AUSTRALIAN EARTHQ UAKE ENGINEERING MANUAL Third Edition, 1993 Editors: H. Max Irvine and Graham L. Hutchinson Published by TECHBOOKS, Melbourne‘All rights reserved. All material in this manual is copyright, and must not be reproduced in contravention of ‘the Copyright Act. The views expressed in the manual are those of the authors alone. The title Australian Earthquake Engineering Manual is a registered business name. © 1990, 1991, 1993 H.M. Irvine, G.L. Hutchinson, LM. Megget, MJ. Pender, J.H. Wood. Information presented in this manual is not intended for use without independent substantiating investigation on the part of the potential user. ISBN 07334 00981 Cover design: © 1993 B.M, Butler Publisbed by: ‘TECHBOOKS, Sydney, Melbourne, Fax: 61-3-344 4616 Phone: 61-3-344 6712AUSTRALIAN EARTHQUAKE ENGINEERING MANUAL Editors: Max Irvine and Graham Hutchinson Contributors: Graham Hutchinson Protessor of Civil Engineering The University of Melbourne Max Irvine Professor of Civil Engineering The University of New South Wales Les Meggett Senior Lecturer in Civil Engineering The University of Auckland Michael Pender Professor of Geotechnical Engineering The University of Auckland John Wood Phillips & Wood Ltd Consulting Engineers, Wellington Third Edition, 1993Table of Contents Chapter 1 Chapter 2 21 2.2 23 Chapter 3 Ba 32 33 Chapter 4 4a 42 421 422 423 424 425 43 Chapter 5 5.1 52 53 531 532 54 54] 542 Chapter 6 61 611 6.1.2 62 Introduction Seismicity and Ground Motions Regional seismicity . ce Characteristics of ground motions... 2... peers Influence of local site conditions . Response of Structures ‘The effect of ground motion on structures . . Structural configurations .. 2... : ‘Comparison with other environmental loadings . prada . Dynamic Analysis Techniques Natural frequencies and modeshapes . oe eee 1 Analysis techniques... . . Equations of motion for SDOF systems Response to an impulse... . . Response to an earthquake . Elastic Response Spectra . . : : Design Curves and Inelastic Response Spectra. ss see 24 Example... . eee Re ore teers er reeag References... 0... : Peete Design Philosophy and the Code Approach Introduction , eee eeaceee ete Earthquake loading and the design process... . Design criteria and analysis options Design criteria . . : ee pee : Analysis options... 2. : aad Real time dynamic analys Poa Response Spectrum approach eee Quasi-static Code approach Australian Earthquake Codes . Building Regulations ASI170.4 Minimum Design Loads on Structures Part: Earthquake Loads Design of Buildings Incoduction . . - Distribution of lateral foroes Horizontal torsion . . Building structure types secs Page iPage ii 621 6.2.2 623 624 625 626 6.27 6.28 6.3 63.1 6.3.2 6.3.3 6.3.4 63.5 6.3.6 63.7 638 64 641 642 643 644 645 64.6 Chapter 7 TA 72 721 722 1.23 73 Frames oo... pees es 9 walls Coupled shear walls Combined walls and frames. Braced frames . . . peneeeerests Bearing wall systems . . - Tilt up stotures . 2. eee Vertical accelerations ..... 2.0 .++- Energy dissipation techniques . . : ee Beam and column plastic hinges . . . . ere Beam hinge versus column hinge mechanisms . . . Wall hinges ee reeeieeesieie ‘Squat shear walls. SESE EE Ce eee east a Coupled shear walls... EEE eee terete Seismic isolation Eee See Braced dampers... 2... +++ Active control, Detailing for limited’ ductility... . . Reinforced concrete... 6... 1 s+ Confinement of end regions Beam - column joints Diaphragms ee eee eee Stmotural steel... eee reece Reinforced masonry... ee Bee el eee ee Walls a oer cea Beam-column joints... 2... es ‘Strengthening of risk bul Domestic Housing Non-structural elements and connections... 2... sees eee ss 62 Infill panels . Veneers . . s : ee Glazing . : pects eee Precast panels and their connections HE a etter Suspended ceilings Interstorey deflections . . . Building separation Other seismic hazards. age® References n Notation. . SEE EE eer eet eet ccc n Foundation Design Induction... 6... ee Cyclic soil properties... . - - Cyclic undrained strength . . Cyclic undrained stifiness . . Assessment of strength and stiffness from penetration resistance . , 1 1 Earthquake effects on a soil deposit... «742 18 751 152 Chapter 8 81 8.11 812 8.1.3 B14 Unsaturated sand. , Footings and raft foundations . . Ultimate capacity ee Foundations subject to vertical load Effect of footing shape Effect of footing depth . eciate Foundations subject to movement . . . ae Foundations subject to horizontalload . ee 25 General bearing capacity equation . . . a 25 Different definitions of factor of safety . . . eee 2 Factor of safety with respect to base movement ............., 29 Local and general shear failure in bearing capacity eee a Design factors of safety . . Tie beams Stiffness... ca The work of Gazetas . Dynamic effects . . Pile foundations... 2. Stiffness... 0... Vertical stiffness Lateral stiffness Inertial interaction»... , ‘Ultimate capacity of pile foundations . Vertical capacity ...... 2. 53 Shaft resistance for piles in clay . - 55 Shaft resistance for piles insand 0.22.2... 37 Bare resistance for pilesinclay 2.0... ee ee eee eee 58 Bare resistance for piles in sand . eee 59 Uplift resistance of belled piles : f 59 Lateral capacity of piles... . : 60 ‘The Broms method ......... 6 Piles in cohesionless soils... 2... eee eee 61 Piles in cohesive soils . More recent work . Capacity of pile groups . eRe eel Caes eee eee SHC 6 Pilerafts 2. eee eee SHH be. ‘Moment capacity of pile groups Peete 7 Free-standing free-head groups . : nm Free-standing fixed-head groups Beret 174 Pile rafts, : ote 74 Lateral capacity of pile groups : 1D Combined capacity, eee ec 15 References See 6 Retaining Structures and Slope Stability Retaining structures... . . ae eee eet el Imvoduction 6.0.2... 0. 1 Wall flexibility and pressures . 2 Seismic coefficients Sheet 3 Load combinations and stability analyses 4 Page i8.15 8.16 B16 82 821 82.2 823 Chapter 9 91 91d 9.2 921 922 923 924 925 93 94 941 942 943 944 95 95.1 95 953 954 955 956 9.6 9.6.1 9.62 9.6.3 964 965 9.6.6 Earthquake forces and pressure distribution «6... 2... 3 Rigid wall... . : 6 Stiff wal eee eee 7 Flexible wall eae eee Displaceable wall... 2.6.00 eee eee ao: Forced wall oo... ee eee ee : 10 Water pressures. . sepa ea oe eee cena Application 1 various types of walls... = - n Free standing walls founded on soil. . See ete euer teeta Free standing walls founded on rock or piles... 0. eee e eee 2 Tied-back and reinforced earth walls 2. ees eevee ee 12 Basement walls . . . n Bridge abutments... 4. + Slope stability... 6... 0 Introduction Wedge failure . . Circular failure surface . ps 16 References... 0.22. ee Figures... 0... eee eee a Design of Bridges Introduction . : eee 1 ‘Current design code provisions . . . . ieee: ‘Common types of damage . . « EES Re eee Societe Spans falling from supports... . « 5 Brittle failures in piers . : 5 Foundation failures... 2.00. + + 6 Abutment failures 6... 6. eee eee ee ‘ 7 Joint damage . ‘i 8 Dynamic response . . . 7 8 Design principles ce : 2 Design level earthquake... . . : ee 12 Design life and retum periods 6... sees cial Design loads... . a eee 14 Detailing and strength capacity principles . 16 Structural form . : See Piers of unequal sliffness . ‘ eee eeeeeee Location of plastic hinges . eae 219 Shape of piers... + ee 20 Single span bridges . . « Eo ee ee 2 ‘Two and three span bridges oe eee 22 Multiple span bridges q 23 Connection and joint details ss BeRsisgsdneneasiee | Rigid connections... . ate Seer Pr aa Flexible connections ote Bee a Restrainers and linkage systems a pier tops EES Restrainers and linkage systems at abutments . . se 6 Holding down devices... .-. 0.008 eet Bearing design Seer 26 Page iv97 O71 9.72 973 974 975 9.7.6 977 978 Chapter 10 10.1 10.2 102.1 102.2 10.2.3 10.2.4 10.3 103.1 10.3.2 10.4 105 106 Design of substructures... . Ductitity factors Confinement and ductility ea Length and potential plastic hinge region . , are Spacing of transverse reinforcement... .. . . ae Anchorage of confinement reinforcement... . 2.2... Splices in longitudinal reinforcement Flexural strength of potential plastic hinge regions Strength of piers cess References . ee Figures Perierasbiritensy Bins, Silos and Tanks Introduction... ee Eee Ee Ground-supported containers for bulk solids... . , Earthquake pressures . Combined pressures... 2... Period of vibration... . Foundation flexibility... , Peete Ground-supported tanks for liquids... 2... . Spring-mass analogy ......... Pressure distributions... . Energy dissipation in ground supported tanks ‘Shell design procedures for ground supported tanks Sepeneseirerenitee) Elevated structures... 0... eee 28 References eee eee PagevPage vi ‘This page is intentionally left blank.1 | Introduction ‘The recent Newcastle earthquake has focused attention on the importance of earthquake resistant design and construction Principles being widely understood by the engincering profession in Australia. Although damaging earthquakes are Felatively rare, it is obvious that the country has a level of seismicity that can no longer be ignored. Standards Australia has been reviewing the Earthquake Code AS 2121 (1979). The new Standard “Minimum Design Loads ‘on Structures Part 4: Earthquake Loads (A$1170.4)" was issued in 1993 and its likely effect is that many Australian engineers will be required 10 apply the principles of earthquake engineering in the design and construction of civil engineering structures in the future, ‘The purpose of this manual is to explain some of the background to seismicity and the nature of ground motions on the one hand, and of earthquake engineering analysis and design Practice on the other. ‘The material presented represents the distillation of many years of experience, both in research and in practice, on the part of the authors. It also represents a distillation of the very large amount that has been leamed by the intemational professional ‘community over the last four decades concerning the principles of earthquake resistant design and construction. The manual is by no means comprehensive but, by the same token, it believed that most issues of potential importance to Australian engineers, given the Australian context, have been addressed, In writing this manual the authors have attempted to strike the right balance, a balance in which important basic pri emphasised while, at the same time, recognising that great detail or advanced techniques, which may be appropriate in other parts of the world (where the levels of seismicity are higher) need not, indeed should not, be unduly emphasised For the fact is that, regardless of Newcastle 1989, and the very substantial economic cost that disaster represents, damaging earthquakes have been, and are likely 10 be, rare events in Australia, at least as far as our current knowledge is concemed This does not mean that most new civil enginecring structures should not be designed with appropriate levels of earthquake resistance in mind - they should be - but it does mean that the community and their technical representatives - in this instance earthquake engineers - must carefully weigh the costs and benefits of the extra measures of resistance supplied. The vexed ‘question of what to do about existing structures, many of them old and/or of historical value, is an area where, yet again, the community must ukimately decide the stance it wishes to adopt. ‘The role of the carthquake engineer here is again paramount, Page 1-1Australian Earthquake Engineering Manual ‘The Newcastle Earthquake of 1989 has been a catalyst for action on the part of the community and the profession Government departments, industry, including the insurance industry, are seeing 10 it that, in the future, it is most unlikely that the tisk of severe damage from earthquakes will be ignored or down played quite as much as the style of some engineering Gesign and construction in Australia over the last two or three decades would seem to indicate. In the future more attention will be paid to providing details for civil engineering structures by which a suitable measure of earthquake resistance can be guaranteed. Design practice will change and so, too, will construction practice. The changes will be gradual, but the authors believe the process is inevitable. ‘The manual may be read in conjunction with two other publications, the first is a text on structural dynamics called “Suuctural Dynamics for the Practising Engineer”, E & FN Spon, London 1993, by