aashto seismic isolation design requirements for highway bridges

21
AASHTO SEISMIC ISOLATION DESIGN REQUIREMENTS FOR HIGHWAY BRIDGES By Ronald L. Mayes, 1 Ian G. Buckle, 2 Trevor E. Kelly, 2 and Lindsay R. Jones 4 ABSTRACT: In October 1990, the American Association of State Highway and Transportation Officials (AASHTO) adopted Guide Specifications for the Seismic Isolation Design of Highway Bridges. This paper overviews the basic concepts and design principles of seismic isolation and discusses the objectives and philosophy of the provisions. A summary of the provisions is presented, and the paper con- cludes with a procedure to compare the performance of isolation systems with different damping values. INTRODUCTION The collapse of a highway bridge during an earthquake will in many cases sever vital transportation routes at a time when they are most needed—to provide emergency services to or help evacuation from a stricken area. The loss of the bridge as a transportation link may potentially result in a greater loss of life than the immediate effects of collapse. Historically, bridges have proved to be vulnerable to earthquakes, sus- taining damage to substructures and foundations, and in some cases being totally destroyed as superstructures collapse from their supporting elements. In 1964, nearly every bridge along the partially completed Copper River Highway in Alaska was seriously damaged or destroyed. Seven years later, the San Fernando earthquake damaged more than 60 bridges on the Golden State Freeway in California. Guatemala, Japan, New Zealand, and Chile have also experienced seismic damage to modern bridges in recent years. The San Fernando earthquake of 1971 taught engineers a great deal about the seismic resistance of bridge structures and resulted in the development of improved provisions for the design of new highway bridges (Standard 1977). This earthquake also demonstrated the potential inadequacy of past design procedures in providing seismically resistant bridges. Since most existing bridges in service today were designed using pre-1971 design pro- cedures, it follows that many of the nation's highway bridges in seismically active areas may have insufficient strength to resist seismic loading. In 1981, the development of "Seismic Design Guidelines for Highway Bridges" (1981) by the Applied Technology Council was another significant step forward in the seismic-code development process. These requirements were national in scope and intended for use in the design of new structures. In 1983, they were adopted by the American Association of State Highway and Transportation Officials (AASHTO) as guide specifications (Guide 1983), 'Pres., Dynamic Isolation Systems Inc., 2855 Telegraph Ave., Suite 410, Berkeley, CA 94705. 2 Deputy Dir., Nat. Ctr. for Earthquake Engrg. Res., State Univ. of New York at Buffalo, Red Jacket Quadrangle, Buffalo, NY 14261. 3 Vice Pres. Engrg., Dynamic Isolation Systems Inc., Berkeley, CA. 4 Executive Vice Pres., Dynamic Isolation Systems Inc., Berkeley, CA. Note. Discussion open until June 1, 1992. To extend the closing date one month, a written request must be filed with the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible publication on January 2, 1991. This paper is part of the Journal of Structural Engineering, Vol. 118, No. 1, January, 1992. ©ASCE, ISSN 0733-9445/92/0001-0284/$!.00 + $.15 per page. Paper No. 1149. 284 J. Struct. Eng. 1992.118:284-304. Downloaded from ascelibrary.org by UNIVERSIDADE FEDERAL DE PARANA on 08/22/14. Copyright ASCE. For personal use only; all rights reserved.

Upload: fcambro13

Post on 14-Nov-2015

50 views

Category:

Documents


2 download

DESCRIPTION

Overview the basic concepts ans design principles of seismic isolation and discuses the objetives and philosophy of the provisions.

TRANSCRIPT

  • AASHTO SEISMIC ISOLATION DESIGN REQUIREMENTS FOR HIGHWAY BRIDGES

    By Ronald L. Mayes,1 Ian G. Buckle,2 Trevor E. Kelly,2 and Lindsay R. Jones4

    ABSTRACT: In October 1990, the American Association of State Highway and Transportation Officials (AASHTO) adopted Guide Specifications for the Seismic Isolation Design of Highway Bridges. This paper overviews the basic concepts and design principles of seismic isolation and discusses the objectives and philosophy of the provisions. A summary of the provisions is presented, and the paper con-cludes with a procedure to compare the performance of isolation systems with different damping values.

    INTRODUCTION

    The collapse of a highway bridge during an earthquake will in many cases sever vital transportation routes at a time when they are most neededto provide emergency services to or help evacuation from a stricken area. The loss of the bridge as a transportation link may potentially result in a greater loss of life than the immediate effects of collapse.

    Historically, bridges have proved to be vulnerable to earthquakes, sus-taining damage to substructures and foundations, and in some cases being totally destroyed as superstructures collapse from their supporting elements. In 1964, nearly every bridge along the partially completed Copper River Highway in Alaska was seriously damaged or destroyed. Seven years later, the San Fernando earthquake damaged more than 60 bridges on the Golden State Freeway in California. Guatemala, Japan, New Zealand, and Chile have also experienced seismic damage to modern bridges in recent years.

    The San Fernando earthquake of 1971 taught engineers a great deal about the seismic resistance of bridge structures and resulted in the development of improved provisions for the design of new highway bridges (Standard 1977). This earthquake also demonstrated the potential inadequacy of past design procedures in providing seismically resistant bridges. Since most existing bridges in service today were designed using pre-1971 design pro-cedures, it follows that many of the nation's highway bridges in seismically active areas may have insufficient strength to resist seismic loading.

    In 1981, the development of "Seismic Design Guidelines for Highway Bridges" (1981) by the Applied Technology Council was another significant step forward in the seismic-code development process. These requirements were national in scope and intended for use in the design of new structures. In 1983, they were adopted by the American Association of State Highway and Transportation Officials (AASHTO) as guide specifications (Guide 1983),

    'Pres., Dynamic Isolation Systems Inc., 2855 Telegraph Ave., Suite 410, Berkeley, CA 94705.

    2Deputy Dir., Nat. Ctr. for Earthquake Engrg. Res., State Univ. of New York at Buffalo, Red Jacket Quadrangle, Buffalo, NY 14261.

