CTBUH Technical Paper

Transcription

1 CTBUH Technical Paper Subject: Paper Title: Author(s): Affiliation(s): Structural Engineering Control Effect of Hydraulic Dampers Installed in High-rise Building Observed during Earthquakes Shimizu, Kan Orui, Satoshi Kurino, Haruhiko Omika, Yukihiro Koshika, Norihide Kajima Corporation, Tokyo, Japan Publication Date: 8 Original Publication: CTBUH 8th World Congress, Dubai. March 3-5, 8. Paper Type:. Book chapter/part chapter. Journal paper 3. Conference proceeding 4. Unpublished conference paper 5. Magazine article 6. Unpublished Council on Tall Buildings and Urban Habitat/Author(s)

2 Control Effect of Hydraulic Dampers Installed in High-rise Building Observed during Earthquakes Kan Shimizu, Satoshi Orui, Haruhiko Kurino, Yukihiro Omika 3 and Norihide Koshika 4 Senior Research Engineer, Supervisory Research Engineer, 3 Standing Advisor, 4 General Manager, Kajima Corporation KI Building 6-5-3, Akasaka, Minato-ku, Tokyo, Japan kan-shimizu@kajima.com, orui@kajima.com, kurino@kajima.com, omika@kajima.com, koshika@kajima.com kan-shimizu@kajima.com orui@kajima.com Kan Shimizu Kan Shimizu received the M. S. degree in civil engineering from Waseda University, Tokyo, Japan in 997. He is currently a senior research engineer in Kobori Research Complex of Kajima Corporation. He has been involved with about twenty high rise-buildings in Japan, planning the structural control devises and conducting seismic response analyses. Haruhiko Kurino Haruhiko Kurino received the M. S. and D.Eng. degree in structural engineering from University of Tokyo, Tokyo, Japan in 99 and 4, respectively. He is currently a supervisory research engineer in Kobori Research Complex of Kajima Corporation. He has studied structural control algorithms and developed hydraulic control devices, including a semi-active oil damper which has been widely used in high-rise buildings in Japan. CTBUH 8th World Congress 8

3 Control Effect of Hydraulic Dampers Installed in High-rise Building Observed during Earthquakes Kan Shimizu, Satoshi Orui, Haruhiko Kurino, Yukihiro Omika 3 and Norihide Koshika 4 Senior Research Engineer, Supervisory Research Engineer, 3 Standing Advisor, 4 General Manager, Kajima Corporation KI Building 6-5-3, Akasaka, Minato-ku, Tokyo, Japan kan-shimizu@kajima.com, orui@kajima.com, kurino@kajima.com, omika@kajima.com, koshika@kajima.com Abstract This paper estimates the dynamic characteristic, especially damping, of an actual high-rise building with two kinds of oil dampers, conventional passive oil dampers and semi-active switching oil dampers, from records of the building s responses during large earthquakes. It also assesses the effect of the semi-active oil dampers by comparing the estimated damping with that predicted by a seismic design model. The building is located in Niigata prefecture, Japan, which has recently been subjected to three large earthquakes: the Mid Niigata prefecture Earthquake in 4, the Noto Hanto Earthquake in 7 and the Niigataken Chuetu-oki Earthquake in 7. During these earthquakes, valuable records were obtained of the responses of the building and behavior of the semi-active oil damper. From the records, equivalent damping ratios are estimated at 6.4% to 6.7% in the transverse direction in which the semi-active oil dampers are applied and 4.3% to 4.8% in the longitudinal direction in which the passive oil dampers are applied. The high damping property of the structural controlled building is also verified. The damping provided by the semi-active oil dampers is estimated at about 5.5%, which was about.7 times that predicted by a seismic design model with passive oil dampers. Thus, the effect of the semi-active oil dampers is also verified. Keywords: structural control, semi-active oil damper, damping, observed record, earthquake Introduction In the last two decades a lot of research and application has been carried out on structural control devices (Spencer, 3, etc.). Recently many high-rise buildings in Japan are being equipped with some kind of structural control device. However, there are few seismic response records from buildings equipped with these device. Therefore, their effects in actual buildings in actual earthquakes have been insufficiently verified. A high-rise building with two kinds of oil dampers, semi-active and passive, recently experienced three large earthquakes and valuable records of its response and the behavior of the semi-active oil dampers were obtained. This paper reports the high damping characteristics of this building and the effects of the semi-active oil dampers. Outline of high-rise building The building was constructed in 3, in Niigata prefecture. Figure shows its exterior. It is 4m high and has 3 floors above ground and one underground. The lower floors are used as offices and the upper floors are used as hotel, as shown in Figure. Columns are concrete-filled steel tubes and the beams are steel. Oil dampers are mainly designed to reduce building responses under severe earthquakes. However, uncomfortable transverse vibrations were expected during strong winds because of its slender shape. Therefore, the dampers in the transverse direction were upgraded to semi-active oil dampers, which also work for very small vibrations and absorb twice as much energy as conventional oil dampers. Figure. Exterior of the building ~9F ~8F ~F Semi-active Oil Damper Passive Oil Damper 5F 46.4m Figure 3. Plan Semi-active Oil Damper ~5F ~4F Figure. Section ~8F ~9F 3m Hotel Office CTBUH 8th World Congress 8