H M Irvine. This work provides background reading in structural dynamics which is of fundamental importance in the study and practice of earthquake engineering. Further details of this work may be found on the back cover. ‘The second publication is @ bulletin prepared by Munich Reinsurance in Australia on the Newcastle Earthquake of 1989. ‘This excellent document presents a comprehensive overview of Newcastle and the problems associated with the earthquake of 1989, as well as giving a thoroughly readable account of the seismicity of Australia. It also outlines other significant seismic events in the recent history of Australia This edition of the manual has been specially issued to coincide \with the release of the Australian Earthquake Loading Standard, AS 1170.4. To meet this deadline, it has been necessary to publish chapters 7 to 10 in a slightly different format. This in no way detracts from their technical content and the publishing style will be made consistent in future editions. ‘This manual is presented in 10 chapters, The first four are concemed with background material. Chapter 2 gives a brief overview of Australian seismicity and the nature of ground motions and, in Chapter 3, general matters 10 do with the response of structures to ground motions are considered, Earthquake loadings are compared with other environmental loadings, particularly wind loading. Comments are made on structural configurations. In the case of buildings this reference relates particularly to the symmetry, or lack of it, im the Door plan and in the lateral resisting elements that provide the skeleton in the structure Page 1-2Introouction ‘Chapter 4 provides an introduction to dynamic analysis, almost exclusively dynamic analysis of single degree of freedom systems responding to earthquake and earthquake-like loading, Response spectra and the concept of ductility are discussed at some length, Various points made are illustrated with appropriate examples, ‘The final six chapters concentrate on design issues. Chapter 5 outlines a design philosophy believed appropriate for Australia and talks about the approach to design adopted in the Australian Standard. Chapter 6 concems the design of buildings and contains sections on building types and the means by which energy may be dissipated in the structural systems undergoing moderate to strong shaking. Advice is given on the detailing of structural ‘members and sub- assemblages in order to achicve limited ductility. ‘Masonry structures are considered and there is a section on the strengthening of risk buildings, always an important consideration when damage has occurred in urban areas. Finally, another important area, often overlooked by engineers, relates 10 the seismic resistance and performance of non- structural elements such as cladding, pamtitions, pieces of special equipment and the like. The question of the provision of suitable connections for such non-structural clements is a crucial one. In Chapter 7 foundation design is considered. There is extensive material on earthquake effects on soil deposits and this is followed by advice on the seismic design of the three fundamental types of foundation namely, footings, raft foundations and pile foundations. Here the questions of stiffness and ultimate capacity are of great importance. It is the foundation systems of civil engineering structures which first intercept and interact with the arriving ground motions. This is different from other types of environmental loading. Accordingly, the integrity of foundation systems assumes a high priority in the overall seismic design process. ‘The related topics of slope stability and the behaviour of retaining structures is considered in Chapter 8, while Chapter 9 contains a lengthy exposition on the design of bridges. Bridges are vital elements in @ community's transportation network and it is imperative that important bridge structures are able 10 function adequately in the aftermath of an otherwise damaging event.Australian Earthquake Engineering Manual The final chapter, Chapter 10, considers the eanhquake performance and design measures necessary to meet the resistance requirements of bins, silos and tanks that are subjected to earthquake loading. These structures are either ground-suppored or elevated and contain fluids or granular materials. They are a particularly important class of structure in ‘Australia given the large resources and agricultural bases to our economy. Max Irvine Graham Hutchinson Sydney Melboume July 1993, Page 1-424 Seismicity and Ground Motions Regional seismicity ‘The most common cause of earthquakes is thought 10 be the violent slipping of rock masses along major geological fault lines in the earth’s crust, or lithosphere. These fault lines divide the global crust into about twelve tectonic plates which are rigid, relatively cool, slabs about 100km thick. Tectonic plates float on the molten mantle of the earth and move relative to one another at rates of 10-100mm/year. ‘The basic mechanism causing earthquakes in the plate boundary regions appears to be that the continuing deformation of the crustal structure eventually leads to stresses/strains which exceed the material strength. A rupture will then initiate at some critical point along the fault tine and will propagate rapidly through the highly stressed material at the plate boundary. In some cases the plate margins are moving away from one another. In those cases, molten rock appears from deep in the ‘earth to fill the gap, often manifesting itself as volcanoes. If the plates are pushing together, one plate tends to dive under the other and, depending on the density of the material, it may resurface in the form of volcanoes. In both these scenarios there may be volcanoes and earthquakes at the plate boundaries, both being caused by the same mechanism of movement in the earth's crus. New Zealand and parts of the Indonesian archipelago are representative of this. Another possibility is when the plate boundaries slide sideways past each other, essentially retaining the local surface area of the plate. The San Andreas fault is representative of this type of plate interaction. Itis believed that some three quarters of the world's earthquakes are accounted for by this rubbing-sticking-slipping mechanism with ruptures occurring on faults on boundaries between tectonic plates. Earthquake occurrence maps tend to outline the plate boundaries. Such earthquakes are referred to as inter plate earthquakes. Australia is located on the Indo-Ausiralian tectonic plate well away from the boundaries. The nearest plate boundaries to Australia are associated with the Pacific plate with a common boundary passing through New Zealand in the Papua New Guinea, Indonesia, and the Philippines in the North, fast and through Earthquakes do occur at locations away from the plate boundaries. Such events are known as intra-plate earthquakes *See, for example, “This Dynamic Planet: The World Map of Volcanoes, Earthquake and Plate Tecionics” published by ‘Smithsonian Institution and the U.S. Geological Survey, 1989 Page 2-1Australian Earthquake Engineering Manual 2.2 and they are much less frequent than inter-plate earthquakes. They are also much less predictable than events at the plate margins and they have been observed to be far more severe. For example, the Eastern United States, which is located well away from the tectonic plate boundaries of Califomia, has recorded the largest eanthquakes in the history of European settlement in that country. These major intra-plate earthquakes occurred in the middle of last century in South Carolina on the East Coast and Missouri in the Interior, However, because of the low population density at the time, the damage caused was minimal. It is significant to note, however, that these intra-plate earthquakes, although very infrequent, were larger than the ‘moderately sized inter-plate earthquakes that frequently occur along the plate boundaries in California. (It is thought that because tectonic plates are not homogeneous or isotropic, areas of Jocal high stress ate developed as the plate attempts to move as a rigid body. Accordingly, rupture within the plate, and the consequent release of energy, are believed to give rise to these intra-plate events), Similarly, in Australia significant intra-plate earthquakes have occurred although in relatively unpopulated areas. The 1989 Newcastle earthquake was in fact a relatively small intre-plate event, but the events in Tennant Creek in 1988 were quite large by world standards and would have caused considerable Gamage in a major urban area of the country. It is important, therefore, to view the seismicity of Australia in the correct context, The continent is indeed surrounded by countries straddling the edges of the tectonic plate on which ‘Australia is located. These countries, New Zealand, Indonesia, Papua New Guinea and the Philippines all suffer from relatively frequent inter-plate earthquakes, Australia, on the other hand, is subject to much less frequent intra-plate events. The Newcastle event has demonstrated that the seismicity of Australia cannot be entirely ignored and earthquakes should be treated as another environmental loading to be accounted for during design Characteristics of ground motions ‘The point in the carth’s crustal system where an earthquake is initiated (the point of rupture) is called the hypocentve, or focus of the earthquake. The point on the earth’s surface directly above the focus is called the epicentre and the depth to the focus is the focal depth. Earthquake-occurrence maps usually indicate the location of various epicentres of past earthquakes and these epicentres are located by seismological analysis of the effect of carthquake waves on strategically located receiving instruments called seismometers, When an earthquake occurs several types of seismic wave are radiated from the rupture, The most important of these are theSeismicity and ground motions body waves (P and S waves). P waves (or Primary waves) are essemtially sound waves travelling through the earth causing particles to move in the direction of wave propagation with alternate expansions and compressions, They tend to travel through the earth with velocities of up to 8000m/sec (up to thiny times faster than sound waves through air), S waves (or Secondary waves) are shear waves with particle motion transverse to the direction of propagation. S waves tend to travel at about 60% of the velocity of P waves so they always arrive at seismometers after the P waves. The time lag between arrivals often provides seismologists with useful information about the distance of the epicentre from the recorder. As seismic waves approach the surface the amplitudes of vibration tend to increase as the stiffness (or impedance) of the rock decreases near the surface, especially in the case of the sediments or alluvium (ie, amplification of vibration occurs). ‘This may be demonstrated using simple energy conservation principles. Also, the velocity of the seismic waves is considerably slowed near the surface and the S and P waves are refracted towards the vertical (see Figure 2.1). This results in the longitudinal P-wave motion being mainly vertical, if recorded on soft rock, and the comesponding S-wave motion is mainly horizontal ‘The total strain energy released during an earthquake is known as the magnitude of the earthquake and it is measured on the Richter scale. It is defined quite simply as the amplitude of the recorded vibrations on a particular kind of seismometer located at a particular distance from the epicentre. The amount of earthquake energy released is logarithmically related to the Richter number such that each unit increase in the Richier number means that the energy released increases 32 times and the amplitude of ground motion is increased ten-fold, The only practical limit to the scale is the fact that to produce magnitudes larger than about 9.5 the earth would probably split in two! In general “strong motion” earthquakes are associated with Richter magnitudes greater than 5 - such earthquakes have the potential to seriously damage structures that are close to the epicentres of such events, especially if the focal depth is low. In Australia carthquakes tend to be shallow focus events occuring at depths of the order of 10 km. The resulting surface motion is therefore devoid of the cushioning effect provided where ruptures occur at great