulation results of the fourth floor oil dampers, Considering the situation of dampers on the fourth the ratio of the maximum interstory drift to the sum of floor, the oil container, piston, and abutment were not stroke and cushion limits can be calculated and used damaged. However, the sealing material had worn as a normalized index to determine the deformation JAABE vol.14 no.1 January 2015 Liyu Xie 187 limit of other floors (2.6 for the fourth floor). With this displacement/velocity capacity under the considered normalized index applied to the first floor, the destructive earthquake level. Improper design of dampers limit displacement was determined to be 59.8 mm, which should be avoided. Dampers are not made to survive is shown as black dashed lines in Fig.13. The simulated catastrophic earthquakes. The limit states of dampers interstory drift of the damper on the first floor in the EW should be considered during design, especially when direction on March 11, 2011 is plotted in Fig.13. The earthquakes beyond the considered level are likely to dampers of the first floor were most likely destroyed in shake the passively controlled structure. about 90 seconds during the 3/11 Earthquake. The failure event has prompted researchers and Of course, there are some flaws in this damage engineers to review the design philosophy of passively simulation. Continuous impacts of the central cylinder controlled buildings. Considering the limit state of against the abutment caused the U-type abutment to the dampers during the design is necessary at this open. The number of collisions between the central point because future catastrophic events may damage cylinder and the abutment before the breakage of the the dampers. This study will benefit the earthquake abutment and oil dampers cannot be determined. In engineering community. this study, the breakage is assumed to be caused by the impact of maximum displacement. Furthermore, the References 1) Cao, M., Tang, H., Funaki, N. and Xue, S. (2012) Study on a real mathematic model of the oil damper does not incorporate 8F steel building with oil damper damaged during the 2011 Great the limit state of dampers. Thus, the maximum interstory East Japan Earthquake. 15th World Conference on Earthquake drift in the simulation is not the same as in the real case. Engineering. Lisbon, Portugal. 2) Chang, K. C., Soong, T. T., Oh, S. and Lai, M. L. (1995) Seismic 5. Conclusion and Discussion behavior of steel frame with added viscoelastic dampers. Journal of structural engineering, 121(10), pp.1418-1426. This paper reports the first failure event of oil 3) Chang, K. C., Tsai, M. H. and Lai, M. L. (2001) Shaking table dampers during the 2011 Great East Japan Earthquake. study of a 2/5 scale steel frame with new viscoelastic dampers. Rebuilding the numerical model of oil dampers and the Structural Engineering and Mechanics, 11(3), pp.273-286. steel structure is possible through a monitoring system 4) Chang, K. and Lin, Y. (2004) Seismic response of full-scale and two-step identification. Based on the identification structure with added viscoelastic dampers. Journal of Structural Engineering, 130(4), pp.600-608. and simulation, the damage process of the oil dampers 5) Funaki, N., Kang, J. and Kawamata, S. (2001) Vibration response was investigated. of a three-storied full-scale test building passively controlled by The main cause of the damper failure was the liquid dampers sealed by viscoelastic material. 16th International insufficient stroke range of pistons. The total allowable Conference on Structural Mechanics in Reactor Technology. m o v i n g r a n g e o f o i l d a m p e r s wa s 2 4 m m f o r t h e Washington DC, USA. 6) Ji, X., Hikino, T., Kasai, K. and Nakashima, M. (2013) Damping dampers on the first floor in one direction (the sum identification of a full-scale passively controlled five-story steel of the 16 mm stroke limit and 8 mm cushion limit). building structure. Earthquake Engineering and Structural Thus, the interstory drift angle of the first floor was Dynamics, 42(2), pp.277-295. 0.3% when the damper reached the displacement 7) Kasai, K., Ito, H., Ooki, Y., Hikino, T., Kajiwara, K., Motoyui, limit. Furthermore, the elastic limit of the interstory S., Ozaki, H. and Ishii, M. (2010) Full scale shake table tests of 5-story steel building with various dampers. Proceedings of the drift angle in the Japan Steel Building Code is 0.5% 9th U.S. National and 10th Canadian Conference on Earthquake for minor earthquakes. The authors can infer that the Engineering. Toronto. original design of this administration building intended 8) Kasai, K., Mita, A., Kitamura, H., Matsuda, K., Morgan, T. A. to use oil dampers to control structural responses and Taylor, A. W. (2013) Performance of seismic protection to minor earthquakes while the structure responds technologies during the 2011 Tohoku-Oki Earthquake. Earthquake Spectra, 29(s1), S265-S293. elastically. However, the unprecedentedly intense 9) Kawamata, S., Funaki, N. and Itoh, Y. (2000) Passive control of earthquake caused large interstory drift, which the oil building frames by means of liquid dampers sealed by viscoelastic dampers could not withstand and consequently broke material. 12th World Conference on Earthquake Engineering. the dampers. Auckland, New Zealand. The limit states of oil dampers were studied based on 10) Lai, M. L., Chang, K. C., Soong, T. T., Hao, D. S. and Yeh, Y. C. (1995) Full-scale viscoelastically damped steel frame. Journal of the damage scenario. The oil dampers experienced the Structural Engineering, 121(10), pp.1443-1447. displacement limit state when the allowable limit was 11) Miyamoto, H., Gilani, A. S., Wada, A. and Ariyaratana, C. (2010) surpassed and the central cylinder pushed against the Limit states and failure mechanisms of viscous dampers and the abutment. In the future, more experiments should be implications for large earthquakes. Earthquake Engineering and conducted to determine all the possible limit states of Structural Dynamics, 39(11), pp.1279-1297. 12) Soong, T. T. and Spencer Jr, B. F. (2002) Supplemental energy oil dampers, and a numerical model that incorporates dissipation: state-of-the-art and state-of-the-practice. Engineering the limit states can be established to test the collapse Structures, 24(3), pp.243-259. state under intensive earthquakes. 13) Symans, M. D., Charney, F. A., Whittaker, A. S., Constantinou, Passive energy dissipation devices are deployed M. C., Kircher, C. A., Johnson, M. W. and McNamara, R. J. to reduce the inelastic energy dissipation demand (2008) Energy dissipation systems for seismic applications: current practice and recent developments. Journal of Structural induced on a structure, especially requested by a Engineering, 134(1), pp.3-21. major earthquake. The passive devices shall have a 188 JAABE vol.14 no.1 January 2015 Liyu Xie

Journal

Journal of Asian Architecture and Building Engineering – Taylor & Francis

Published: Jan 1, 2015

Keywords: oil damper; simulation; steel building; damage; the 2011 Great East Japan Earthquake

There are no references for this article.