4 Figure 3 shows the distribution of the dampers. 7 semi-active oil dampers are set in the transverse direction and 4 passive oil dampers are set in the longitudinal direction. Outline of semi-active oil damper Figure 4 shows the semi-active oil damper installed in the building. One of the features of this damper is that the whole system which consists of a controller and an oil damper equipped with sensors is closed as shown in Figure 4. This makes it as easy to use this semi-active oil damper as it is use to a conventional oil damper. Table shows its specifications. Controller changed to a very small value (C min ) and the dashpot absorbs the energy accumulated in the spring. When the force decreases to point C, C(t) is changed to C max again and the damper continues this cycle. The force-displacement relation of a semi-active oil damper under this control law is shown in Figure 7. The force-displacement relation of a passive oil damper with an optimum damping coefficient is also shown in Figure 7. These areas show the energy absorption capacity of the damper. The semi-active oil damper can absorb twice as much energy as the passive oil damper. Brace Column Beam Brace, Damper Damper Damper Figure 5. Mechanical model of oil damper installed in a building F F F FB B B B Table. Specifications of semi-active oil damper Item Oil Damper Figure 4. Installed semi-active oil damper Maximum design force F max Relief force F R Maximum piston stroke Stiffness k d Size Power consumption Specification,5kN,3kN mm 5MN/m 37mm,435mm Approximately 5W Oil dampers are usually installed in the inter-story spaces of buildings with braces as shown in Figure 5. Their mechanical model is described as a Maxwell model. Under the constraint of the Maxwell model, a conventional passive oil damper with a linear damping coefficient behaves like a spring when the damping coefficient is too large and doesn t generate any force when it is too small. Therefore, there is an optimum damping coefficient that produces the maximum energy absorption capacity for the passive oil damper. On the other hand, the semi-active oil damper s control law maximizes the energy absorption capacity under the constraint of the Maxwell model by switching the damping coefficient, as described in Kurino et al. (3). Figure 6 show the behavior of the Maxwell model and its elements under this control law. The damping coefficient (C(t)) is usually kept large (C max ). When vibration starts C(t) is kept to C max from point A to B, so the dashpot doesn t move and the spring accumulates energy. At point B, at which the velocity changes direction, C(t) is F -F A Cm ax Cm ax D Semi-active oil damper Cm in Cm ax Cm in k C C C x xk A A Cm ax D D amaxwell model bspring element cdashpot element Figure 6. Behavior of semi-active oil damper k F k Cmax k Damper characteristics and performance have been verified through full-scale device tests (Kurino et al., 3), and forced vibration tests (Tagami et al.,, Shimizu et al., 4). It is now being applied or planned for more than buildings and this number will increase. Observed response during earthquake Acceleration sensors were set on the roof floor mainly to observe the building s response under strong winds such as typhoons. The 5 th floor semi-active oil damper s force and stroke were also observed to check its behavior. On October 3 4, the building was subjected to the Mid-Niigata Prefecture Earthquake whose magnitude was 6.8 and epicentral distance was about 73km. In order to estimate the dynamic characteristics of the building from records of after Cmin Cmin Cmax Passive oil damper with optimum C Figure 7. Force-displacement relation of dampers under harmonic excitation x CTBUH 8th World Congress 8