depth. ‘The magnitude of an earthquake, which reflects the size of an earthquake at its source, by itself is not sufficient to indicate whether structural damage can be expected at a particular site, ‘The distance of the structure from the source has an equally important effect on the response of a structure, as do the local gtound conditions. The local intensity of a particular earthquake is measured on the subjective Modified Mercalli scale which ranges from 1 (barely felt) to 12 (total destruction). ‘The Figure2.1 Refraction of P and SwavesAustralian Earthquake Engineering Manual Pagoe-4 Modified Mercalli scale is essentially a means by which damage may be assessed after an earthquake, In a given ocation, where there has been some experience of the damaging effects of earthquakes, albeit only subjective and qualitative, regions of varying seismic risk may be identified ‘The Modified Mercalli scale is sometimes used to assist in the delineation of these regions. A particular earthquake will be associated with a range of local intensities which generally diminish with distance from the source, although anomalies, due to local soil and geological conditions, are quite common, ‘Traditionally, earthquake motions have been recorded using seismometers which essentially consist of a pendulum inside rigid housing. It is possible, by appropriately choosing the stiffness of the pendulum arm and the associated damping, to ensure that, when the housing is subjected 10 an earthquake motion, the movement of the pendulum mass, relative to the housing, is directly proportional to the acceleration being experienced by the base of the housing (see Ref. 4.5). By providing the instrument with high damping, one eliminates phase distortion in combining the various components of the incoming waves. High damping also permits a wider frequency range in the input motion to be accurately recorded. Modem seismometers (or seismographs) are sophisticated instruments utilising, in pan, electromagnetic principles. These instruments can provide digitised or graphical records of carthquake- induced accelerations in both the horizontal and vertical directions at a particular site. Unfortunately, the distribution of instruments in Australia is limited. In the past, seismometers have usually only been installed around large reservoirs 10 check on the possible local seismicity induced by the reservoirs themselves. Consequently there is no “national grid” of seismometers, and information from seismicity maps must be interpreted with a great deal of care as the data precision, accuracy and resolution normally varies over the area of the map and with time, It should be noted that even in highly seismic areas of the world, appropriate networks of seismometers are often absent. The basic lack of data world- wide, and particularly in Australia, means that information ‘conceming the influence of such factors as magnitude, distance and local soil condition on the characteristics of earthquake ‘motions is still rather scarce, Accelerometers provide records of earthquake accelerations at a particular site and the records may be appropriately integrated to provide velocity records and displacement records. Peak accelerations, velocities and displacements are all in tum significant for structures of differing stiffness (see Chapter 3 - The effect of ground motion on structures). Because of the double integration required, ground displacements can become strikingly inaccurate unless care is taken. Typically. in ‘Australia, peak ground accelerations of the order of 0.152Seismicity and ground motions 2.3 would be considered high (see Section 4.2.5 for up to date information in this regard) and the duration may be as much as 20 seconds, although 5 - 10 second durations seem more common. The frequency content of the ground motion is measured by the number of zero crossings of the acceleration record and is thought to be largely @ function of the earthquake source mechanism. Also, there is a tendency for the predominant period in the ground motion to lengthen the further the site is from the epicentre. This occurs because shorter period vibrations attenuate (or dissipate) more rapidly than long period vibrations, It should be noted that earthquake motions are rarely conducive to the establishment of resonant vibration conditions in structures. This sets them apart from some other major environmental loadings (see Chapter 3 - Comparison with other environmental loadings). Influence of local site conditions Local geological and soil conditions will have a significant influence on the ground motions experienced at the site during an earthquake, be it a major or minor event. ‘These conditions affect the earthquake motions experienced (and hence the structural response) in one, or more, of the following ways: a) interaction between the bed rock earthquake motion and the soil column will modify the actual ground accelerations input to the structure. This manifests itself by an increase in the amplitude of the ground motion over and above that at the bed rock, and a filtering of the motion so that the range of frequencies present becomes narrow, with the high frequency components being eliminated. This condition particularly arises in areas where soft sediments and alluvial soil overly bed-rock. The degree of amplification is dependent on the strength of shaking at the bed rock. Because of non-linear effects in the soil the amplification ratio is less in strong shaking than under base motions of lower amplitude b) the soil properties in the proximity of the structure contribute significantly to the effective stiffness of the structural foundation. This may be a significant parameter in determining the overall structural response, especially for structures that would be characierised as stiff under other environmental loadings. ©) the strength (and response) of the local soil under earthquake shaking may be critical to the overall stability of the structure. (A detailed discussion of the effects of an earthquake on a soil deposit is presented in Chapter 7). It is also important that information on relevant geological features such as faulting be assessed. Geological information on suspected active faults near the site can assist in providing a Page 2.5Austraan Earthquake Engineering Manual Page2-6 basis for evaluating the intensity of a likely earthquake. It is, usual to use this information, together with the regional seismicity data, to determine the likely level of seismic activity It is @ universal truth that damage to buildings and civil engineering structures is more severe when they are founded on soft soils than when founded on rock. It has been argued by some that, in a country such as Australia, where the risk is low and insufficient information exists to be definitive with respect to the validity of isoseismals (which delineate regions of equal expected intensity), it would be more appropriate to distinguish only between their two “soil” conditions, no matter where, One benefit of this is that maps of the local geology, including soil deposits, exist and these can be matched with existing and proposed developments in order to quickly ascertain the levels, of seismic loading appropriate for new or remedial design.Ei Response of Structures It is helpful t0 identify the major categories of structures to which the analysis and design principles embodied in this ‘manual apply 2s their applicability is structure-dependent. It is also useful to identify common characteristics and briefly discuss the skeletons that enable structures to function. Buildings and other engineered structures may have to resist earthquake attack from any direction. The maintenance of structural integrity in an earthquake will assume greater or lesser importance depending on the function of a structure (particularly any post-earthquake disaster relief function). ‘Structures may provide protection, or support, they are used as containers, they may occur singly or as part of a linked system and they may form a part or portion of another structure, ‘One major category of engineered structures takes the form of cantilevers and this group is particularly susceptible to earthquake loading. Many buildings and most towers and chimneys fall into this category. Suctures of horizontal extent, such as bridges, long span roofs, pipelines and transmission lines form another major category, as do retaining structures such as water tanks (both ground supported and elevated) and bins and silos used for storing a multitude of materials. One of the particular difficulties associated with retaining and containing structures lies in the determination of the extent to which the contained mass participates in the structural response under earthquake attack. It is also essential to ensure that machinery incorporated in structures in any of the above categories is adequately designed to resist carthquake loading. In particular, the mountings of machinery and electrical equipment are susceptible to damage. This is especially so if the equipment is located in pans of supporting structures which may themselves be strongly excited during an earthquake. Modem buildings contain a lot of expensive sensitive electronic equipment. In purpose-built structures such as telephone exchanges or central banks, the value of the equipment may be an order of magnitude greater than the building itself. This may mean that quite stringent design considerations will exist for structures and equipment and it also raises a fundamental question about placing central control in just one building, rather than spreading the risk over several. 3.1 The effect of ground motion on structures ‘The effect of ground motion on the various categories of ‘structures is dictated almost entirely by the distribution of mass and stiffness in the structure. It is important to appreciate that, Page 3-1‘Australian Earthquake Engineering Manual 3.2 Page 3-2 jn an earthquake, loads are not applied 0 the structure. Rather, ‘earthquake loading arises because of accelerations generated by the foundation level(s) of the structure intercepting and being influenced by transient ground motions. Specifically, the product of the structural mass and the total ‘acceleration produces the inertia loading experienced by the structure. This is an expression of Newton's Second Law. It is important to appreciate that the total acceleration is the absolute acceleration of the structure; namely, the sum of the ground acceleration and that of the structure relative to the ground. If the structure is stiff there is litte, if any, additional acceleration relative to the ground motion, and, therefore, the earthquake loading experienced is essentially proportional to the building mass, ie. Fey aM For structures which are flexible, for example those in the high- rise or Jong-span category, the absolute acceleration is low. This ‘occurs because the ground acceleration and the acceleration of the building relative to the ground tend to oppose one another. In this case the earthquake loading is approximately proportional to the square root of the mass, ie. Fa, 0. MOS For structures in the cantilever category, which are essentially vertical, it is the horizontal accelerations that are significant, whereas for structures that are largely horizontal in extent the effect of the vertical accelerations is dominant. Moreover, if the plan distributions of mass and stiffness are dissimilar in vertical structures, significant twisting motions may arise. ‘The peak ground acceleration is of importance in the response of stiff structures and peak ground displacements are of importance in the response of flexible structures, with peak ground velocity being of importance for structures of intermediate stiffness. Stiff structures tend to move in unison with the ground while flexible structures, such as high-rise buildings, experience the ground moving beneath them, their upper floors tending to remain motionless. Structural configurations Figure 3.1 shows a floor plan of a simple building. A column is located at each comer and a “nominal” shear wall, which may actually take the form of an internal stair or lift well, is also shown. When an earthquake strikes, the columns and the shear wall will provide the lateral resistance to the earthquake- induced loading ‘The force resulting from the earthquake may be considered to act through the centre of mass of the floor (see Figure 