    3Vice Pres. Engrg., Dynamic Isolation Systems Inc., Berkeley, CA. 4Executive Vice Pres., Dynamic Isolation Systems Inc., Berkeley, CA. Note. Discussion open until June 1, 1992. To extend the closing date one month,

    a written request must be filed with the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible publication on January 2, 1991. This paper is part of the Journal of Structural Engineering, Vol. 118, No. 1, January, 1992. ASCE, ISSN 0733-9445/92/0001-0284/$!.00 + $.15 per page. Paper No. 1149.

    284

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • and later became AASHTO's standard specification for seismic design (Standard 1991). The ATC also completed another significant project en-titled "Seismic Retrofitting Guidelines for Highway Bridges" (1983). Since seismic retrofit is still as much an art as a science, these guidelines have not been adopted as code requirements; however, they are being used by several states in their retrofit work. In 1987, a Federal Highway Administration-funded project resulted in a document entitled Seismic Design and Retrofit Manual for Highway Bridges (Buckle et al. 1986), which was an imple-mentation manual of the new and retrofit design requirements developed by the ATC. This document also provided for the first time a summary of good and bad structural form. This issue is key to the good seismic per-formance of highway bridges but is difficult to quantify. Consequently, this section of the manual is a valuable starting point for the seismic design process.

    The October 1989 Loma Prieta earthquake, 70 mi south of San Francisco, California, was another reminder of the vulnerability of our existing bridges. This earthquake caused minor damage to 80 bridges, and 10 needed tem-porary supports, 10 were closed due to major structural damage, and three collapsed ("Loma Prieta" 1990). Damage and replacement costs are esti-mated to be between $1.8 billion and $2.0 billion. This figure includes the cost to complete the Caltrans retrofit program on the state highway system, but it excludes the societal costs of closed or restricted transportation routes.

    Thus, the problem of the seismic safety of our existing highway bridges is widespread and of sufficient magnitude to warrant national attention. Although bridge failures due to earthquakes have been confined to Alaska and California, many of these failures occurred at relatively low levels of ground motion, e.g., most of the bridge failures in the 1989 Loma Prieta earthquake were at acceleration levels of 0.15-0.25 g. Seismologists have estimated that 37 of the 50 states and Puerto Rico have the potential for ground motions of magnitudes greater than or equal to that which has resulted in serious bridge damage in past earthquakes. Comparatively high levels of ground shaking can be expected in 15 of these states as well as in Puerto Rico. The potential for bridge failure in many states may also be aggravated by a previous lack of emphasis on seismic design. Certain struc-tural details, especially at bearings, have proved extremely vulnerable to damage in past earthquakes.

    Within the context of the increasing awareness of the earthquake problem, a relatively new technology called seismic isolation has emerged as a practical and economical alternative to conventional design. This concept has re-ceived increasing academic and professional attention ("Proceedings" 1986; Seismic 1989; Billings and Kirkcaldie 1985; Blakeley 1979; Buckle et al. 1986; Buckle and Mayes 1987,1989,1990; "Loma Prieta" 1990; Kelly 1986) and is being applied to a wide range of civil engineering structures. To date, there are several hundred bridges in New Zealand, Japan, Italy, Iceland, and the United States that use seismic isolation principles and technology for their seismic design. The basic intent of seismic isolation is to increase the fundamental period of vibration such that the structure is subjected to significantly lower earthquake forces.

    One of the major impediments to the implementation of seismic isolation has been the lack of code requirements (Mayes et al. 1990). With liability issues being a major concern to design professionals in today's litigious society, many firms have been unwilling to use the technology without the availability of professionally acceptable code provisions. Thus the October

    285

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • 1990 adoption of seismic isolation design requirements for highway bridges by AASHTO (Standard 1991) is a key step forward in the more widespread use of seismic isolation.

    This paper reviews the basic principles of seismic isolation, summarizes AASHTO's guide specifications for seismic isolation bridge design, and provides a methodology for comparing different isolation systems.

    BASIC CONCEPTS

    There are three basic elements in any practical isolation system:

    A flexible support so that the period of vibration is lengthened suffi-ciently to reduce the force response.

    A damper or energy dissipator so that the relative deflections across the flexible support can be limited to a practical design level.

    Rigidity at low (service) load levels such as wind and braking; forces.

    Flexibility An elastomeric bearing is not the only means of introducing flexibility

    into a structure, but it certainly appears to be the most practical and the one with the widest range of application. The idealized force response with increasing period (flexibility) is shown schematically in the acceleration re-sponse curve of Fig. 1. Reductions in base shear occur as the period of vibration is lengthened. The extent to which these forces are reduced is primarily dependent on the nature of the earthquake ground motion, the soil type, and the period of the fixed-base structure. However, as noted, the additional flexibility needed to lengthen the period will give rise to relative displacements across the flexible support. Fig. 2 shows an idealized displacement response curve from which displacements are seen to increase with increasing period (flexibility).

    Energy Dissipation Relative displacements can be controlled if substantial additional damping

    is introduced into the structure at the isolation level. This is shown in Fig. 3, as is the smoothing effect of higher damping.

    One of the most effective means of providing a substantial level of damp-ing (greater than 20% equivalent viscous damping) is hysteretic energy dissipation. Fig. 4 shows an idealized force-displacement loop where the enclosed area is a measure of the energy dissipated during one cycle of motion. Mechanical devices that use the plastic deformation of either mild steel or lead to achieve this behavior have been developed. Mild-steel bars in torsion and cantilevers in flexure have been tested, refined, and are now included in several bridge structures (Blakeley 1979; Billings and Kirkcaldie 1985). Similarly, lead extrusion devices and lead-rubber (elastomeric) bear-ings have also been developed and implemented ("Design" 1990; Buckle and Mayes 1987, 1989).

    Many engineering materials are hysteretic by nature, and all elastomers exhibit this property to some extent. By the addition of special-purpose fillers to elastomers, it is possible to increase their natural hysteresis without unduly affecting their mechanical properties. Such a technique gives a useful source of damping, but so far it has not been possible to achieve the same level of energy dissipation as is possible with a lead-rubber elastomeric bearing.

    286

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • c _o '+-< TO i -

    _0> O) u u <

    * Period Shift

    Period FIG. 1. Idealized Acceleration Response Spectrum

    Period Shift

    c

    u

    _ro Q.

    Period FIG. 2. Idealized Displacement Response Spectrum

    287

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • CD U U <

    Acceleration Response Spectrum

    BB.