5 shocks, an acceleration sensor was quickly added to the first basement floor. Records of after shocks in November 8 were thus obtained. There were two more large earthquakes in 7. The first was the Noto Hanto Earthquake, whose magnitude was 6.9 and epicentral distance was about km. The next was the Niigataken Chuetu-oki Earthquake, whose magnitude was 6.8 and epicentral distance was 57 km. During this earthquake, the maximum damper force reached 64kN, which was the maximum force of this damper in an actual building. Acc.(cm/s ) Acc.(cm/s ) (a) Roof floor acceleration (Longitudinal direction) structural control device in an actual building are rare. Figure 8 shows the records during the Niigataken Chuetu-oki Earthquake. The force-displacement relation shown in Figure 8(e) shows the typical parallelogram shape produced by this control law. Table shows the maximum acceleration of the roof floor and the first basement floor and the maximum force of the semi-active oil damper. Figure 9 shows the location of the building and the epicenters. Table. Maximum acceleration of records Earthquake BF Long. (cm/s ) BF Tran. (cm/s ) RF Long. (cm/s ) RF Tran. (cm/s ) Mid Niigata (main) Mid Niigata (after) Force (kn) Noto Hanto Chuetu-oki (b) Roof floor acceleration (Transverse direction) Acc.(cm/s ) Building (c) First basement floor acceleration (Longitudinal direction) Acc.(cm/s ) (d) First basement floor acceleration (Transverse direction) (e) Force-stroke relation of 5 th floor semi-active oil damper Figure 8. Records of the Niigataken Chuetu-oki earthquake in 7 These records are very valuable because such large responses of a structural controlled building and a Figure 9. Location of building and epicenters Estimate of building s dynamic characteristics In this section, dynamic characteristics such as natural frequency and equivalent damping ratio are estimated. In particular, damping is discussed and the effect of dampers is verified. Only the first mode, which is the dominant vibration mode of the building, is considered. Before discussing damping during earthquakes, damping during microtremors is demonstrated, because the vibration level of microtremors is small enough to neglect the effect of dampers. Equivalent damping ratios of the building during microtremors were estimated by the Random Decrement (RD) technique (Tamura et al., 993). Time histories of free vibrations obtained by the RD technique are shown in Figure. Estimated damping ratios were.97% in the transverse direction and.% in the longitudinal direction. CTBUH 8th World Congress 8