3.1) Similarly, the force resulting from the contributions of the resisting elements (ie. the columns and shear wall) may be considered to act through the centre of stiffness (or shear centre)Response of structures of the floor. The two points are not coincident because the distribution of column and shear wall stiffness is asymmetric with respect 10 the distribution of mass. In addition, various “accidental” eccentricities may be introduced in the real structure by live loads - which alter the distribution of mass - and by intemal partitions, etc, which also contribute to resisting the carthquake induced loading. It may be seen from Figure 3.1, therefore, that under earthquake loading the building is not only subjected to the expected lateral motion but is also subjected 10 a twisting or torsional effect. Statics would imply that the torsion has @ magnitude of Fe, where ¢, is the distance between the centre of mass and the centre of stiffness for the particular floor. However, because of the dynamic characteristics of buildings, influenced by both the magnitude and the distributions of mass and stiffness, the torsional cffect is amplified and the real torsion be considerably greater than the simple product of F and e, Earthquake codes around the world, including the Australian code, recognise this effect and include an amplification factor for torsional effects Even a building which may appear symmetrical from the design centre of mass ceaue of stifiness eanquake drawings will have “accidental” eccentricities associated with it, Figure3.2 Floor plan of a Earthquake codes usually recognise the presence of accidental simple building eccentricities - in the new Australian Earthquake Standard AS1170.4 the “accidental” eccentricity is taken as 5% of the maximum lateral dimension of the building perpendicular to the horizontal loading direction under consideration. This concurs with other eanhquake codes of practice. WA : extgate (a) Symmewic - (b) Bocenuic - wwist (c) Becentric - oe ee ee ee ocmcot mes emanate nin con Figure 3.2 Plan views of simple idealise building models A multi-storey building, with perfectly regular, identical, floor plans as shown in Figure 3.2(a) is considered symmetrical if the centre of stiffness of lateral resisting elements is co-incident with the same vertical line that connects the centre of mass of each floor slab. However, because of the presence of accidental eccentricities, such multi-storey buildings tend to have different eccentricities associated with each floor. Accordingly, since such eccentricities vary in magnitude and sense from floor to floor, the overall effect may not be as severe as the code would Page 3-3Australian Earthquake Engineering Manual imply. That is there would be compensating negatives and positives, Figure 3.2(b) shows a typical floor plan for a common type of building with marked eccentricity in that the service core of the building has been located against one face. In responding 10 earthquake attack from the direction indicated, (the resultant force of which passes through the centre of mass) the building will obviously twist as well as translate, ie the building is torsionally eccentric. However, it is possible to effectively remove the twisting effect by placement of walls as shown in Figure 3.2(c) and, although the eccentricity still remains, it does not directly affect the response provided the walls are properly designed and detailed. The extra walls effectively remove the twisting effect by developing an opposing torque which, because these walls are stiff, does not allow significant rotation in plan. Another commonly occurring form of structure are 3 10 4 storey walk-up blocks of flats constructed either as unreinforced masonry structures or as lightly framed structures with in-fill walls. Usually, the ground floor of such buildings is extremely torsionally flexible whille the upper floors tend to be quite rigid. ‘This form of construction is very common in Australia and may not perform well, even in minor earthquakes, because of the torsional motion that will be induced, Figure 3.3 shows an elevation of a building and another, perhaps more subtle, way in which eccentricities may be introduced into a structure, Even though each floor may be nominally symmetric, the overall effect of the stepping of the floors in elevation results in a gross asymmetry, with all its consequences, when under earthquake attack. It has already been noted that towers may be particularly susceptible to the effects of earthquakes. As a rather extreme example picture a thin-walled steel tower. The tower supports large and rather massive batery of lights and has a cut-out, in the form of an access door at its base. In order to minimize torsional effects from earthquake loading (and, of course, wind loading) when the ditection of attack is parallel to the face of the battery, the cut-out should be on the opposite face of the tower to that of the battery. This helps ensure that the shear centre and the centre of “loading” are more or less on the same vertical line. By introducing a cut-out, the tower has suffered a significant local loss of torsional stiffness so it is as well to attempt to minimize the torques generated by earthquake attack. Significant changes in the vertical distribution of mass and stiffness in @ structure can dramatically affect the structure's response to earthquake attack. Industrial buildings tend to have few secondary components thus, in many cases, minimising twisting effects due to accidental eccentricities. However, such structures may have quite dramatic changes in mass and Figure 3.3 Elevation of a ypical stepped building eeResponse of structures stiffness, particularly if, as shown in Figure 3.4(a), a heavy silo is supported in the structure. Figure 3.4(b) shows the dramatic changes in scismically-induced shear forces that may occur because of the silo mass and the stiffer members required to support it ‘Theoretically, dramatic reduction in a storey stiffness (soft storey concept) provides a spongey layer which effectively isolates a structure from ground motion. In particular, an otherwise stiff structure could be protected from short period Ge, high frequency) ground motions by making the bottom storey relatively flexible. Indeed, this sometimes happens as architectural requirements often call for large column free ‘spaces at ground floor level. Experience has shown, however, that many modem buildings of this type perform badly in earthquakes and the so-called soft-storey concept has now been shown 10 be flawed, both theoretically and practically. It appears that soft storeys should, if possible, be avoided because the earthquake response in terms of forces is concentrated in relatively few structural elements, the capacity of which may be insufficient to meet the demands placed on them. ‘Notwithstanding this, the not unrelated technique of base — Curve 1— Absolute summation of first three modes Curve 2— SRSS combination of first three modes Curve 3— Mode 1 contribution sor Curve 4 — Mode 2 contribution ° 200-400 68D «100 (a) Framing. system (®) Seismically-induced shear forces Figure 3.4Australian Earthquake Engineering Manual 3.3 Page 3-6 isolation, which has grown in popularity over the last two decades (chiefly as a result of research and development in New Zealand and Califomia) has many attributes. Some of the characteristics of base isolators are described in Chapter 6. Comparison with other environmental loadings In the Australian context the major environmental loadings likely to be encountered by structures include thermal loadings, wind and wave loading, and earthquake loadings. Cyclonic ‘winds tend to dominate structural design in the northem half of the continent, and the potential severity of wave loading on off- shore structures in Bass Strait, for example, is now well known. Moreover, a temperature range of at least 50° Celsius, causing significant thermal loadings in some types of structure, is possible in most pans of the country. These loadings, or combinations thereof, would unquestionably be regarded as significant in a world context. On the other hand, on the best available evidence, the Australian land mass is an area of relatively low seismic activity, This implies a low risk of (damaging) earthquake ‘occurrence, although it does not necessarily follow that low risk ‘means no risk, nor that low risk implies small magnitude events. For example, the three consecutive Tennant Creek (intra-plate) earthquakes of January, 1988, had Richter Magnitudes of approximately 6.7. ‘The most obvious comparison to be made from a designer's point of view is usually that between the lateral loads caused by ‘wind and earthquake. Wind loading has a mean component with ‘fluctuations about that mean: earthquake loading is represented by fluctuations about a zero mean. Extremes of wind loading, which may be as much as three or four times the loading associated with the mean result, are possible and a significant contribution to this extreme is often supplied by the resonant component in the turbulence of the wind, Resonance refers to a condition in which the periodicity fof forcing is identical to that of the structure, with 2 consequential amplification of response that is Limited only by the level of damping the structure. A typical wind contains a wide range of frequency components in its turbulence so it is always possible that the peak response has a resonant component. Earthquake ground motions are characterised by a series of rather random spikes, with the range of frequencies present (ie, the range of intervals between zero crossings on the ground acceleration record) being somewhat narrower than for normal ‘wind turbulence. Structures thet are stiff will move essentially in unison with the ground motion. For more flexible structures response is analogous to that from a series of impulses, with the —Response of structures dominant frequency in the response being that of the stwuctuse itself. This frequency, the natural frequency of the structure, is dependent on the mass and stiffness of the system Wind loading depends on exposed area: earthquake loading depends on the (hidden) mass of the structure. Structures attract wind loadings which increase steadily with the major dimension (height or span, say) The earthquake loading experienced by such structures increases much less rapidly with the result that, for high rise structures, wind loading is almost always the dominant lateral loading. This assumes clastic responses for both regimes of loading. Wind loading depends on topography and, in urban areas, on the proximity of other buildings. Earthquake loading, on the other hand, depends to a marked degree on the foundation materials. It is universally observed that buildings founded on soft soils perform more poorly than those founded on rock, Page 3-7Australian Earthquake Engineering Manual Page 3-8at Dynamic Analysis Techniques Natural frequencies and modeshapes ‘The first step in any dynamic analysis of a civil engineering structure, once a satisfactory model has been constructed, is the evaluation of certain properties of the structural system which are independent of loading, at least insofar as response remains within the elastic regime, or where the tisk of local or global instability is low. ‘Two quantities of importance are the natural frequencies and the associated natural modes of vibration: the adjective natural signifying that their evaluation depends on determination of system properties alone, in this case the properties of stiffness and mass, The relationship between the two is manifested in a structure vibrating with a characteristic, unchanging shape when forced to do so at a frequency which is one of the natural frequencies of the system. ‘Any structural system has an infinity of natural frequencies and associated modes. These may be ordered from the lowest to the hhighest - the lower being those typically associated with movement (recall, the ubiquitous example of the street lighting standard in the shape of an T’ and the way it moves in the wind), while the highest will be the ringing modes associated with the passage of sound-waves in and around the structure Fortunately, in most of civil engineering practice, and in almost the whole field of earthquake engineering, it is only the first few that are of structural interest. Very often we are interested only in the fundamental mode: