CURRENT STATUS OF BUILDING PASSIVE CONTROL IN JAPAN

Transcription

1 The 14 th World Conference on Earthqake Engineering CURRENT STATUS O BUILDING PASSIVE CONTROL IN JAPAN K. Kasai 1, M. Nakai 2, Y. Nakamra 3, H. Asai 4, Y. Szki 5, and M. Ishii 6 1 Professor, Strctral Engineering Research Center, Tokyo Institte of Technology, Japan 2 General Manager, Advanced Strctral Engineering Department, Takenaka Corporation, Japan 3 Depty General Manager, Institte of Technology, Shimiz Corporation, Japan 4 Chief, Technical Research Institte, Obayashi Corporation, Japan 5 Chief, Advanced Strctral Engineering Department, Takenaka Corporation, Japan 6 Senior Engineer, Strctral Engineering Department, Nikken Sekkei Ltd., Japan kasai@serc.titech.ac.jp, nakai.masayoshi@takenaka.co.jp, ytaka.nakamra@shimz.co.jp, asai.hidekats@obayashi.co.jp, szki.yoske@takenaka.co.jp, ishiim@nikken.co.jp ABSTRACT : The present paper describes three key isses as follows: (1) A major research program on passive control of bildings: Shaking table tests sing E-Defense will be condcted in ebrary 29, for a fll-scale 5-story steel bilding with or withot dampers. (2) Code and specifications: Japanese code reqires nonlinear time history analysis for bildings with dampers, or energy-based analysis when sing steel dampers. Unlike the code rles leading to iterative design, JSSI specifications give a direct design method (DDM) for any target performance set by designers or clients. (3) Damage-ree Design: Under the government spport, a large team of researchers/designers has developed a damage-free design method sing dampers and frames of the so-called sper high strength steel. The method is an extended version of the DDM, and its inclsion in the ftre code is being investigated. KEYWORDS: Passive Control, JSSI Passive Control Manal, Bilding Standard Law, Shear Beam Model, Damage-ree Strctre Design, Sper High Strength Steel 1. INTRODUCTION Since the 1995 Kobe earthqake, the Japanese social desire for adopting passive control schemes has increased considerably. The schemes are typically sed for major bildings, and even for many small residential bildings, in order to better protect the bilding and its contents. In these schemes, the damper connected to the strctral frame dissipates the seismic inpt energy, thereby redcing the kinetic energy and vibration of the bilding. A variety of dampers are being prodced by more than twenty manfactrers and more than ten general constrction companies in Japan. Nmeros technical papers on passive control are also presented in varios Japanese symposims. With the above backgrond, Japan has prodced the largest nmber of passively-controlled bildings, and is believed to have condcted the most extensive research to realize varios control schemes. This paper explains the present and ftre of this Japanese technology, especially by referring to major activities related to research and design. It addresses the three key isses sch as; extraordinary experimental project to validate passive control performance sing different types of dampers, the state of the crrent code and specifications regarding passive control, and a new direction to promote the so-called damage-free design method even against the catastrophic earthqake. These three isses are discssed in the following sections. 2. VALIDATION O PASSIVE CONTROL TECHNOLOGY 2.1. l-scale Tests of 5-Story Bilding with s The first isse is validation of the passive control technology. Becase the history of passive control is short, the technology has never been attested nder the major and catastrophic earthqakes, while it is increasingly sed in Japan. Therefore, it is extremely important to validate this technology by realistic experiments, before