6 Freq.(Hz) Freq.(Hz) Acc.(cm/s ) Dampig Ratio Damping Ratio RF Disp.(cm) (a) st mode natural frequency (Longitudinal direction) (a) Transverse direction Identified (b) Longitudinal direction Figure. Random decrement and free vibration decay response of microtremer Observed record Figure. Comparison between the identified record and observed record (The Niigataken Chuetu-oki earthquake in 7) RF Disp.(cm) (b) st mode natural frequency (transverse direction) 3 RF Disp.(cm) (c) st mode damping ratio (Longitudinal direcrion) 4 h=.97% h=.% RF Disp.(cm) (d) st mode damping ratio (transverse direcrion) Figure. Estimated dynamic characteristics Microtremor Microtremor On the basis of damping of conventional steel structures shown in Satake, 3, these results were inferred as damping without dampers. Next, the dynamic characteristics under earthquake vibrations are estimated. First, the time history of the roof floor acceleration is filtered to extract the first vibration mode. Then a Single Degree of Freedom (SDOF) system that makes the mean-square error the smallest is identified as shown in the following equation x( t) 4hfx ( t) x( t) X BF ( t) T () ( X ( t) x( t)) dt min RF RF where x(t) is the response of a SDOF system, h is the damping ratio of a SDOF system, f is the natural frequency of a SDOF system, X B F is the filtered acceleration record at the first basement floor, T is the duration of the earthquake, X RF is the filtered acceleration record at the roof floor, and RF is a first mode participation function at the roof floor, for which the participation factor of the seismic design model shown in Figure 4 is used. As a result, the natural period and the damping ratio of the SDOF system are identified as those of the building. In Figure, the dotted line shows the filtered acceleration record at the roof floor, and the full line shows the identified time history of the record of the Niigataken Chuetu-oki Earthquake in 7. These lines show a significant match. Figure shows the identified natural frequencies and damping ratios of each record. The amplitudes are the maximum displacement of the roof floor. Identified damping ratios are from 6.4% to 6.7% in the transverse direction and from 4.3% to 4.8% in the longitudinal direction. Compared with the microtremor case shown by the full line in Figure (c), (d), the high damping characteristic of the building with oil dampers are verified. We also used another method to check the reliability of the identified dynamic characteristics. The previous method depends on a physical quantity and uses an equation of motion but this method doesn t depend on a physical quantity. First, the transmissibility is identified as single-input-single-output ARX model: input is filtered acceleration at the first basement floor and output is filtered acceleration at the roof floor. Then the natural frequency and corresponding damping ratio are calculated from the pole of the identified ARX model (AFAK, 99). The identified results are 6.3% to 7.3% in the transverse direction and 4.4% to 5.5% in the longitudinal direction. The difference between the two methods is less than 5%. Here, the observed effect of the dampers is compared with that expected by the seismic design model. Figure 3 shows the seismic design model, in which each column and beam is modeled as a beam element. Bending deformation, shear deformation and axial deformation are considered at columns and bending and shear CTBUH 8th World Congress 8

7 deformations are considered at beams. Panel elements at which only shear deformation is considered were set at each column/beam connection. Oil dampers are modeled as a Maxwell model, described in Figure 5. Figure 4 shows the mode shape of each direction. dampers is almost the same as that predicted by the seismic design model and the damping added by the semi-active oil dampers are almost.7 times that predicted by the seismic design model with passive oil dampers. Therefore, the effect of semi-active oil dampers is clearly verified. Floor 3 Figure 3. Seisimic design model : st mode : nd mode :3 rd mode Floor 3 Additional discussion (Wind case) This section demonstrates the effect of the semi-active oil dampers during strong winds. In 4, typhoons struck Japan. This was a record since the Japan Meteorological Agency started to take statistics in 95. One of them, Typhoon 6, had a maximum daily wind velocity of 9.5m/s, which was recorded at Niigata Local Meteorological Observatory. This was the maximum daily wind velocity in 4 in Niigata, and it continued for about hours. Damping in the transverse direction during this typhoon was also estimated by an RD technique as 5.3%. Figure 5 shows the force-stroke relation of the 5 th floor semi-active oil damper during the typhoon. It also shows a typical parallelogram shape from a very small stroke, under.5mm. Next the without damper time history is estimated. The absolute acceleration response, x w, without semi-active oil damper is calculated by equation () considering only the first mode. x w H H s s ( if ) H ( if ) x w ( if ) Hw( if ) s f / f hs f / f i f / f h f / f i where H s (if) is the st mode transfer function with semi-active oil damper, H w (if) is the st mode transfer function without damper, x s is the observed record, f is the st natural frequency, h s is the st equivalent damping ratio with semi-active oil damper, and h w is the st equivalent damping ratio without damper. w () (a) Longitudinal direction (b) Transverse direction Figure 4. Mode shape (Participation function) Complex eigen value analysis was conducted to estimate damping added by passive dampers. Damping factors of dampers were set to 5MNs/m. To accurately estimate the effect of dampers, damping of the structure isn t set. The complex eigen value analysis yielded values of 3.% for each direction. By subtracting damping estimated during microtremors from damping estimated during earthquakes, the damping added by the dampers in an actual building is obtained. The values thus obtained were 5.4% to 5.7% in the transverse direction and 3.% to 3.7% in the longitudinal direction. Comparison of the results verified that the damping added by the passive Figure 5. Force-stroke relation of 5 th floor semi-active oil damper under typhoon Results estimated by the RD technique were used as parameters:.3hz for f,.53 for h s, and.97 for h w. Figure 6 compares the top floor acceleration time history of Without damper estimated by equation () against the observed acceleration record. Without CTBUH 8th World Congress 8