at most, the first three modes are required. We shall see the way in which these fundamental properties of the system are fed into the dynamic analysis process when we ‘come to look at response to transient Loads, A single degree of freedom model is shown by Figure 4.1. That single degree of freedom is the displacement A(¢) representative of a simple structure in which one displacement quantity to describe the motion. Page 4-1Australian Earthquake Engineering Manual Page 4-2 Figure 4.1 Schematic of a single degree of freedom model If the mass M is pulled to the right an amount 4, , and released from rest, it will undergo sinusoidal harmonic motion of the form 2nt Ai) = A,sin== £20 (1) = Aysin= or =A, sin 2rft or=4,sinor where the characteristic time of the system, the period 7, is related to the natural frequency J, and to the natural circular frequency «, by o=2h wi where the units of w are radians/second, and those of f are cycles/second. There are 2x radians in a circle, so dividing radians/second by 2x gives cycles/second. ‘Structures for which earthquake loading may be important have (fundamental) periods in the range 0.1 second to 10 seconds, approximately. Sydney Tower is a structure characteristic of the upper end of this range, while a Paddington terrace fits in at the other end. In the absence of damping this free vibration motion would be repeatable as shown by Figure 4.2. ‘Overain on this is the true free vibration response — that when damping is included. Even for quite large levels of damping, such as that shown here, the period of the structure is relatively unchanged, but it is a matter of common observation that damping, even the small levels typically inherent in civil engineering structures stressed below the elastic limit, eventually wipes out free vibrations. So, in the absence of dissipative agents such as damping (which is the third important system property after mass and stiffness), we may establish the formula for the natural period as follows: Eqn. 4.1 Eqn. 42Dynamic analysis techniques Exponential dacay Undamped structure Ag 4 Damped structure Time, t Figure 4.2 Effect of damping on free vibration At point A the strain energy stored is a maximum being, since the mass is momentarily stationary, dee Egn.4.3 Es On the other hand, at B, the situation is reversed with the spring, momentarily unloaded and the kinetic energy at its peak, which is FM Bw Eqn. 4.4 Further since, energy is conserved we have that Eqn. 4.5 and because Eqn. 4.6 we have Bigg, = 282% Eqn. 4.7 Ga leading to the period 7, which is the time scale of this process, namely Eqn 48 Units of mass in tonnes, and of stiffness in kN/m, lead to the unit of the period being seconds, as required, In assuming the spring and mass system of Figure 4.1 t0 represent a structure we have set aside the crucial question of just how an actual structure is rendered into this single degree equivalent, Page 4-3Australian Earthquake Engineering Manual Therefore, for purposes of illustration, we shall treat a simple cantilever structure, a structure which is slender, and for which bending action is the dominant means of resistance of the Joading. The structure in mind is the reinforced concrete wind shield for the flues of a coal-fired power station, The structure is a thin-walled tower with the wall thickness increasing as the base is approached in order to maintain stress levels, under self ‘weight and lateral loads, which do not vary significantly over the height of the structure. ‘The structure is shown by Figure 4.3. 2R Ao | T it te) 4 A@=Ao GZ TAS PRESS Figure 4.3 A tower and its associated fundamental mode of vibration. ‘The curve of deflection in response to distributed lateral loads for such flexing structures is given to a very good degree of precision by the parabola zy a@=A, Mal 63) and this represents an excellent choice for the (approximate) shape of the fundamental mode of vibration. We are thus choosing the tower tip deflection A, as the degree of freedom of this system, a system undergoing free vibration of the form Eqn. 4.9 Page 4-4 | |soa(g) ‘The peak velocity of the tower is given by 2 eae Cel ne sina«(2) and so the peak kinetic energy is 1p ey KES $l ane(2) ) dz (Ioc(zy «| oe M, 007 3 1 2 1 2 where M, is the effective mass of the tower. It is a great deal Jess than the total mass which is Me flin(2) oz the ratio M:M, being 5:1 for a tower of uniform mass, rising to 15:1 for one in which the wall tapers completely. The reason is, of course, that not all the tower is vibrating with the same amplitude as its tip. At another phase in the vibration the strain energy is greatest, and this occurs when L phenol PA) send feral ZA) oc 2 14 [FM) dea, or S.E.==—K,i 2 where K, is the effective stiffness of the tower. On account of conservation of energy we have Ke w= [xe mM, where we have, now, the crucial information relating 10 the appropriate stiffness and mass to use in making an appropriate single degree of freedom model of the structure. While most information of importance to the earthquake engineer is contained in the first mode of vibration and its associated natural frequency, all structures have @ multitude of modes. Dynamic analysis techniques Eqn. 4.50 qn. 4.11 Eqn. 4.12 Eqn. 4.13 Eqn. 4.14 Eqn. 4.15 Page 4-5Australian Earthquake Engineering Manual For cantilevers such as mulistory buildings, the first three ‘modes of vibration look like those shown in Figure 4.4, 10 r os 06 Height Ratio 04 02 oo 4 Ot Figure 44 Typical modes for multstory buildings in which botk bending and shear are important (from Ref. 48) ‘The modes indicate that some shearing action is occurring (notice, for example, the curvature of the first mode which is largely reversed from that for true flexure), and is typical of multistory buildings in which frame action is important. For pure shearing structures the periods of the higher modes are approximately Tr I, het.h-2 Peg eS while for pure flexing structures, the ratios are, approximately, i, i th Ret PeOt Ears Depending on the type of structure (ie, the degree of sheai/flexure in the modes), estimates of reasonable quality can ‘be made as to the periods of the higher modes. For example, T; = 7;/6, would be a good estimate for the period of the second mode of the slender chimney wind shield, whereas a value of T) = T,/4.5 would be an appropriate estimate for a typical 20-25 storey building in which the lift and service core provided the majority of the resistance to lateral loads. The walls of such cores exhibit @ reasonable amount of shearing deformation in cheir response, as well as the bending traditionally associated with shear wall structures. Eqn. 4.16 Eqn. 4.17 Page 4-6Dynamic analysis techniques 4.2 Analysis techniques 4.24 Equations of motion for SDOF systems Figure 4.5 is representative of a typical structural element, such as a pict which carries tributary portions of span at a hammer head, say. mass, M MY | stitiness, K | damping, C ayn ‘The mass of the structure, M , the stiffness, K, and the damping level, C, are all assumed known. The ground displacement as a result of strong shaking, «,(z), is assumed known, something which can never be the case since we are talking of the future. ‘The displacement of the mass relative to the foundation, A(t), is the quantity to be calculated. © Mode! Figure 45 SDOF Model of a typical civil engineering structure To begin to find this we must first write the equation of dynamic equilibrium (when this equation is solved we find A(e) and so, for example, the spring force KA() — or, for that ‘matter, the seismic shear force in the pier may be found) From Newton's second lew we have M(ii, +A)+CA+KS=0 Eqn. 4.18 where the first term is the inertia force (note that we must use the total acceleration (+8) when writing this term), the second is the damping force (this is 2n idealisation of the actual mechanisms of energy dissipation ina vibrating sinucture — the phrase “equivalent viscous damping coefficient” is sometimes used to describe C), and the third term is the spring force. Page 4-7Australian Earthquake Engineering Manual ‘This equation is usually rearranged into the form &+2tah + 0% =~ii where @ = {K/M the natural circular frequency in radians per second = C/(2VKM) .the fraction of critical damping Critical damping (= 1) refers to the situation when, for example, the mass is just able to retum to its position of equilibrium, after being drawn to one side and released, in one gradual non-oscillatory movement, A self-closing door is a 00d (qualitative) example of a critically damped system. Damping levels are problematic as they are certainly dependent ‘on material type, vibration amplitude levels and, ultimately, during strong shaking, on damage levels. Values of § of 0.0/ - 0.02 - 0.03 might be typical of well - detailed structures of steel ~ prestressed concrete - reinforced concrete under modest shaking, However, when cracking and some limited inelastic action occurs, as it probably will under moderate to strong shaking, values of 0.02 - 0.04 - 0.06 might well be judged appropriate, Stiff structures founded on soft soils show lengthened natural periods on account of the foundation system, which also provides enhanced damping perhaps 0.05, or more, under strong shaking. The mechanics of this damping includes the energy that is radiated away from the structure into the surrounding soil and, also, inelastic behaviour in the soil. For significant inelastic action that occurs in a controlled ductile manner in a structure a different approach is considered appropriate and we shall address this later. Records of strong ground motion taken with an appropriate accelerometer suggest that the forcing, Mil,(c), may be considered as an essentially random series of spikes, or impulses. From Figure 4.6, where a well known U.S, accelerogram (and integrated ground velocity and displacement curves) is portrayed, it may be noted that in this record of, say, 30 seconds duration, some 10-15 seconds covers the duration of strong shaking. Peak ground accelerations of 0.3g, or more, occur and there are several “peaks”, of alternating sign, in the vicinity of the largest value, Notice also that ground displacements of 100 mm, or more, occur ‘The acceleration trace gives an indication of the higher frequency (je. shorter period) content of the earthquake. The ground displacement trace shows the long period components in the ground motion. The words “essentially random”, used 10 describe the ground acceleration impulses, is a fair statement but it does mask the fact that there are 2 (limited) range of periods that characterize earthquake ground motions, from say Eqn. 4.19 Page 4-88 Acceleration ‘omvsec/sec 0.1 seconds to 10 seconds, as may be seen from Figure 4.6. This straddles the same range as the natural periods typical of civil engineering structures. ynamic analysis techniques IMPERIAL VALLEY EARTHQUAKE MAY 18, 1940 - 2097 PST A001 40,007.0 EL CENTRO SITE IMPERIAL VALLEY IRRIGATION DISTRICT COMP SOOE © Peak values: Accel = 241.7 em/secisec Velocity = 23.4 emisee Dispi = 10.9. cm 5 é 85 0 s a An L\ vey 20 ° 4.2.2 10 20 30 Time - Seconds Figure 4.6 Typical records of ground motion (from Ref. 45) ‘The higher frequency (shorter period) components in the ground ‘motion tend to damp out with distance from the source. Thus flexible structures may be set swaying by distant earthquakes which are imperceptible to those occupying stiff dwellings. Response to an impulse We stan here the process of working out the response of @ structure to earthquake loading, Suppose, starting time at r=0, a force F(t) ( applied for a time + Force, F(t) “Mil,(i) ) is Figure 4.7 Delivery of an impulse 40 50 Page 4-9Australian Earthquake Engineering Manuat Page 4-10 Define I= [PFU at. Eqn. 4.20 and if 1 << T(=2n/o) , we have an impulse. We suppose that the impulse is, say one spike of the acceleration record. During the time of the impulse we have the equation of motion Berod+0%s -FO~ Eqn. 4.21 M Integrating this, over a short time ¢, gives ee 1 Eqn. 4.22 = Afro axel A) -B) = PFO Ox= Fe ‘because there will not have been time for the oscillator to move. ‘After time t= t there is no extemal forcing and, so, the solution to an impulse is to be found as that of an oscillator undergoing free vibrations after an initial velocity (from the impulse - momentum relationship) of ier ieear Eqn. 4.23 b@= 77 [Hose The solution is i ot a Eqn. 4.24 Ae sinofi-@r , where 20 Maqi C ‘The