2 The 14 th World Conference on Earthqake Engineering occrrence of sch earthqakes. Prsant to this, a fll-scale bilding with dampers will be examined in ebrary and March 29 sing the E-Defense, the world s largest three-dimensional shaking table. As shown in ig. 1, the bilding is 5-story with two bays in each direction. De to the redction in bdget, the bilding is made smaller than originally planned and described elsewhere (Kasai et al. 27a, 28a, 28c). In spite, the test is still by far the largest and the most realistic, among those condcted for passively-controlled bildings. The plan dimension is 1m 12m, and total height from center line of the fondation beam is 16.3 m. Seismically active weight of the sperstrctre is 4,7 kn, inclding all strctral/non-strctral components and a portion of live load. The frame members of the sperstrctre consist of wide-flange beam sections of 4 mm deep, and sqare box colmn sections of 35 mm 35 mm. The expected steel yield strength for the beam and colmn shall be 358 MPa and 325 MPa, respectively. All the beam and colmn connections will be a flly-restrained type. The steel deck with concrete on top will be considered and flly composite beams will be created. Note that the beam flange is hanched to increase yield rotation and to delay onset of yielding (Kasai et al. 27a). Some stories of the bilding will be provided with glass crtain wall, pre-cast light-weight crtain wall, partitions with doors, several types of ceilings with sprinkler systems, and mechanical eqipment. The stdy on the behavior of non-strctral components is a part of the major US-Japan collaborative research program. The 5-story bilding will have 12 dampers (ig. 1) of the same type with three to for different sizes. The test will be repeated for different types. or major damper types are considered: they are steel, viscoelastic, viscos, and oil dampers, and their characteristics will be explained in Section 3. In order to assre performance of the damper to be sed in the bilding as well as to validate analytical model, dampers of three different sizes per each type (ig. 2) were dynamically tested at Tokyo Institte of Technology (Kasai et al. 28a, 28c). The damper capacities were in the range between 5kN and 15kN, and sinsoidal and random deformation tests were performed. Deformations of varios components of the damper-brace assembly were measred, and are sed to estimate properties of the analysis models Target Performance of Bilding Specimen The specimen is designed to be the high vale-added bilding whose strctral and non-strctral components are protected from a major seismic attack. The story drift angle is reqired not to exceed.5 (=1/2) rad. nder the so-called level 2 (design basis) earthqake, which is mch more stringent than.1 (=1/1) rad. sally considered for a conventional bilding. rthermore, nder the catastrophic earthqake of grond velocity 2.5 times larger, drift angles wold be abot.1 rad. and the frame wold remain almost elastic with no damage. The high performance has been predicted by extensive time history analyses, considering the for different types of dampers. The fll-scale experiments mentioned above will provide extremely important data to verify the analysis and the design method. Note also that a blind analysis contest, like the one in 27 for X Y ( X2-rame ) ig. 1 ll-scale 5-story Bilding with s (E-Defense Tests)

3 The 14 th World Conference on Earthqake Engineering Viscoelastic Oil Steel Viscos (US) ig. 2 ll-scale s Tested Prior to Bilding Experiments steel bilding collapse (Ohsaki et al. 28), will be held as the world-wide competition for the accracy of analytical prediction prior to the fll-scale experiment. 3. CURRENT CODE AND JSSI SPECIICATIONS 3.1. Japanese Code and JSSI Specifications The second isse is standardization of the technology and design method. ig. 3 shows design procedres of steel bildings stiplated by the Japanese code called Bilding Standard Law, and new options being proposed by Association of New Urban Hosing Technology (ANUHT), Japan. or the bildings exceeding 6 meters in height, the code reqires time history analyses, design review by the panel members, and special permission by the ministry of land, infrastrctre, transport, and torism (MLIT). As for the shorter bildings that constitte majority of the bilding stock in Japan, the designers are allowed to select one of the for methods shown in ig. 3. However, methods (1) and (2) are intended for conventional strctres, and are not sitable for those with dampers. Method (3) incldes the so-called energy-balance method that incldes only a case sing steel dampers. In this manner, the code is not yet providing a simplified method covering all typical damper types. Height 6m No (1) Allowable Stress and Ultimate Strength Method Yes (2) Response and Limit Capacity Method Alternative Methods (3) Other methods eqal or sperior to (1) or (2) (4) Time history analysis No reference to expected response. No concept of damper and response control. No concept of damper and response control. Seismic Isolation Energy Balance Method Advanced evalation scheme. Improved Response and Limit Capacity Method Damage-ree Design (Passive control system with elastic frame ) = options being proposed by ANUHT. ig. 3 Crrent Japanese Code Procedres and Proposed Additions