8 semi-active oil damper, the maximum response of the building would double and uncomfortable vibration would continue for a long time. AFAK, E. (99). Adaptive Modeling, Identification, and Control of Dynamic Structural Systems : Theory. Journal of Structural Engineering, ASCE, Vol.5, No., pp , November 99 (a) With Damper (Observed) (b) Without Damper (Simulated) Figure 6. Roof floor acceleration time history (transverse direction) Conclusion Three large earthquake records were obtained from a high-rise building equipped with semi-active oil dampers in the transverse direction and conventional passive oil dampers in the longitudinal direction. Estimated damping verified the high damping property of the building with oil dampers. Equivalent damping ratios were from 6.4% to 6.7% in the transverse direction and from 4.3% to 4.8% in the longitudinal direction. The effect of the dampers was also discussed with reference to a seismic design model. By comparing the damping ratio added by the dampers estimated from the earthquake records with that estimated from the seismic design model with passive oil dampers, it was verified that the passive oil dampers showed almost the same results as expected, and the semi-active oil dampers showed high performance in adding damping, which was about.7 times as large as that expected by a passive oil damper. References SPENCER, Jr. B. F., and NAGARAJAIAH, S. (3). State of the Art of Structural Control. Journal of Structural Engineering, ASCE, Vol. 9 No.7, pp , July 3 KURINO, H., TAGAMI, J., SHIMIZU, K., and KOBORI, T. (3). Switching Oil Damper with Built-in Controller for Structural Control. Journal of Structural Engineering, ASCE, Vol. 9 No.7 pp TAGAMI, J., KOSHIDA, H., KURINO, H., SUGIYAMA, T., SUWA, M., and MORI, F. (). Forced Vibration Test of an -story Building with Semi-active Switching Oil Damper. Proceedings of the 3rd World Conference on Structural Control, Como, Italy. April 7-,. Vol. pp.75-8 SHIMIZU, K., YAMADA, T., TAGAMI, J., and KURINO, H. (4). Vibration Tests of Actual Buildings with Semi-active Switching Oil Damper. Proceedings of the 3th World Conference of Earthquake Engineering, Vancouver, B.C., Canada, August -6, 4. Paper No. 53 TAMURA, Y., SASAKI, A., and TSUKAGOSHI, H. (993). Evaluation of Damping Ratios of Randomly Excited Buildings Using the Random Decrement Technique. Journal of Struct. Constr. Engng, AIJ; 454 pp. 9-38, in Japanese SATAKE, N., SUDA, K., ARAKAWA, T., SASAKI, A., and TAMURA, Y. (3). Damping Evaluation Using Full-Scale Data of Building in Japan. Journal of Structural Engineering, ASCE, Vol. 9 No.4, pp , April 3 CTBUH 8th World Congress 8