peak displacement occurs at a time of = 7/4 which is usually well after the impulse has been delivered. ‘The value of the peak displacement is Pasi Eqn. 4.25 Mo ‘occurring at 1 = 7/4, provided damping is small. In this case the peak occurs shortly after delivery of the impulse and before damping has had much opportunity to reduce response. ‘The peak force iS Ping, = KOmax = OF Eqn. 4.26 Returning to the example of the wind shield we wish to interpret the result above. Figure 4.8 shows ground acceleration applied to the base of the tower and the equivalent loading distributed up the tower, the distribution factor being the shape of the first mode of vibration, eeeuynamec analysis techniques « —/ f (ay (it) a0 KEE * WE Figure 48 An impulse applied to the chimney. We arrive at this because the work done by this loading is force times displacement which, for the whole tower, is: Y Eqn. 4.27 WD = P(N), = ~Liimeri,)a,(2) a 7 so that the effective loading is ae -({fms (Jes Eqn. 4.28 = Meg ily where M,, is the effective earthquake mass of the tower. It, 100, differs from the total mass of the tower and differs from M, as well, For a uniform tower M:M,, = 3:1, while for one in which the wall thickness tapers completely M:M,,= ‘The equation of motion in such cases is M, Av2tod+o% = £0 2 Mey pees where @ = fK,/M General definitions of M,, and M, are Eqn 4 M,= fima(gce) 2 eee Mag = [meee where (2) is the modeshape. This analysis could therefore apply to a high rise building, for example, and we need not have confined our attention to just the first mode, The equation of motion above applies 10 any mode provided ©, @(2),.M, and My, are correctly interpreted, In matrix notation, the equivalent expressions are: Page 4-11Australian Earthquake Engineering Manuat Page 4-12 M,=6™Mo My = 6™ where Mis the mass matrix, @ is the eigenvector, and QT its transpose Example 'A reinforced concrete power station chimney shield has dimensions of h = 200 m, diameter of 17.5 m and wall thickness varying from 750mm at the base, to 250mm at the top. Calculations show that K, = 13400 kNm 1M, = 3400 tonnes Myq = 6250 tonnes and itis assumed that the concrete is not cracked. ‘The tower suffers a jolt of peak intensity (at base) of 0.3 g with a sawtooth distribution of duration 0.3 seconds. Find the peak base shear in the tower due to this impulse. Now o 2 rad [sec YM, so the natural period of the fundamental mode of vibration of the tower is T = 2nfur=3.1 sec Note that 1/T=0.3/3.1= 1/10 so the loading is impulsive. Remember time is not usefully judged in absolute terms in ‘dynamics; by comparing it to the natural period of vibration one has a relative scale which becomes meaningful. ‘The equivalent impulse delivered is 1 46250 0.3%9.81x0.3 2760 KN-sec the peak tower tip displacement, relative to the base. is and this peak displacement ocous at a time of 7/4 = 3, 1/4 = 0.8 seconds after the impulse is delivered, ‘The distribution of loading that gives rise to that deflection is @ matter of some interest: knowing the associated shear force and Eqn. 4.31 Eqn. 4.32 Eqn. 4.33 Eqn. 4.34 See eee ee tee eee eeeDynamic analysis techniques bending moment distributions is obviously an important matter, This Ties at the heart of earthquake engineering analysis; the determination of equivalent static actions from which design may proceed, (See Figure 4.9) Figure 4.9 Equivalent static actions that generate the same peak displacement response ‘The tower vibrates freely after the impulse is delivered with peak accelerations at each level given by 2 aa (2 Saul a o() ‘Thus the peak inertia forces have a distribution given by : (2) nu (2) with units of N/m, say ‘The base shear is = My 0” Day or S,= Mg nas ‘The overturning moment is Ma = Jf2m(2107 dau (2) ae 2M, 3, where Eqn. 4.36 Eqn. 4.36 Eqn. 4.37 Eqn. 4.38 Page 4-13Australian Earthquake Engineering Manual Page 4-14 ia Eqn. 4.39 jg the distance to the centre of effort of the inertia loading. Now Bgay = ©? Mipax = 4% 0.41 = 1.64 m / sec? Eqn. 4.40 or about 16% of gravity, and the base shear is S, = 6250 x 1.64 = 1.03 x 10° KN ‘Eqn. 4.41 ‘The total weight of the tower, (chimney flues) is W =2.25%108 KN Eqn. 4.42 0 the ratio of base shear to self weight is 5,103 |g o4s Eqn. 4.43 W 22.5 ‘The base shear coefficient is about 5%. If the structure has been perfectly rigid the peak base shear would have been the product of the total mass of the tower and the peak acceleration from the impulse, namely 4 2.25210" 9 39.81 Pee 9.8: = 6.810" kN which is almost seven times greater than the actual peak. The true peak base shear is much less because the structure, although massive, has a relatively long fundamental period. ‘As fas as the tower's overtuming moment is concemed, calculation shows that in this case Th 149m Eqn. 4.45 10 and, so Mo, = 140. 1.03 x108 = 1.44 «10° KN -m Eqn. 4.46 “The bending stresses associated with this loading are x08 Eqn. 4.47 oy = t Mas = Rt x1" 9950.96 . Z ~ wx(8375) x750 = 49.1 MPa while the axial stress, due to self weight of the chimney, of weight 1.32108 N, at the tower base, isuynamic analysis tecnniques 132 x10% EE = 3. MPa Eqn. 4.48 2x 8375x750 8, ‘The peak tensile stress is 9.1-3.3=5.8MPa gn. 4.49 which indicates that some cracking of the cross section is likely (cracking on both faces, incidentally, because of the oscillatory nature of the loading). The stress situation will be different, and possibly more severe, at other points in the tower. ‘The reader is urged to redo the latter part of this analysis for a section at mid-height. Because cracking will occur under this seemingly innocuous loading (which is, in fact, more severe than the design wind Joading) the natural period will be lengthened slightly and the inertia loads generated will be less. That is, the cracking, once begun, will reduce response peaks. Controlled light damage could be considered beneficial, if anything! 4.2.3 Response to an earthquake We may use the result for response to an impulse to construct a general solution for the response of a SDOF oscillator. A typical trace of ground acceleration looks like 2 seties of impulses (although, as has been said, there is a significant range of frequencies apparent in the input - recall that the trace of ground displacement often shows the longer period components) and, 1 so, for present purposes, we can assume the general solution has been generated specifically to address the response of the oscillator to ground acceleration, Figure 4.10 shows input which an be divided up into slices. One such slice is shown shaded. Force Figure 4.10 Definition diagram for construction of Duhamet's Integrat An impulse of magnitude AN = FEAT; Eqn. 450 ‘occurring at 1, gives rise to displacement of the oscillator at a subsequent time of Page 4-15‘Australian Earthquake Engineering Manual non ge S608) sino fT UE (¢— 4) B ino) — (t- 4). 2% Note that i,j) 4; is the incremental, or “initial”, velocity imparted to the oscillator at time +. That is, from the impulse - momentum relationship MA(t) or Aft,) aii (4) 8% =i, (1) 8% When we add up all the contributions from all the “slices” from t=20tot=twe have Aw= “Seah (rye sino —F (0-2) de, 120 ‘This is Duhamel's Integral and it is a formal solution to the oscillator equation for general input acceleration. ‘The formal solution is not much use unless simple numerical methods are employed using, say, Duhamel's Integral as the starting point. One useful, simple and accurate way of calculating A() at a series of time intervals is the linear force ‘method (see Ref 4.3). If the input acceleration is known at each (constant) time step, t, the response is A,=0 w-(breta (1+Gor)de= (Zip + 3g +a} (2a), (0-02) on'3 where k = 2, These answers will be accurate provided that wt <<1. In practice if the time step, t, is less than, say, one-tenth of the period, 7, solution accuracy is assured. ‘Typically, the input might consist of, say, 30 seconds of strong motion, If intervals of 0.02 seconds are used, some 1500 “calculations” will have to be done in order to find Apex and this for one value of natural frequency. @, and one value of damping, 6 ‘Add to this the fact that the input record used will, in all probability, bear litte relationship to a future event at the site, especially in Australia where so few records exist, and one sees the problems of time-history analysis, A lot of effort is required to obtain just one piece of information (peak response, say) and in any event, the problem is non-deterministic since the furure loading is unknown. Eqn. 4.51 Eqn. 4.52 Eqn. 4.53 Eqn. 4.54 Page 4-16Dynamic analysis techniques The sort of output that such numerical schemes generate is shown by Figure 4.11 which is the response of a 25 storey building to the Taft 1952 earthquake. The response quantity is the building tip displacement 03 02 4 ool 04 02 03 Figure 4.11 Response of a 25 storey building to the Taft 1952 earthquake ‘The plot demonstrates that the building responds at its own natural frequency to this “random” input (the building period is about 2.5 seconds), and that there are several peaks near the maximum tip displacement of 300 mm. Longer period structures such as high-rise buildings tend to behave like this in earthquakes. Example Retuming to the chimney shield example, it is noted that for response confined to the first mode of vibration, we have to solve aes, Eqn. 4.55 A+ 200A + 0 Thus, from Duhamel’s Integral, the solution for the tip displacement is Meg/Me t! (a) bets A(t) = “ope b g(r)" and the ratio M,,/M, is the multiplication factor by which the tower is made equivalent to a SDOF oscillator. It is usually called the panicipation factor. sinool— 2 (e— hat eats For the particular tower here Eqn. 4.57 so, the input acceleration needs to be scaled up by this amount in order to make it relevant to the tower, Example We can use the recursion formula given earlier to solve siep-by- step the case of the response of the tower to the sawtooth impulse. The period of the tower is 3.1 seconds but we take the Page 4-17Australian Earthquake Engineenng Manual Pago 4-18 lime step 10 be T= 0.15 {if we took t= 0.3 seconds, we would wipe out the input!), so that +/7°= 1/20 which is plenty small enough. Damping is set at zero 10 be conservative, that 6 +0. ‘The right hand side of the equation has to be scaled by 1.84 as mentioned above. Thus we have as input Eqn. 4.58 .84 x 0.39.81 = 5.42 m/sec li, =0 and, since 7 = 0.3, we have Ay =O Eqn. 4.59 Ay =[Fxo+dxs 42)x0 15? = 0.0203 m tt 2 1 2 ay= arc tas 424 Ex0 x0.15* +1.91x 0.0203-0=0.120 m a= (3x5 s2+2xo42x0)xoas" +191 0.120-0.0203 = 0.229 m A, =0+191x0. 229 -0.120 = 0.318 m Ag = 1.91% 0.318 - 0.229 = 0.878 m A, = 1.91x0.378-0.318 = 0.404 m 910.404 - 0.378 = 0.394 m ‘The peak response is B= 040 m, occurring at a time of 6x0.15 =0.9 seconds. These results compare favourably with the early, approximate, analytical results. ‘We can retum now to the mainstream of this example which is to calculate shear force and bending moment diagrams corresponding to the peak seismic load. For a given earthquake the peak response for an oscillator (of period 3.1 sec and fraction of critical damping, say, 0.02) may bbe labelled Apa, The peak tip displacement of our tower is, Meg Eqn. 4.60 imax = 1-84 Aray M, ‘This translates into a distribution of peak acceleration up the tower of Me, Vz mg, Od 347184 ( 55) 7.36Boue(si5) m/sec? "= 200/ 200 ‘The mass per unit height of the tower isDynamic analysis techniques 4.2.4 150|1 (ss) tonnes / m Thus the equivalent static loading to be applied to the tower, in order to represent its peak tip displacement response under the earthquake. of 1.84 yay iS ro ~ (35)) x raz) Snax 2010-2) (5) an For example, at the tip of the tower where, say, the maximum displacement is 0.5m under a particular earthquake, the equivalent lateral load is 1) 0.5 1.10 rei 3) 1-5 = 150 KN x 5) Ugg 7 SO KN In such a case the distribution of equivalent static loading, representing the worst effect of this “earthquake” is shown by Figure 4,12. This should be studied in conjunction with Clause 6.2 of AS11704 1993 Minimum Design Loads on Structures Part 4 Earthquake Loads. 0.5m 150 50m 7 0.64 x 10" 0.17 x 10° 50m, “ 1.05 x 10" 0.63. 