4 The 14 th World Conference on Earthqake Engineering 粘性ダンパー Viscos オイルダンパー Oil Viscoelastic 粘弾性ダンパー Steel(riction) 鋼材ダンパー riction Shear/low せん断, 流動抵抗 Resist. low 流れ絞り抵抗 Resist. Panel, 面型, Box, 箱型, Cylinder 筒型 Cylinder 筒型 α = C = C 1 or C 2 Shear せん断抵抗 Resist. Cylinder, Brace 筒型, 面型 Panel,, その他 etc. Axial/Shear 塑性履歴抵抗 Yielding Cylinder, 筒型 Brace, 面型 Panel,, その他 etc. = K ( ω ) + C ( ω ) = K f ( ) Slip Resist. Brace, Panel = K f () ig. 4 ive Types of s Considered by JSSI Manal On the other hand, an accrate and simple design procedre is available from specifications in Manal for Design and Constrction of Passively-Controlled Bildings that is pblished by Japan Society of Seismic Isolation (JSSI 23, 25, 27). The researchers in China also prodced Chinese translation of the manal (Jiang 28). The design procedre was originally proposed by Kasai et al. for elastoplastic damper and viscoelastic damper (Kasai et al. 1998). It was then extended to consider more damper types (ig. 4), a variety of frame configrations (ig. 5), higher mode effects, and other key items (Kasai et al., 25, 26b, 27b, 28b). The manal reflects sch research findings, and provides the design procedres for five different types of dampers (ig. 4). The dampers are briefly described below: The three damper types shown at left of ig. 4 are velocity-dependant. Viscos damper tilizes flow resistance of the polymer liqid. Its force is proportional to the fractional power of velocity, leading to the hysteresis loop of combined ellipse and rectangle. Oil damper tilizes flow resistance of the oil with low viscosity. The damper typically has a relief mechanism to switch viscos coefficient to a small vale when sbjected to a large velocity, making the hysteresis to switch from an elliptical shape to a rectangle shape. Viscoelastic damper tilizes moleclar motion of a polymer for energy dissipation. Hysteresis loop is an inclined ellipse, and the inclination angle and the fatness of the loop depend on the excitation freqency and the temperatre. The two damper types shown at right of ig. 4 are deformation-dependant. Steel damper tilizes yielding of steel material for energy dissipation. It shows a rond crve bonded by bi-linear lines, and can be analytically modeled by sing readily available constittive rles for steel materials. riction damper tilizes slip resistance between two metallic srfaces, or between metal and friction pad whose material is analogos to that Directly Connected System Wall Type Brace Type Shear Link Type Indirectly Connected System Std Type Bracket Type Connector Type rame Model, Kf Ka Special System Colmn Type Beam Type Toggle Type (a) Model Kd (b) Brace Model Kb ig. 5 Varios rame Configrations ig. 6 Shear Beam Modeling for lexre Behavior

5 The 14 th World Conference on Earthqake Engineering R a μ rame only (, ) R d (a) Constant S pa Case typically sed for a car brake. These elements mst be compressed sfficiently to prodce the friction force, and the hysteresis is of an ideal rectangle shape. Varios frame types crrently sed in Japan (ig. 5) are also considered. or design and efficient time history analyses, they are commonly transformed into a mlti-degrees-of-freedom (MDO) shear beam system sch as shown in ig. 6b, by sing a niqe transformation method proposed by Kasai and Iwasaki (26a). The method can create a reasonably accrate shear beam system, even for a frame developing considerable flexral deformation (ig. 6a). Based on these, and by reflecting as accrately as possible the difference of hysteresis loops among the five damper types, the design procedres for the mlti-story passively-controlled bilding were developed. They will be briefly explained below, by referring to a steel damper type as an example Performance Crves and Direct Design Method for Target Performance. R d (b) Constant S pv Case ig. 7 shows the performance crves representing the mlti-story bilding by an eqivalent single-degree-offreedom (SDO) system. The system may be considered as a single-story system having eqivalent frame, damper, and brace like the MDO system in ig. 