    Period

    FIG. 3. Response Spectra for Increasing Damping

    Plastic Deformation

    Displacement

    Hysteresis Loop

    FIG. 4. Hysteretic Force-Deflection Curve

    288

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • Rigidity under Low Lateral Loads While lateral flexibility is highly desirable for high seismic loads, it is

    clearly undesirable to have a structural system that will vibrate perceptibly under frequently occurring loads, e.g., wind or braking loads. Mechanical energy dissipators may be used to provide rigidity at these service loads by virtue of their high initial elastic stiffness. As an alternative, some isolation systems use a separate wind restraint device for this purposetypically a rigid component that is designed to fail at a predetermined level of lateral load. A listing of the alternative sources of flexibility and energy dissipation are given in Table 1. More details on these concepts have been covered in two comprehensive documents ("Proceedings" 1986; "Seismic" 1989).

    SEISMIC ISOLATION DESIGN PRINCIPLES

    The design principles for seismic isolation are shown in Fig. 5. The top curve on this figure shows the elastic forces that will be imposed on a conventional fixed-base structure (from the new AASHTO standard spec-ifications) for a rock site if the structure has sufficient strength to resist this level of load. The lowest curve shows the forces for which the AASHTO standard specifications requires a multicolumn bent bridge to be designed, and the second lowest curve shows the probable strength, assuming the structure is designed for theAASHTO forces. The probable strength is 1.5-2.0 times higher than the design strength because of the design load factors, actual material strengths (which are greater in practice than those assumed for design), conservatism in structural design, and other factors. The dif-ference between the maximum elastic force and the probable yield strength is an approximate indication of the energy that must be absorbed by ductility in the structural elements. However, when the bridge is isolated, the max-imum forces are reduced considerably due to period shift and energy dis-sipation. The forces on a seismically isolated structure are shown by the small dashed curve in Fig. 5. If a seismically isolated bridge is designed for the AASHTO forces in the period range of 1.5-3.0 sec as shown in Fig. 5, then the probable yield strength of the isolated bridge is approximately the same level as the maximum forces to which it will be subjected. Therefore, there will be little or no ductility demand on the structural system, and the lateral design forces are reduced by approxiamtely 50%.

    TABLE 1. Alternative Sources of Flexibility and Energy Dissipation Flexible mounting systems

    (1) Unreinforced rubber blocks Elastomeric bearings

    (Reinforced rubber blocks) Sliding plates Roller and/or ball bearings Sleeved piles Rocking systems Suspended floors Air cushions Steel springs

    Damping device/mechanisms (2)

    Plastic deformation of metal Friction High-damping elastomers Viscous fluid damping Tuned mass damping

    289

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • FOR ISOLATED BRIDGES (Period)

    FIG. 5. Earthquake Forces

    CONSIDERATIONS FOR SEISMIC ISOLATION

    Seismic isolation may be applied to both the design of new bridge struc-tures and the retrofit of existing structures (Blakeley 1979; Billings and Kirkcaldie 1985; Buckle and Mayes 1987, 1989, 1990). In general, imple-mentation is straightforward, since most bridges have bearings to accom-modate thermal movements, and the substitution of isolation bearings for these standard hardware items is routine.

    For new construction, the reduction in elastic column forces by factors of 3-10 will provide cost savings of up to 10% in both the columns and foundations, particularly if piled footings are used (Billings and Kirkcaldie 1985). There will also be substantial long-term reductions in the repair costs of seismic damage. For existing bridges, seismic isolation is an effective solution to the three most common deficiencies in bridges built before the mid-1970s. These are:

    Vulnerability of existing steel rocker and roller bearings and their connections.

    Insufficient strength and ductility of columns and substructures. Inadequate support length for girders.

    The last item is the area in which most of the current U.S. effort in bridge seismic retrofit has been carried out. The California Department of Trans-portation (Caltrans) has pioneered this work and has so far concentrated on the provision of positive connections between the superstructure and supporting substructure (Zelinski 1985). Measures used to date include lon-gitudinal joint restrainers, transverse bearing restrainers, and vertical mo-tion restrainers. Other concepts that have been proposed include bearing seat extensions and the use of shear keys or stoppers.

    290

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • An elegant solution to the three problem areas noted can also be achieved by replacement of vulnerable bearings with seismic isolation bearings. This not only solves the problem of inadequate bearing strength but also provides a solution for the other two problem areas due to the force reduction and displacement control features of these bearings (Buckle and Mayes 1987, 1989). Another feature of elastomeric-based isolation bearings is the ability to control the distribution of lateral forces, such as wind and seismic forces, to the substructures ("Design" 1990). Like seismic isolation design, the force control feature impacts the global design of the bridge, since lateral forces can be directed away from weaker substructures.

    The benefits of seismic isolation for bridges may be summarized as follows:

    1. Reduction in the elastic forces to which a bridge will be subjected by a factor of between 3 and 10 (based on curves 1 and 2 of Fig. 5 and a period shift, due to isolation, from 0.4 to 2 sec).

    2. Elimination of the ductility demand and thus of damage to the piers. 3. Control of the distribution of the seismic forces to the substructure elements

    with appropriate sizing of the elastomeric bearings. 4. Reduction in column design forces by a factor of at least two compared

    to conventional design (based on curves 4 and 2 of Fig. 5 and a period shift, due to isolation, from 0.4 to 2 sec).

    5. Reduction in foundation design forces by a factor greater than 2.5 com-pared to conventional design (based on the fact that conventional design requires higher design forces for the foundations than for columns).

    PERFORMANCE IN PAST EARTHQUAKES

    A significant amount of research and development ("Proceedings" 1986; Seismic 1989; 1990), shake-table testing (Kelly et al. 1977,1980,1985; Kelly and Hodder 1981), computer modeling, and real-world applications have been performed in the past 15 years. Use of the technology is becoming more widespread since design codes have been adopted for bridges, build-ings, and hospitals. A report by Dynamic Isolation Systems {Performance 1991) documents the performance of constructed facilities in real earth-quakes through 1989.