10° 50m 18 1.22x 10° 1.19 x10" Som. EE 1.25% 10 1.80% 10° Peak Relative Equivalent External Shear Force. Bending Moment Deflection Loading (kNim) Diagram (kN) Diagram (kN-m} Figure 4.12 Equivaten: static actions for a 200 m high chimney responding to an earthquake. Elastic Response Spectra Because of the effort required to evaluate peak response for one Structure responding to one (fundamentally artificial) earthquake, and in view of the necessity of considering several seismic “events” in order to provide a better estimate of Possible loading, the concept of a response spectrum has been developed. Figure 4.13 shows a record of peak relative displacement response of a SDOF oscillator can be plotied for a given Eqn. 4.62 Eqn. 4.63 Eqn. 4.64 Page 4-19‘Australian Earthquake Engineering Manual earthquake, given damping, and a range of periods, typical of structures. EL CENTRO, S0GE COMPONENT MAY 18, 1980 04g = ° ge Lee eee eeeeeeeee peer ° 10 20 30 Timo, seconds 10 Hae * 2.48 T=05 sec g- 0.02 8 - g = a T=1s0c 2 E 0.02 = 3 & > Eros Ht ° 10 20 30 Tine, seconds 20 Deformation fr displacement) 3 18 F Response Spectrum Bag | Se2percen 4 aos — ° I ° 1 2 a Natural vibration poriod, T seconds Figure 4.13 Compilation of (relative) displacement response spectra (rom Ref. 44) ‘This is the relative displacement response spectrum for this carthquake for structures exhibiting this level of damping Response spectra are particularly important in practice: the first use was in the design of landing gear for aircraft, where runway roughness is obviously of concern. Changing the value of damping and repeating the other steps produces another curve (higher damping tends to reduce the jagged nature of curves of maximum displacement response). Using a different earthquake as input makes for a whole new round of changes, and so on, Page 4-20Dynamic analysis techniques The peak relative displacement is usually called S, and the peak strain energy of the oscillator is SE.=2KS3 Eqn. 4.65 seciu< ss or SE. or SE. Where we define the pseudo-relative velocity spectrum as S, = Ss Eqn. 4.66 and we see an “equivalence” again between strain energy and kinetic energy. The peak force generated by the oscillator is Fi =KS, Eqn. 4.67 ot Fan =MXS, M or Fax = MOS, or Faux = MS, ‘Where we define the pseudo-absolute acceleration spectrum as 5S, = 0S, Eqn. 4.68 ‘These definitions are very useful in practice. Figure 4.14 shows Sav, and S, for El Centro, 1940, with 2% damping. Interpreting these definitions for a structure, such as our tower, that has been rendered into its SDOF equivalent, proceeds as follows. The participation factor M,q/M, is used to scale up Sx S, and S, since they all apply only to a genuine SDOF oscillator. Page 4-21Australian Earthquake Engineering Manual 20 1 15 (a) & < b) 15 T T i wo 4 fe © a 0s 4 fig seta Vaasa Raaios anos ease ° 1 2 3 Natural vipration period, T sec Figure 4.14 Diagrams for (a) Displacement, (b) Pseudo-velociy,(c) Pseudo-acceleration response spectra. El Centro, 1940, 2% damping, (rom Ref 4.4) ‘The peak strain energy is Eqn. 4.69 and the peak force in the oscillator is Page 4-22Dynamic analysis techniques ‘The peak base shear for the tower is (See Eqn. 4.37 in this chapter) M, or 5,=—<2F, mi and we see that we must apply the participation factor to the peak equivalent force in the oscillator in order to obtain the true peak base shear, ‘Two examples now follow. Example Recall the diagram of Figure 4.5 which is of a single stem reinforced concrete footbridge pier. Suppose the pier is 7m high, from top of footing to centre of gravity of bridge superstructure, The tributary mass at the cross-head is 40 tonnes, together with some 7 tonnes of cross-head and part of the pier itself, and the (uncracked) pier stiffness is 1000 kN/m. Assume damping of 2% critical. ‘The natural period is i a Rajoe = 2] K 1000 From the previous diagram of the response spectrum for El Centro 1940 itis seen that T= 1.36 seconds S45 inches = 125mm but, given that the calculation of the period is approximate, Sy could lie between 100 and 150mm, as may be seen from the sesponse spectrum This corresponds to a base shear of Eqn. 4.70 Eqn. 4.71 Eqn. 4.72 Eqn. 4.73 Page 4-23Australian Earthquake Engineering Manual 4.25 ‘The base shear cocfficent is 5,5, Ww 470 0.27 that is about 27% of gravity. Such Joading is quite severe since the peak bending moment of 7 x 125 = 875 KN-m is to be resisted by a single stem, We shall retum to this example when inelastic response spectra are considered. Example ‘The tall chimney has a period of 3.1 sec, and for this period (and 2% damping assumed), and the El Centro 1940 response spectrum, Sy 15 inches 375mm ‘which means that the tip displacement is M, a 54 = 1.84 x 375 = 690 mm because we have to make an equivalent SDOF model of the tower. “The peak base shear is 2 (#4) K, Sq = (1.84)? -13400-0.3 M, = 70 10% KN which is about 8% of the total weight of the tower. ‘The overtuming moment is My = 7S, =140x1.7x 108 = 2.38108 kN-m the peak bending stresses at the base of the tower are +15 MPa, and there would be no doubt of significant cracking occurring under this loading. This will tend to reduce the peak lateral loads and also necessitates a more sophisticated approach to the calculation of stresses! Design Curves and Inelastic Response Spectra Results for one earthquake give, at best, some slight indication of response if, in the future, an earthquake of the same peak ground acceleration, similar frequency content and duration scrikes. Since this is improbable, results of response to several similar earthquakes could be assessed and smoothed curves obtained. Eqn. 4.74 | | Eqn. 4.75 Eqn. 4.76 Eqn. 4.77 Page 4-24Dynamic analysis techniques For example, an extreme case of such smoothing would be to take the pseudo-velocity spectrum for El Centro, 1940, and replace it by the curve in Figure 4.15, Sy mise or 08 period, T see Figure 4.15 An idealised (elastic) design pseudo-velocity response spectrum This bears some relationship to the actual result (look back at Figure 4.14) and, it could be argued, is a smoothed design curve, More specifically, the associated pseudo-acceleration response spectrum is then shown by Figure 4.16. Si misec* ae | eee secre period, T Figure 4.16 The associated idealised (clastic) design pseudo- acceleration response spectrum From here it is @ short step 10 curves such as those given in ‘codes of practice, like that in Figure 4.17. This, in fact, is the basic design curve for Auckland, New Zealand (according to NZS 4203 1976). Reference should also be made to figure 7.2 on page 46 of AS11704, 1993 which is similar to Figure 4.36 above ‘The question that immediately arises when one compares the curve in Figure 4.17, with the one in Figure 4.16, is why the peak acceleration has been limited to 0.1g when it is clear that peak response accelerations of $0 — 100% of g could be expected for lightly damped structures with periods less than 1 second? Page 4.25Australian Earthquake Engineering Manual Page 4-26 t Design acceleration cwsec* 12° poriod, T sec Figure 4.17 A design response specirum ‘There are two parts to the answer, The first part relates to damping which during strong shaking is likely to be somewhat greater than the 2% used for the response spectrum shown. A value of 5% would be more realistic and this reduces the peaks and jagged nature of the curve significany. The other part, which is not unrelated, involves the issue of design philosophy. It is widely accepted that it is uneconomical to design structures to respond elastically to moderate-to-severe shaking. Rather it is better to adopt a limit state approach by insisting, say, that, while a structure must remain standing after a strong motion event, controlled damage (with consequential significant energy dissipation) is acceptable. By allowing the structure to yield, in a controlled fashion, we limit force levels in the structure and, if the structure is ductile, the energy the earthquake supplies may be dissipated by inelastic action. Thus yielding is equivalent to the introduction of a fuse: force levels are limited to predetermined levels, but the structure must be detailed in order 10 be able to sustain several load reversals at this level. This is the Damageability Limit State Gee Chapter 5 - Design criteria and analysis options) One way of establishing the approximate level of force reduction possible is t0 note with respect to Figure 4.18 that if the structure has the possibility of exhibiting a simple elasto- plastic curve, the structure may be considered ductile, In ductile structure, or sub-assemnblage, the resistance, R, may be sustained at displacements that are several times those at first yield, A,Dynamic analysis techniques Resistance yield J Peek deflection x unloading Deflection 4, ro Figure 4.18 A simple bi-linear elasto-plastic curve of response. representative of ductile performance With reference to Figure 4.18, the area under the curve, up to the instant of maximum displacement, is 1 Eqn. 4.78 Rb max 7 Ry or 2Rq,| 24m) 2A, 1M op, or s(n -1 7K een) where 1 (=Angx/y) _ is the ductility factor and S(=R/M) _ is the reduced pseudo - spectral accelleration The peak energy supplied is Lat oo Eqn. 4.79 2K"* tus? 2 and equating the two yiclds the modified (je, reduced) pseudo- spectral acceleration, S,” , namely oe Eqn. 4.80 Pac and the resistance to be supplied is os, = MS Eqn. 4.81 Ro MS, = ee This formula tends to fit the results of more detailed non-linear elasto-plastic analysis best for structures of shorter period, For longer period structures the simpler formula remy = HS Eqn. 4.82 seems better and it, too, can be shown to have some physical basis; being, in this instance based on the assumption of equal deflections imespective of whether purely elastic or elasto- plastic behaviour is assumed. Notwithstanding this, there is evidence that the first formula for R represents an upper bound Page 4-27Australian Earthquake Engineering Manual Page 4-28 so that for example, for given resistance the ductility demand is, if anything, less than that given by that first formula (Ref 6.6) It is seen, therefore, that structures in which controlled, cyclic, {inelastic response is possible in their various parts may be designed to lower force levels (ductility factors of w= 4 - 6 are Jooked for in order for structures to survive moderate-to-strong shaking) and the energy “supplied” by the earthquake is dissipated by this ductile behaviour, while force levels in the structure do not exceed predetermined levels. This philosophy underlies much of modem practice in carthquake engineering in the earthquake-prone areas of the world. The concept is not inappropriate 10 Australia, indeed, it is contained in AS1170.4, 1993, The philosophy is bound to reccive more attention in ‘Australia in the aftermath of the Newcastle Earthquake, albeit in the context of limited ductility for many structures, in all probability. See Chapter 6. Example “The single stem reinforced concrete footbridge pier (recall the ‘example on page 4.23 of this chapter) has an ultimate moment of resistance of about 2.8108 N-mm. The cantilever stiffness of the pier, when some allowance is made for cracking, is about 750 kN/m and the total mass, including an allowance for the pier and cross-head, is about 47 tonnes. A pseudo spectral velocity of 5, = 0.65 M/sec is the (severe) design case and the question is, what inelastic displacement results? The ultimate moment corresponds to about 8% g as a lateral design acceleration. ‘The total kinetic energy to be absorbed is $xd7 0.68 = 9 KN m In terms of moment-rotation behaviour we have M=Fh=KAh and in our case Kg = 750 700° = 3.68 x 10'° N- mm / rad ‘The rotation of the pier when the design ultimate moment (the approximate ultimate strength of the pier) is reached is, 2.8108 tte 3.68% 10 71.6 x10 rad. Eqn. 4.83 Eqn. 4.84 Eqn. 4.85Dynamic analysis techniques ‘The energy absorbed at the plastic hinge in reachi maximum r0Lation, Braga, iS the 1 Eqn. 4.87 My 0,(2u-1) 7 where | = 0,,9,/8, = rotational ductility factor for the pier. ‘Thus, in order to find 1 we have, from the appropriate form of the energy balance tee Hee Eqn. 4.68 My: 6,(24 =1) «Last 99 $52,810 «7.610 or 2n— or ‘This ductility factor might be achieved but only by confining the high moment region at the base of the stem with closely spaced, well anchored. ties. Loss of confinement by tie fracture or pull-out is often the means by which failure of reinforced concrete structural members occurs in strong shaking. See Chapter 6. ‘The coresponding peak lateral displacement of the pier cross- head is Simax = 5.2% 7.6 X10 x 7000 Eqn. 4.89 = 280 mm. ‘The moment due to this shift in the line of action of the tibutary