6a. The crves show both displacement redction ratio R d and force (or acceleration) redction ratio R a, which are the vales of the peak responses normalized to those withot dampers (JSSI 23, 25, 27). The crves are obtained by expressing mathematically the effective vibration period and damping ratio in terms of the balance of the parameters mentioned above, and by combining them with the idealized smooth response spectrm. ig. 7a assmes the constant psedo-acceleration spectrm, as often considered in designing short bilding, and ig. 7b assmes the constant psedo-velocity response spectrm, typically considered for moderate to tall bildings. or the steel dampers considered, K a /K f and μ govern the response redction: the former is a ratio of the added component elastic stiffness to the frame elastic stiffness, and the latter the dctility demand to the added component, where added component is a series combination of the damper and the elastic brace. The crves clearly indicate, and promote nderstanding of, the strong effects of balance among the frame, damper, and brace. The crves also clearly indicate necessary stiffness ratio of damper or added component relative to the frame, once the target drift and desired dctility demand are specified. rthermore, the ratio can be applied to size the damper in the MDO system whose frame stiffness is explicitly known at each story. A special method is also available for assring reasonably niform story drift distribtions, even when the frame withot dampers tends to deform non-niformly (Kasai et al., 25, 26b, 27b, 28b). Note also that the crrent code provides methods (ig. 3) to check adeqacy of a given strctre, ths, the design procedres are iterational, involving redesign and rechecking. In contrast, the method introdced here does not only address wide range of passive control, bt also has a significantly different concept. It directly obtains the design soltion for a given target performance, and wold be more sitable for the performance-based engineering. The method, therefore, will be called the Direct Design Method for Target Performance (DDM). R a.5 μ ig. 7 Performance Crve for Steel Case rame only (, )

6 The 14 th World Conference on Earthqake Engineering Stress σ(mpa) 1, Level of σ YS65 y to HSA7 Achieve θ y = 8 1/6 HT62 1/75 6 SN49 HT44 1/1 4 SN4 1/ ig. 8 ε(%) Strain Sper High Strength Steel Considered Acceleration (cm/s 2 ) K-NET Oshika 23 (6.3) K-NET Hiroo 23 (6.) K-NET Chokbets 23 (6.3) K-NET Tokamachi 24 (6.2) K-NET Nagaoka-shisho 24 (6.1) K-NET Anamiz 27 (6.3) K-NET Kashiwazaki 27 (6.3) JMA Kobe 1995 (6.4) JMA Ojiya 24 (6.3) JMA Kawagchi 24 (6.3) Tokyo-Koenji (6.6),Artif. Tokyo Bay -Nihonbashi (6.5),Artif. Design Basis Eq.Level,Artif ξ = Period (sec) ig. 9 Catastrophic Earthqakes Considered 4. DEVELOPMENT O DAMGE-REE STRUCTURE AND DESIGN SPECIICATIONS 4.1. Crrent Development by ANUHT Committees The third isse is extension to damage-free strctre and corresponding design specifications. The Japanese government set p a major research/development project, targeting realization of rban infrastrctre that is free from damage, even nder the catastrophic earthqake. The project has been sbcontracted by ANUHT for the period of 25 to 29, and has organized a large team of researchers/designers from niversities, constrction companies, and design firms, in order to develop a new bilding system sing steel dampers and steel frames. The system is expected to be damage-free, against the earthqake of the JMA (Japanese Meteorological Agency) seismic intensity p to 6.4, which is eqivalent to the modified Mercalli intensity of abot X. The team investigated into the economical sper high strength steel that can be sed for the members of the frame. The newly developed steel called HSA7 has the strength more than 8 MPa and is considered to have adeqate properties, bt other sper high strength steel materials are also being investigated. By sing this material, the yield drift angle of the frame increases to 1/75 (=.133) rad. or more (ig. 8). As will be explained, the bilding drift angle will be kept within this limit by sing steel dampers. ig. 9 shows acceleration response spectra of the grond motions that were recorded dring the recent major earthqakes in Japan. The thick and smooth line indicates the design spectrm corresponding to times the Japanese design basis earthqake, and it shows