    There are more than 125 seismically isolated structures in regions of moderate to high seismicity throughout the world. Many of these structures are instrumented with strong-motion accelerographs, and several have ex-perienced moderate-to-strong earthquake shaking, with peak ground ac-celerations at the site in excess of 0.2 g. Thus, there is a growing data base of information on the actual earthquake performance of seismically isolated structures (Chapman and Kirkcaldie 1990; Dowrick 1988; Kaneko et al. 1990; "Base-Isolated" 1990; Tamura et al. 1988). The recorded performance of these structures confirms the ability of seismic isolation to significantly reduce the structure's tendency to amplify ground motion.

    The largest levels of ground acceleration have occurred in Richter mag-nitude 6.3, 6.8, and 7.1 events. The Shimizu office building in Japan, which incorporates a lead-rubber isolation system, has been subjected to 63 dif-ferent earthquakes with magnitudes varying between 4.2 and 6.8. The two largest events produced peak ground accelerations of 0.27 g and 0.22 g. The Mark II Detector at the Stanford Linear Accelerator Center, which incorporates a lead-rubber isolation system, was 32 mi from the epicenter of the magnitude 7.1, Loma Prieta earthquake. A peak ground acceleration

    291

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • 0.29 g was recorded on the nearby Stanford campus. The elastic response spectrum computed from nearby records closely follows the Uniform Build-ing Code (UBC) 0.35-g SI spectral shape in the 1-2-sec period range.

    The five-span Te Teko bridge in New Zealand, which incorporates a lead-rubber isolation system, was 6 mi from the epicenter of the 1987 Richter magnitude 6.3 Edgecumbe earthquake. The peak ground acceleration at the bridge site was estimated to be in the 0.35-0.40-g range, since the nearest record 15 mi from the epicenter had a peak ground acceleration of 0.33 g. The elastic spectra computed from the ground motion record closely follows the UBC 0.25-g SI spectra shape in the 1-2-sec period range.

    The force-deflection characteristics of a practical isolation system are shown in Fig. 4. The nonlinear force-deflection relationship provides a built-in wind-resisting mechanism. The yield level of the isolation system that governs the higher initial stiffness segment of the system is a design variable and in most cases varies between 2% and 10% of the weight of the structure. Because of the need to provide this higher initial stiffness, it is expected that at low levels of ground acceleration (less than 10%), isolated structures will behave as conventional, fixed-base structures. The isolation system will become more effective as the ground acceleration increases. This is clearly demonstrated in the records of both minor and moderate events.

    OBJECTIVES OF PROVISIONS

    In the development of the seismic isolation design requirements for bridges, there were three basic objectives, as follows: (1) To be as consistent as possible with the recently adopted AASHTO standard specifications for conventional seismic design; (2) to be as consistent as possible with the recently adopted Uniform Building Code provisions for seismically isolated buildings; and (3) to be applicable to a wide range of possible seismic isolation systems.

    The first objective necessitated that the requirements fit within the seismic performance category (SPC) concept of the new seismic design provisions. This concept provides a gradation of requirements from minimal require-ments for the lowest category, SPC-A, with an acceleration coefficient less than 0.10, to the most stringent requirements for the highest category, SPC-D, with an acceleration coefficient greater than 0.29. The second ob-jective formed the primary basis for the isolation design requirements. Some modifications were required due to the differences between building and bridge structural form and design loads. Others were required to provide consistency with the first objective.

    The third objective necessitated that the requirements remain general and, as such, rely on mandatory testing of isolation system hardware to confirm the engineering parameters used in the design and to verify the overall adequacy of the isolation system. Some systems may not be capable of demonstrating acceptability by test and, consequently, would not be permitted. In general, acceptable systems will: (1) Remain stable for re-quired design displacements; (2) provide increasing resistance with increas-ing displacement; (3) not degrade under repeated cyclic load; and (4) have quantifiable engineering parameters (e.g., force-deflection characteristics and damping).

    Both static and dynamic analysis procedures are included (depending on the SPC). They are based on the same level of seismic input and require the same level of performance from the bridge. The design basis earthquake

    292

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • load corresponds to a level of ground motion that has a 10% probability of being exceeded in a 50-year time period, i.e., 475-year return period.

    PHILOSOPHY OF PROVISIONS

    Seismic isolation provides a significant reduction in the elastic seismic forces the bridge must resist when compared to conventional fixed-base design. As a consequence, there are two possible design philosophies that can be used, and both are included in the AASHTO isolation guide spec-ifications (Standard 1991). The first is to take advantage of the reduced forces and provide a more economical bridge design than conventional construction. This option uses the same response modification factors (R-factors) as the recently adopted AASHTO standard specifications and thus provides the same level of seismic safety. The advantage of this design option is that if seismic forces are governing the design of the bridge, cost savings up to approximately 10% of the total bridge cost can be realized (Billings and Kirkcaldie 1985).

    The second design option is to provide a bridge with much better seismic performance characteristics than that of a conventional design using the AASHTO standard specifications. The intent of this design option is to eliminate or significantly reduce damage (inelastic deformation) to the sub-structure and abutments. In this case, an R-factor of 1.5 will ensure essen-tially elastic response by eliminating the ductility demand on the substruc-ture. In bridges, this design option can generally be achieved for similar or less cost than a conventional design. Furthermore, it provides protection for earthquakes that may exceed the 475-yr design event. A longer return-period event closer to a maximum credible earthquake (e.g., 10% proba-bility of exceedance in 250 years) may be as high as double the 475-yr event in some seismic zones according to the latest U.S. Geological Survey (USGS) 2,300-year return period map (NEHRP 1988). For conventional design, the elastic forces will increase proportionally for this higher level event, whereas in an isolation design, the forces will increase only as a function of the stiffness of the elastomer (approximately 20%). Thus, seismic isolation pro-vides significant protection from earthquake damage for an earthquake considerably in excess of the design event.

    METHODS OF ANALYSIS

    The basic premise of the seismic isolation design provisions (consistent with those for buildings and hospitals) is twofold: (1) The energy dissipation of the isolation system can be expressed in terms of equivalent viscous damping; and (2) the stiffness of the isolation system can be expressed as an effective linear stiffness. These two basic assumptions permit both the single- and multimodal methods of analysis to be used for seismic isolation design.