the constant acceleration and velocity spectra of 1,8 cm/s 2 and 248 cm/s, respectively, nder damping ratio of.5. By converting the vales to initial damping of.2 typically considered for steel bilding, one obtains the constant vales of S pa = 2,32 cm/s 2 and S pv = 32 cm/s, which will be sed to prodce modified performance crves as described below: 4.2. Modified Performance Crves for Direct Design Method ig. 1 shows projection of the performance crve shown earlier (ig. 7). Advantage of this format is that the displacement and acceleration responses are not normalized, and story drift angle and maximm acceleration response are readily fond. Straight-line deformed shape of the bilding is assmed and eqivalent height is considered to be 2H/3, when estimating the drift angle θ. Note that the figre corresponds to a case of selected frame vibration periods T f and bilding height H. or other vales of T f and H, the responses shown mst be scaled by sing the rles indicated in ig. 1. As a next effort to make the crve more convenient, only the dctility demand μ = 5 will be selected. This is becase μ 5 leads to almost the smallest set of story drift and acceleration for a given K a /K f vale (igs. 7 and ig. 1a,b). The modified performance crves are given in ig. 1c and 1d. The crves can give the designers opportnities to explore combinations of frame and dampers. The conventional steel moment resisting frame tends to show the period T f.3h for a tall bilding, and increases p to.5h approximately, as bilding is shorter. The modified performance crve can

7 The 14 th World Conference on Earthqake Engineering S pa ( cm/s 2 ) μ S pa = 232 cm/s 2 25 (ξ =.2) rame only T f =.5H 1 3. H = 16 m (a) θ ( rad ) S pa ( cm/s 2 ) 25 S pv = 32 cm/s (ξ =.2) 2 μ rame only 15 T f =.5H H = 4 m (b) θ ( rad ) S pa ( cm/s 2 ) 25 Tf =.3 H.4 H.5 H (H = 16 m) S pa = 232 cm/s 2 (ξ =.2) 1 (when μ=5) 5. 5 (c) θ ( rad ) Constant S pa Case : or other S pa and H vales, θ and S pa will be β 1 β 2 and β 1 times, where β 1 =S pa /232 and β 2 =H/16. S pa ( cm/s 2 ) 25 2 Tf =.3 H 15.4 H 1.5 H (H = 4 m) (when μ=5) (d) θ ( rad ) Constant S pv Case : or other S pv and H vales, θ and S pa will be β 1 and β 1 /β 2 times, where β 1 =S pv /32 and β 2 =H/4. ig. 1 Modified Performance Crve for Convenient Evalation S pv = 32 cm/s (ξ =.2) propvide different options to meet the target drift angle, i.e., one can se a stiffer (i.e., short period) frame and softer (i.e., smaller) damper, or vice versa. Acceleration will differ between the options (ig. 1c, d) Example for Damage-ree Strctre Design A design of the damage free strctre will be briefly described. ig. 11 shows the acceleration spectra of 7 grond motions considered. The frame is 8-story and its height H = 33 m. The frame vibration period T f = 1.39 s,. Ths, constant psedo-velocity and T f.4h will be considered in the modified performance crve of ig. 1d. Note that the psedo-velocities of some grond motions exceed the code-specified vale of 32 cm/s nder damping ratio.2 (ig. 11). Among sch motions, an artificial motion that is 2.5 times the BCJ-L2 (Level 2 grond motion given by the Bilding Center of Japan) shows very smooth spectrm crve of 36 cm/s (ig. 11). Since the crve can represent somewhat conservative level for all the motions plotted, the design considers S pv = 36 cm/s, as well as the target drift angle of 1/75 (=.133) rad. mentioned in Sec Ths, β 1 = S pv /32 = will be sed for the performance crve in ig. 1d. The actal θ-vale will be β 1 times the vale in ig. 1d. Ths, for sing the performance crve in ig. 1d, one can consider.133/β 1 =.118 rad. as the fictitios target angle. Correspondingly, the crve for T f =.4H gives K a /K f 2.8 in ig. 1d. By converting the K a /K f vale of the SDO system to MDO system, one can obtain the reqired stiffness of the added component and damper (ig. 6). rom the target drift angle, one obtains the corresponding deformation of the added component, and dividing it by the reqired dctility demand μ = 5 gives the yield deformation, and conseqently the reqired yield force of the added component. ig. 12 shows performance of ths-designed strctre. The direct design method (DDM) for the target story drift appears to be reliable. Clearly, the damage-free strctre nder the catastrophic motion is feasible, as evidenced by the small drift and well-controlled performance shown in ig. 12.