    For sliding systems without a self-centering mechanism or for pure elasto-plastic isolation systems, the equivalent viscous damping concept is no longer valid. The equivalent viscous damping formula produces a value that is independent of the coefficient of friction for sliding systems or the yield point for elasto-plastic systems. Furthermore, because these systems lack a restoring force, the total design displacement may be underestimated. Con-sequently, it is necessary to perform a nonlinear time-history analysis for all seismic isolation systems that have no self-centering mechanism.

    293

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • Statically Equivalent Seismic Force and Coefficient For the design of conventional bridges, the form of the elastic seismic

    coefficient Cs is

    c-1-^ a) where S = soil type, with values that range from 1.0 to 1.2 to 1.5 for different soil types; A = acceleration coefficient and depends on the location of the bridge in the seismic risk map; and T = the fundamental period of vibration. Although the ground response spectrum decreases approximately as \IT for longer periods, the form given does not decrease as rapidly as 1IT. In fact, at a period of 2.0 sec, Cs will be approximately 50% greater than the ground acceleration response spectra. The two major reasons for introducing this conservatism in the design of longer period (tall columns, long spans) conventional bridges is stated in the commentary of the AASHTO Standard Specifications (1991) as follows:

    1. In longer period conventional bridges, high ductility demands will be con-centrated in a few columns.

    2. Instability of a conventional bridge is more of a problem as the period increases.

    For seismic isolation design, the elastic seismic coefficient is directly re-lated to the elastic ground response spectra. This is because the intent of seismic isolation design is to introduce flexibility and damping in specifically designed and tested elements with the goal of eliminating or significantly reducing the ductility demand on the substructures. Consequently, the con-servatism of the seismic coefficient required for long-period (long span, tall column) conventional bridges is not necessary for short-span, regular col-umn-height isolated bridges. The form of the seismic coefficient is therefore slightly different from that for a conventional design and, for 5% damping, is given by

    C, = ^ f (2) where A = the acceleration coefficient; , = the site coefficient for seismic isolation design, and the \IT factor accounts for the decrease in the ground response spectra ordinates as T increases. The specific 5, values for the isolation requirements reflect the fact that above a period of 1.0 sec, there is a 1.0 to 1.5 to 2.0 relationship for the spectral accelerations for soil types I, II, and III, respectively. Once again, Cs should not exceed a value of 2.5 A.

    If the effects of damping are included, the elastic seismic coefficient is given by

    c- = i w

    where B = the damping term for the isolation system. Note that for 5% damping, B = 1.0.

    The quantity Cs is a dimensionless design coefficient, which when mul-

    294

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • tiplied by g produces'the spectral acceleration. This spectral acceleration SA is related to the spectral dispalcement SD by the relationship

    SA = *2SD (4) where u> = the circular natural frequency and is given by 2TT/T. Therefore, since SA = Cs g

    sA = H g (5) and

    _ 1 AS, 9J9ASiT. ^ = ^ra9=^m {6)

    Denoting SD as D, which is the displacement across the elastomeric bearings, (6) is approximated by

    D - ^ f . (7)

    An alternate form for C, is possible. The quantity Cs is defined by the relationship

    F = C,W (8) where F = the earthquake design force, and W = the weight of the structure. Therefore

    F_ _ Sfceff x D Cs

    w~ w " w

    where S/ce(f = the sum of the effective linear springs of all isolation bearings supporting the superstructure segment. The equivalence of this form to the previous form is evident by recalling that S/ce(f = u>2W/g, from which

    oo2W D 4TT2 1 9J9AS,T c*

    = -^

    xw = -TxmAx^T- ^ Cs = ^ (10b)

    BT y ' Thus, in summary, (7) and (9) are used to determine the statically equiv-

    alent seismic force. The isolated period of vibration is given by

    T =

    2 7 \ / ? ? ^ v 2Xrff g

    The base shear V, which is equal to the statically equivalent seismic force, is obtained by substituting (7) and (11) into (8) V = F = J,Keft D (12a)

    A V V=~-W (126)

    295

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • Single-Mode Spectral Analysis The single-mode method of analysis given in Standard Specifications (1991)

    is also appropriate for seismic isolation design. In fact, use of the method is simplified with seismic isolation. Steps 1, 2, and 3 of the procedure are not necessary since the use of an isolation system will ensure a simple rigid-body deformation pattern of the superstructure.

    In step 4 of the procedure, the value of peix), the intensity of the equiv-alent static seismic loading, is determined as pe(x) = w{x)-C, . .(13) where Cs is calculated by (9), and w(x) = the dead load per unit length of the bridge superstructure. In step 5 of the procedure, the loading pe(x) is applied to the superstructure to determine the resulting member forces and displacements.

    Multimode Spectral Analysis The guidelines given in Standard Specifications (1991) are also appropriate

    for the response spectrum analysis of an isolated structure with the following modifications:

    1. The isolation bearings are modelled by use of their effective stiffness prop-erties determined at the design displacement D (Fig. 6).

    2. The ground response spectrum is modified to incorporate the damping of the isolation system (Fig. 7).

    The response spectrum required for the analysis must be modified to incorporate the higher damping value of the isolation system. This modified portion of the response spectrum should only be used for the isolated modes of the bridge. It will then have the form shown in Fig. 7.

    Time-History Analysis When a time-history analysis is required for systems with a noncentering

    capability, it is necessary for the time histories to be frequency scaled so that they closely match the appropriate ground response spectra for the site. In addition, the analytical model should incorporate the nonlinear defor-mational characteristics of the isolation system.

    DESIGN DISPLACEMENTS FOR SEISMIC AND OTHER LOADS

    Adequate clearance shall be provided for the displacements resulting from the seismic isolation analysis in either of the two orthogonal directions. As a design alternate in the longitudinal direction, a knock-off abutment detail (Fig. 8) may be provided for the seismic displacements between the abut-ment and deck slab. Adequate clearance for the seismic displacement must be provided between the girders and the abutment.

    The shear deflections in the isolators resulting from braking loads, wind loads, and centrifugal forces will be a function of the force-deflection char-acteristics of the isolators. Adequate clearance at all expansion joints must be provided for these movements.

    DESIGN FORCES FOR SEISMIC PERFORMANCE CATEGORY A

    The AASHTO Standard Specification (1991) for conventional design has only two requirements for SPC-A: All bearing and column connections are

    296

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • Am DISPLACEMENT

    PIG. 6. Characteristics of Elastomeric Bearing with Bilinear Characteristics (Note: Qd = characteristics strength (kips); Fy = yield force (kips); Fmax = maximum force (kips); K = postelastic stiffness (kip/inch); K = elastic (unloading) stiffness (kip/ inch); Kcff = effective stiffness; and A, = maximum bearing displacement)

    1.00 0.40 1.20 1.60 2.00 2.4C PERIOD (sEcont/5) -

    3.20 3.60

    FIG. 7. Development of Composite Response Spectrum

    297

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • Construction Joint Designed to Fail

    FIG. 8. Top of Back-Wall Knockoff Detail

    required to be designed for 0.2 times the dead load, and the minimum support length requirements must be met. In the isolation design require-ments, the isolation bearing connections are required to be designed for F = KeffD, where _Keff = effective linear stiffness of the isolation bearing; and D = displacement of the isolated superstructure using a minimum acceleration coefficient A of 0.10.

    This permits use of the real elastic force reduction provided by seismic isolation and will result in a lower connection design force than conventional design. It should be noted that the acceleration coefficient, which has a maximum value of 0.09 for SPC-A bridges, is specified to have a minimum value of 0.10 if seismic isolation is used. This conservatism will ensure, for most areas designated SPC-A, that the isolation bearings are capable of resisting force levels associated with twice the design earthquake, since most areas of SPC-A have an acceleration coefficient of 0.05.

    DESIGN FORCES FOR SPC-B, -C, AND -D

    Design forces for seismically isolated bridges in SPC-B, -C, and -D are obtained using the same load combinations as for a conventionally designed bridge. The two design philosophies discussed previously are incorporated in the determination of the design forces. The provisions permit the use of the same R-factors as conventional design, with a lower limit on the forces being the yield level of the isolation system. This option permits a more economic design with the same performance level as conventional design. If a higher level of performance is desired, it is recommended that an R-factor of 1 to 1.5 be used to ensure essentially elastic response. For bridges designated SPC-B, foundation design forces are determined based on one-half the R value used for column design; they need not be greater than the elastic forces. This is consistent with the foundation design procedure for

    298

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • conventionally designed bridges. For bridges designated SPC-C or -D, the foundation design forces need not exceed neither the elastic forces nor the forces resulting from plastic hinging in the columns.

    OTHER REQUIREMENTS

    Nonseismic Lateral Forces Since an element of flexibility is an essential part of an isolation system,

    it is important that the isolation system also provide sufficient rigidity to resist more frequently occurring wind, braking, and centrifugal loads. This requires an elastic restraint system with higher initial stiffness than the element of flexibility (see Figs. 4 and 6). Limits on displacements resulting from nonseismic loads need to be satisfactory to the design engineer.

    Lateral Restoring Force The basic premise of the seismic isolation design provisions is that the

    energy dissipation of the system can be expressed in terms of equivalent viscous damping and the stiffness by an effective linear stiffness. The re-quirements of this section provide the basis by which this criterion is met. Systems that do not meet this requirement are not excluded; however, the analysis requirements and vertical load stability requirements are more stringent.

    Vertical Load Stability The isolation system will provide a factor of safety of at least 3 for vertical

    loads (dead load plus live load) in its laterally undeformed state. It will also be designed to be stable under the dead load plus or minus any vertical load resulting from seismic effects at a horizontal displacement of 1.5 times the total design displacement for isolation systems with a lateral restoring force, and 3.0 times the design displacement if there is no lateral restoring force. If the design is based on a maximum credible response spectra, then the 1.5 and 3.0 coefficients shall be reduced to 1.1 and 2.2, respectively. The detailed design requirements of the system will be dependent on the type of system. The multipliers of 1.5 and 3.0 on the total design displacement are based on a design response spectra corresponding to a 475-year return period event. If a maximum credible response spectra is used for the design of the isolation, these multipliers are reduced to 1.1 and 2.2, respectively. In some of the low seismic-rick areas (A < 0.25) of the United States, a multiplier of 2.0 and 4.0 may be more appropriate since a longer return period event (2,300 years) may be up to two times greater than the 475-year event.

    REQUIRED TESTS OF ISOLATION SYSTEM

    The code requirements are predicated on the fact that the isolation system design is based on tested properties of prototype isolators. The testing section of the provisions provides a comprehensive set of tests to both establish the design properties of the system and determine the adequacy of the tested properties. Systems that have been previously tested with this specific set of tests on similar type and size of isolator units do not need to have these tests repeated. Design properties must therefore be based on manufacturers' preapproved or certified test data. Extrapolation of design properties from tests of similar type and size of isolator units is permissible.

    299

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • ELASTOMERIC BEARING DESIGN REQUIREMENTS

    Elastomeric bearings that are used for seismic isolation will be subjected to earthquake-induced displacements D and must therefore be designed to carry the vertical loads at these displacements safely. Since earthquakes are infrequently occurring events, the factors of safety required under these circumstances will be different from those required for more frequently occurring loads.

    Since the primary design parameter for earthquake loading is the dis-placement D of the bearing, the design procedures must be capable of incorporating this displacement in a logical consistent manner. The require-ments of section 14.2 of ttie AASHTO Standard Specifications (1991) limit vertical loads by use of a limiting compressive stress and therefore do not have a mechanism for including the simultaneous effects of seismic dis-placements. The British Specifications BE 1/76 and BS 5400 recognize that shear strains are induced in reinforced bearings by both compression and shear deformation. In these codes, the sum of these shear strains is limited to a proportion of the elongation-at-break of the rubber. The proportion (1/2 or 1/3 for service load combinations and 3/4 for seismic load combi-nations) is a function of the loading type.

    Since the approach used in BE 1/76 and BS 5400 incorporates shear deformation as part of the criteria, it can be readily modified for seismic isolation bearings. The design requirements given in the isolation design requirements are based on the appropriate modifications to BE 1/76 and BS 5400.