8 The 14 th World Conference on Earthqake Engineering Velocity (cm/s) K-NET Chokbets NS 23 (6.3) K-NET Chokbets EW 23 (6.3) K-NET Nagaoka-shisho EW 24 (6.1) K-NET Anamiz W 27 (6.3) JMA Kobe 1995 NS BCJ-L2 (Level 2) 2.5,Artif. Koenji EW (6.6),Artif. Design Basis Eq.Level,Artif. 7 ξ = Period (sec) ig. 11 Velocity Spectra for Seven Grond Motions Story Chokbets NS Chokbets EW Nagaoka-shishoEW Anamiz EW Kobe NS BCJ-L2 2.5, Design Basis Artif. Eq. Koenji EW,Artif. 層せん断力係数 R=1/75 Story Story Drift Angle Dctility Demand ig. 12 Performance of Damage-free 9-story Bilding Designed by Using Performance Crve (No dampers at top story) 5. SUMARRY This paper has discssed three key isses regarding the crrent stats of passive control technology in Japan. They are; a major fll-scale experimental program on passive control of bildings, crrent sitation of the code and specifications on passive control, and new performance engineering and damage-free design sing high strength steel and dampers. ACKNOWLEDGEMENTS Prof. M. Midorikawa of Hokkaido Univ. kindly provided valable information. Mr. Y. Ogawa, Tokyo Institte of Technology, assisted with data processing and drawing. Their help is grateflly acknowledged. The athors also acknowledge contribtions from many members of varios committees of JSSI, E-Defense, and ANUHT. REERENCES Kasai, K.,, Y., and Watanabe, A. (1998). Passive Control Systems for Seismic Damage Mitigation, Jornal of Strctral Engineering, American Society of Civil Engineers, 124 [5], pp JSSI Manal, Design and Constrction Manal for Passively Controlled Bildings, Japan Society of Seismic Isolation, Tokyo, Japan, 1st Ed. 45 pages, 2nd Ed. 515 pages, 2nd Ed. 2nd Print 59 pages, 23, 25, 27 (in Japanese) Kasai, K. and Ito, H.(25). Passive Control Design Method Based on Tning of Stiffness, Yield Strength, and Dctility of Elasto-Plastic, J. of Str. and Constr. Eng., AIJ, No.595, Sep., pp (in Japanese) Kasai, K. and Iwasaki, K. (26a). Redced Expression for Varios Passive Control Systems and Conversion to Shear Spring Model, Jornal of Str. and Constr. Eng., AIJ, No.65, Jly, pp (in Japanese) Kasai, K. Minato, N., and Kawanabe Y. (26b). Passive Control Design Method Based on Tning of Eqivalent Stiffness of Visco-Elastic, Jornal of Str. and Constr. Eng., AIJ, No.61, Dec., pp (in Japanese) Kasai, K., Ooki, Y., Motoyi S., Takechi T., and Sato, E. (27a). E-Defense Tests on ll-scale Steel Bildings: Part 1 Experiments Using s and Isolators, ASCE Str. Congress, Long Beach, CA, May Kasai, K. Ogra, T,., and Szki, A. (27b). Passive Control Design Method Based on Tning of Eqivalent Stiffness of Nonlinear Viscos, Jornal of Str. and Constr. Eng., AIJ, No.618, Ag., pp (in Japanese) Sita, K., Yamada, S., Tada, M., Kasai, K., Matsoka, Y., Matsoka, Y., and Sato, E. (28). Reslts of Recent E-Defense Tests on ll-scale Steel Bildings: Part 1 Collapse Experiments on 4-Story Moment rames, ASCE Strctres Congress, Vancover, B.C., April Ohsaki, M., Kasai, K, Matsoka, Y., and Zhang, J. (28). Reslts of Recent E-Defense Tests on ll-scale Steel Bildings: Part 2 Collapse Simlation and Blind Analysis Contest, ASCE Str. Cong., Vancover, B.C., April Kasai, K., Ooki, Y., Motoyi S., Takechi T., Kajiwara, K., and Sato, E. (28a). Reslts of Recent E-Defense Tests on ll-scale Steel Bildings: Part 3 Experiments on s and rame Sbassemblies, ASCE Strctres Congress, Vancover, B.C., April Kasai, K. Ito, H., and Ogra, T. (28b). Passive Control Design Method Based on Tning of Eqivalent Stiffness of Bilinear Oil, Jornal of Str. and Constr. Eng., AIJ, No.63, Ag., pp (in Japanese) Kasai, K., Ooki, Y., Ito, H., Motoyi, S., and Kaneko, K. (28c). Vale-Added 5-Story Steel rame and Its Components: Part 1 ll-scale Tests and Analytical Modeling, 14WCEE, Beijing, Oct Jiang, T. (translation), JSSI Design and Constrction Manal for Passively Controlled Bildings: Chinese Translation, Manal Translation Committee, China Architectre and Bilding Press, Beijing, China, 28.