    The more conservative aspects of BE 1/76 and BS 5400 have been used. For example, BS 5400 requires the summation of compression, thermal, and rotational shear strains and requires this to be less than 5.0. BE 1/76 requires the summation of only the compression and thermal shear strains and requires this to be less than zJ2. The isolation guide specification {Standard Specification 1991) requires the summation of the three different shear strains (compression, thermal, and rotation) with a limit of EJ2, where EJ2 may not exceed 5.0. DESIGN ISSUES FOR ISOLATION SYSTEMS

    The global design issues for the isolation system involve the desired iso-lation period, the base shear, the damping of the system, and the required margins of safety for the system. There are many interrelated variables involved in an isolation system design. To provide an overview of how these variables impact the structural design process, the new AASHTO seismic isolation design requirements are used. The formulas used in these require-ments are for the response spectra shapes of the 1988 Uniform Building Code and the new AASHTO standard specifications {Standard 1991) for the design of conventional bridges.

    The displacement across the isolators D, the elastic base shear V, and the isolation period T have been derived previously and are given by the following formulas [see (7), (11), and (12fo)]:

    D = ^- (14a)

    T=2^i^h {14b)

    300

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • A S V = KD =-W (14c)

    where D = isolator displacement (inches); A = acceleration coefficient obtained from the seismic-risk map; S = soil-type coefficient (1.0 for soil type Slt 1.5 for soil type S2, and 2.0 for S3); T = isolated period (seconds); B = damping coefficient (1.0 for 5%, 1.2 for 10%, 1.5 for 20%, and 1.7 for 30% damping); K = overall isolator stiffness (GAITr for plain elasto-meric isolators, where Tr = the total elastomer height); V = elastic base shear (kips); W = structure weight, (kips); G = shear modulus of elastomer (kips/inch) and A is the bonded or plan area (inches2).

    The important design variables are impacted by the isolation system prop-erties as follows.

    Displacement The isolator displacement is given by (7), and the two isolation system

    properties that impact isolator displacements D are the isolated period T and damping coefficient B. D increases as the isolation period T increases and decreases as the damping coefficient B increases.

    Base Shear The base shear V is given by (11) and is inversely proportional to the

    isolation system period T and damping coefficient B, i.e., V decreases as both the isolation period T and damping coefficient B increase.

    Design of Isolators The design of an elastomeric-based isolation system is such that the iso-

    lation system period is proportional to the total rubber height Tr and in-versely proportional to the plan area A of the isolator and the elastomer shear modulus G as follows: Tct\/TJG A.

    Thus the isolated period increases as the total rubber height increases, the plan size decreases, and the shear modulus decreases.

    Isolator Safety Margins The margin of safety of an elastomeric base isolation system is a function

    of the plan size, vertical load, the total rubber height, the internal layer thickness, and the isolator displacement. The safety margin increases with increased plan size, lower total rubber height, lower vertical load, and lower displacement.

    PERFORMANCE COMPARISON OF ISOLATION SYSTEMS

    To compare the relative performance of isolation systems with different damping values, a basis for the comparison must be established. There are three possibilities: (1) The isolated structure must have the same base shear; (2) the isolated structure must have the same displacement; and (3) the isolated structure must have the same isolated period.

    If different isolation systems have different damping values, these three options are mutually exclusive, i.e., only one of these criteria may be sat-isfied at any one time. From a structural engineering design perspective, a key design parameter is the equal base shear option, since the structure will have the same degree of protection and safety margin regardless of the isolation system used.

    301

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • Each of these comparison options are evaluated with regard to their overall impact. For the purposes of numerical comparison, a 10% damped (Bw = 1.2) and a 30% (B30 = 1.7) system will be compared, since this is the practical range of currently available systems.

    Equal Base Shear If the elastic base shear of two isolated systems are to be equal, then the

    relationship between the isolated periods and resulting displacements of the two systems, defined by subscripts 1 and 2, can be derived from (7), (11), and (126), as follows.

    Set Ve2 = Vel, T2 = r 7\, and D2 = r2 DL, where r = BJB2, and 5X and B2 are the damping factors for the two isolation systems.

    For numerical comparisons, if V10 = V30, then Tw = 1.4T30, and Dw = 2.0030.

    Thus, for an equal base shear comparison, a 10% damped system will require a 40% longer isolated period, and will experience twice .the dis-placement of a 30% damped system.

    Equal Displacements If the displacements of the two isolated systems are to be equal, then by

    using (7), (11), and {12b) the relationship between the isolated periods and the elastic base shears of the two systems can be derived as follows.

    Set D2 = Du T2 = TJr, and Ve2 = r2 Vel, where r = BJB2. For numerical comparisons, if D10 = D30, then Tw = T30/1A, and Vw = 2.0 V30.

    For an equal displacement comparison, a 10% damped system will require a 40% lower period and will produce twice the base shear of a 30% damped system.

    Equal Isolated Periods If the periods of two isolation systems are required to be equal, then the

    relationship between the resulting displacements and base shears can be derived from (7), (11), and (12b) as follows.

    Set T2 = Tu Ve2 = r Vel, and D2 = r DL, where r = BXIB2. For numerical comparisons, if T1(t = T30, then Vm = 1.4 V30 and Dm =

    1.4 D30. For an equal isolated period comparison, a 10% damped system will

    produce 40% greater displacements and elastic base shears than a 30% damped system.

    In summary, the damping of the isolation system is a key component in controlling the structural design parameters of elastic base shear and the displacement of the isolated structure.

    SUMMARY

    Several practical seismic isolation systems have been developed and im-plemented in recent years as interest in the application of this technique continues to grow. Although seismic isolation offers significant benefits, it is by no means a panacea. Feasibility studies are required early in the design phase of a project to evaluate both the technical and economic issues. If seismic isolation is appropriate from a technical and first-cost perspective, then significant life-cycle cost advantages can be achieved. Thus, seismic

    302

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • isolation represents an important step forward in the continuing search for improved earthquake safety.

    Seismic isolation offers particular advantages to bridge structures. Re-ductions in earthquake loads can be significant, and savings can be achieved in the foundations of new designs as well as improved seismic performance (elastic response). Isolation also offers an elegant solution to many of the common retrofit problems encountered in existing bridges. Limited expe-rience to date has shown that isolation can be adapted and implemented to meet a wide variety of different site and bridge conditions.

    The adoption of the seismic isolation code requirements discussed herein is a key step in the implementation of any new technology. These code requirements should significantly increase the implementation phase of the technology. They provide engineers with professionally acceptable proce-dures, which, when lacking, cause liability issues to dominate decision mak-ing. The present lack of code provisions has been an impediment to the more widespread use of the technology, and thus these new code require-ments are a key s tep forward in the appl ica t ion of this beneficial technology.

    APPENDIX 1. CONVERSION TO SI UNITS

    To convert

    in. ft kips kip-in. lb/ft

    To

    mm m kN N-m N/m

    Multiply by

    25.4 0.305 4.45

    113 14.6

    APPENDIX II. REFERENCES

    "Base-isolated buildings by Sumitomo Construction." (1990). Report, R&D Section, Building Dept., Sumitomo Construction, Tokyo, Japan.

    Billings, I. J., and Kirkcaldie, D. K. (1985). "Base isolation of bridges in New Zealand." Proc, US-NZ Workshop on Seismic Resistance of Highway Bridges (Report No. 12-1), Applied Technology Council, May.

    Blakeley, R. W. G. (1979). "Analysis and design of a bridge incorporating mechanical energy dissipating devices for earthquake resistance." Proc. Workshop on Earth-quake Resistance of Highway Bridges (Report No. A TC-6-l)m Applied Technology Council, 313-342.

    Buckle, I. G., and Mayes, R. L. (1987). "Seismic isolation of bridge structures in the United States of America." Proc. 3rd U.S.-Japan Workshop on Bridge Engrg., Public Works Research Institute, Ministry of Construction, Tsukuba, Japan, 379-401.

    Buckle, I. G., and Mayes, R. L. (1989). "The application of seismic isolation to bridges." Proem ASCE Structures Congr.: Seismic Engrg.Res. and Practice, ASCE, 633-642.

    Buckle, I. G., and Mayes, R. L. (1990). "Seismic isolation: History, application and performanceA world view." Earthquake Spectra J., 6(2).

    Buckle, I. G., Mayes, R. L., and Button, M. R. (1986). Seismic design and retrofit manual for highway bridges. Computech Engrg. Services, Berkeley, Calif.

    Chapman, H. E., and Kircaldie, D. K. (1990). "Seismic design of base isolated bridges incorporating mechanical energy dissipators." Bridge Design and Res. Sem-inar, (RRU Bulletin 87), Road Research Unit, Transit New Zealand, vol. 3.

    "Design of force control bearings for bridges." (1990). Design handbook, Dynamic Isolation Systems, Berkeley, Calif.

    303

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.

  • Downck, D. J. (1988). "Edgecumbe earthquakeSome notes on its source, ground motions, and damage in relation to safety." Bull. New Zealand Nat. Soc. Earth-quake Engrg., 21(3).

    Kaneko, M., Tamura, K., Maebayashi, K., and Saruta, M. (1990). "Earthquake response characteristics of base-isolated buildings." Proc. 4th U.S. Nat. Conf. on Earthquake Engrg., Palm Springs, Calif., May.

    Guide specifications for seismic design of highway bridges. (1983). Am. Assoc, of State Highway and Transp. Officials, Washington, D.C.

    Kelly, J. M. (1986). "Seismic base isolation: Review and bibliography." Soil Dyn. Earthquake Engrg., 5, 202-216.

    Kelly, J. M., and Hodder, S. B. (1981). "Experimental study of lead and elastomeric dampers for base isolation systems." Report No. UCB/EERC-81/16, Earthquake Engrg. Res. Ctr., Univ. of California, Berkeley, Calif.

    Kelly, J. M., Beucke, K. E., and Skinner, M. S. (1980). "Experimental testing of an energy-absorbing base isolation system." Report No. UCBIEERC-80135, Earth-quake Engrg. Res. Ctr., Univ. of California, Berkeley, Calif.

    Kelly, J. M., Buckle, I. G., and Tsai, H. C. (1985). "Earthquake simulator testing of base isolated bridge deck." Report No. UCB/EERC-85/09, Earthquake Engrg. Res. Ctr., Univ. of California, Berkeley, Calif.

    Kelly, J. M., Eidinger, J. M., and Derham, C. J. (1977). "A practical soft story system." Report No. UCBIEERC-77127, Earthquake Engrg. Res. Ctr., Univ. of California, Berkeley, Calif.

    "Loma Prieta earthquake reconnaissance report." (1990). Earthquake Spectra J., May.

    Mayes, R. L., Jones, L. R., and Buckle, I. G. (1990). "Impediments to the imple-mentation of seismic isolation." Earthquake Spectra J., May.

    NEHRP recommended provisions for the development of seismic regulations for new buildings. (1988). Nat. Earthquake Hazards Res. Program, Building Seismic Safety Commission, Washington, D.C.

    Performance of seismically isolated structures in earthquakes. (1991). Dynamic Iso-lation Systems, Berkeley, Calif.

    "Proceedings of a seminar and workshop on base isolation and passive energy dis-sipation." (1986). ATC Report No. 17, Applied Tech. Council, Palo Alto, Calif.

    Seismic engineering: Research and practice. (1989). ASCE, New York, N.Y. Seismic retrofitting guidelines for highway bridges." (1983). Report No. ATC-6-2,

    Applied Tech. Council, Palo Alto, Calif. Standard specification for highway bridges. (1991). 15th Ed., Am. Assoc, of State

    Highway and Transp. Officials, Washington, D.C. Standard specification for highway bridges relating to seismic design. (1977). Div. of

    Struct., Caltrans, Sacramento, Calif. Tamura, K., Yamahara, H., and Izumi, M. (1988). "Proof tests of the base-isolated

    building using full-sized model." Proc, American Society of Mechanical Engi-neers, New York, N.Y.

    Zelinski, R. J. (1985). "California Department of Transportation bridge earthquake retrofitting program." Proc. Workshop on Seismic Resistance of Highway Bridges (ATC Report No. 12-1), Applied Technology Council, San Mateo, Calif.

    304

    J. Struct. Eng. 1992.118:284-304.

    Dow

    nloa

    ded

    from

    asc

    elib

    rary

    .org

    by

    UN

    IVER

    SID

    AD

    E FE

    DER

    AL

    DE

    PARA

    NA

    on

    08/2

    2/14

    . Cop

    yrig

    ht A

    SCE.

    For

    per

    sona

    l use

    onl

    y; a

    ll rig

    hts r

    eser

    ved.