gravity load is Mp.,, = 470 280x 10" = 32x 10° N-mm. Eqn. 4.90 Mpg _ 132x108 _ MS = a 7ex108 = 947 This is a significant fraction and implies that we should have allowed for the “P-A” effect in or original equation, With reference to the Figure 4.19, the gravitational potential energy lost by the superstructure, when a rotation of Oy... OCCUTS ( @ ees Eqn. 4.91 COSA nox) man 1a |) = MghOa, Mgh(i Page 4-29Australian Earthquake Engineering Manual Ute Figure 4.19 Diagram for the plastic “P-A" effect ‘Thus the statement of energy conservation becomes 1 lene? =< ims? Eqn. 4.92 My Ornax ~ My 8) 5 Mah Orns =gMs Let us now solve this quadratic equation, bearing in mind that, since we are looking for a solution which gives a larger value of nan than before, but which gives our earlier value if Mgh is ‘small, we take the negative root of the quadratic, Meh = x 425 x 7000 x 1000 = 1.6510? N- mm Eqn. 4.93 ust =9.9x108 N-mm M6, = 1.05%108 N-mm M,= 2.8108 N-mm so 1.65 10° Ob ay — 2.8 X10" Opn + (9-9 + 1.06) x 10° = 0 or 2.8108 - V2.8? x 10" — 4 x 10,96 x 1.6510" 2.8210! — v2.8" x10" —4 10.56 x 1.65210 2x 1.65 x10" 2.8% 108 — 0.78 x 108 2x 1.65 x10" = 6.1x107 rad so that Eqn. 4.94 81 Thus, a better estimate of the ductility factor p=8 and the likely permanent set on the pier after this earthquake is 77.6 x10 x 7000 = 370 mm ‘The pier would be on the verge of collapse under this severe loading, loading that has been exacerbated by self weight. Page 4-30Dynamic analysis techniques Collapse of such a structure, while not necessarily important in itself, becomes a crucial issue, if for example, the footbridge crosses a freeway, and the pier is in the median strip. This “plastic P-A effect” inclines the structure, if and when it enters the inelastic region of its response, to vibrate about a mean position which is not vertical. More modest loading, more typical of levels of loading in Australia, say with S,=0.35 m/sec, yields a value of 1 = 1.6 signifying limited, repairable damage with little, if any, P-A effect present, A lot of useful earthquake engineering analysis can be done using approaches such as those presented herein. The reference list for this chapter is by no means ‘comprehensive but all of the works are worthy of consultation. Page 4-31Australian Earthquake Engineering Manual 43 Page 4-32 Example Seismic analysis of a 25 storey frame A 25 storey braced frame siee! building has the following properties K,= 18x10 kN/m together with a mass per unit height of ‘m= 20 tonnes/m ‘The mass of the building is uniform over its height of 100 metres. The braced frame has been designed so that stress levels in the chords under lateral loading are more or less constant with height. Thus an appropriate first mod shape is the parabola. alet)=ale (2) Find the following quantities for the equivalent SDOF system: i) The effective mass M, ii) The effective earthquake mass, Muy iii) ‘The natural period iv) The participation factor Using that information, and that contained in the response spectrum for 2% damping, (see figure 4.20), find i) the peak tip deflection ii) the peak base shear iii) the overtuming moment ‘Sketch the form of the equivalent loading diagram, (what is the peak lateral load applied to the building?), the shear force diagram and the bending moment diagram. Calculate the peak inter-storey drift atthe top of the frame and comment with respect to non-structural damage. Calculate the peak acceleration at the top of the frame and ‘comment with respect to the loading on sensitive equipment that might be housed at that level. ee (Ze 24mDynamic analysis techniques g 18 epee eect oeeeee Toe 10 Hy 4 4 os 4 ° See Pee eee ° 1 2 3 Natural vivration period, T soc © cy Figure 420 (a) Deformation (or displacement, (b) pseudo-velocity ‘and (c) pseudoacceleration response spectra. El Centro ground motion - S00 E component, 1940. Damping ratio & = 2 percent from Ref 44 Worked solution i) The effective mass, M, a, = fim(2(2) ae ohoelZ) ml = 122 since mass is uniform 20x 100 = 400 tonnes This is the effective mass tributary to one frame. Page 4-33Australian Earthquake Engineering Manual ii) The effective canthquake mass, Mey f 2 My=[ mie #) de ml : 3 since mass is uniform 20100 3 = 670 tonnes ili) The natural period is, given K,=1.8%108 KN/m, T= 2x(M,/K,) = 2x(400/1800)* =3.0sees, that is a long period structure. iv) The panicipation factor is Mu _ 670 M, ~ 400 =167 which is broadly typical of the first mode of high rise structure. For the first mode Aw (3) shape for fundamental mode, period 3.0 seconds, artcipaton factor 1.67 100 m Ke = 18% 10° kvm M, = 4001 Meg = 6700 From the El Centro ($00°E) 1940 displacement spectrum for 2% damping, Page 4-34Dynamic analysis techniques Sq= 15 inches Peak tip deflection ii) Peak base shear s=(H Me = Melk) = 1.67 X1.8x 10° x 0,635 = 191x109 EN S __ 191 x10? WwW” 20x100x9.81 = 9.7% Base shear coefficient ~ 0.1. This is likely to be a more severe loading than wind, Overtuming moment My =2 Sp fLeme{z) ee Lima =BA/4 =3x 100/4 =75m M,, = 75x 191x 10° = 143% 10° KN-m iv) Sketches M,, ) y Equivalent static loading = tayo" s.(2) 2 =20x167%{ 25) 0.88% 1 GB =55.7KN/m Page 4-35Australian Earthquake Engineering Manual 100 m 191 x 10° kN My = 143 x 10° KN. Equivalont SFO ‘BMD static loading Check = 1.86 10° KN ‘The peak inter-storey drift atthe top of the frame is 2 635 — (28 635 = 50mm \a00 ‘This will place severe constraints on structural details such as windows, panels, and partitions. The frame may need to be stiffened to reduce this potential problem. In pure flexing structures, imer-storey drifts are greatest between the upper floors. In pure shearing structures, inter- slorey drifts are greatest between the lower floors. Since most high rise structures are a combination of these two, inter-storey drifts always warrant special investigation. See Chapter 6. ‘The peak acceleration at the top of the building is, from Newton's second law, fim 35.7/20 = 2,8m/sec? ‘This is 28% of the gravity and may be important for the design of the attachments for heavy cladding. Expensive electronic equipment may also be affected so it may be necessary 10 isolate such equipment on special sub-floors in order 10 preclude the possibility of damage. Page 4-36Dynamic analysis techniques References 44 IM. Biggs, “Introduction to Structural Dynamics”, McGraw-Hill, New York, 1964 42 RW, Clough & J. Penzien, “Dynamics of Structures”, Wiley, New York, 1975 43 HM, Irvine, “Structural Dynamics for the Practising Engineer,” Unwin Hyman, London, 1990 44 AK, Chopra, “Dynamics of Structures: a Primer,” Earthquake Engineering Research Institute, Berkeley, 1981 45 DE. Hudson, “Reading & Interpreting Strong Motion Accelerograms”, Earthquake Engineering Research Institute, Berkeley, 1979 46 NM, Newmark & W.J. Hall, “Earthquake Spectra & Design”, Earthquake Engineering Research Institute, Berkeley, 1982 47 G.V. Berg, "Seismic Design Codes and Procedures”, Earthquake Engineering Research Institute, Berkeley, 1983, 48 G.W. Housner and P.C. Jennings, “Earthquake Design Criteria", Earthquake Engineering Research Institute, Berkeley, 1982. Page 4.37Australian Earthquake Engineering Manual Page 4-38| 5 | Design Philosophy and the Code Approach 5.1 5.2 Introduction Past Australian building practice involved little, if any, provision for resistance to seismic loadings. Accordingly, the relatively minor earthquake at Newcastle in December 1989 (Richter Magnitude 5.6) resulted in a disproportionate level of damage and economic loss to the community. In monetary terms, this earthquake was the most expensive national disaster in Australia's history. No pan of Australia can be guaranteed immune from such an event and similar levels of damage may reasonably be expected in any built-up area. Levels of earthquake resistance supplied in future structures, as well as the refurbishment and/or retrofitting of existing structures, will be determined ultimately by the price sociery is prepared to pay: clearly a balance must be struck for it is not realistic, nor possible, to actually design for all eventualities. In this Chapter we discuss the place and relevance of earthquake loadings in the design cycle, various design criteria including that recommended as appropriate for Australia; before ‘outlining the analysis options open to the earthquake engineer. Earthquake loading and the design process In areas of the world recognised as earthquake-prone, and having a well-monitored history of damaging seismic events, it is essential for a preliminary structural design to be based on some combination of dead, live and earthquake loadings. In these circumstances resistance to earthquake loading will dictate choice of structural configuration. An important feature of such structural configurations will be the existence of skeletons by which some measure of integrity is guaranteed as the structure responds to the rather Aaphazard series of impulses typical of strong ground motions In the design process it should be bome in mind at all times that the design loadings for future earthquakes, however derived, will always be subject to a degree of uncertainty. Because of the low level of seismicity in Australia, preliminary structural designs will almost always be dominated by a combination of dead, live and wind loading, with earthquake loading being applied 10 check if design refinements are required, Such refinements may include altematives to structural configuration, modification of structural elements and improved detailing (such as a more appropriate distribution of flexural and shear reinforcement if the structure is reinforced concrete). The aim is to provide a structure that is resilient to Page 5-1Australian Earthquake Engineering Manual 5.3 5.3.1 Page 5-2 eanthquake attack. If available records of ground acceleration versus time (je, time histories) for a particular region, or locality, can be processed to provide appropriate response spectra, these spectra may then be used either to assess the: earthquake loading that a particular structure might attract, or they may be used as the basis for a complete structural analysis under carthquake loading. Another possibility is that these local records might have been gathered with a view to analysing a particularly important structure (such as a nuclear power plant), and the designer would proceed to carry out real time dynamic analyses of the structure (See Chapter 4, Section 4.2.3, Response to an earthquake). However, for most buildings and engineered structures estimates of design earthquake loadings may be obtained by following 2 series of simple steps, as outlined in a code. The loading provisions in earthquake codes around the world follow similar principles and provide a means of realistically assessing an equivalent set of static loads to be applied to a structure in order to represent an “average” future earthquake. If the structure under consideration has fearures, such as unusual distributions of mass and/or stiffness, which call into question the validity of the simplified code approach, the designer is faced with considerable difficulty in realistically defining an appropriate earthquake loading. (It should be noted that, if loadings can be defined, appropriate techniques for structural analysis are readily available, for the most part) Notwithstanding the obvious difficulty of reliably predicting future earthquake events on the basis of a knowledge of past events, the present lack of adequate records makes this task almost impossible in Australia. The gathering of a statistically significant number of appropriate records of strong ground motions for any given region will take many years. In practice every effort should be made to utilise whatever information and experience is available 10 derive suitable response spectra for structures for which the simplified code approach is inapplicable. Inespectve of these effors, as a very minimum, the designer must carefully consider designing details that ensure a reasonable degree of structural resilience is provided in buildings and other engincered structures. This, presupposes that adequate load paths exist by which seismic loads may pass through the structure. This exercise, in itself, is, important, especially where geometric complexity exists. Design criteria and analysis options Design criteria In areas of the world recognised as being prone to major earthquakes the engineer is faced with the dilemma of beingDesign philosphy and the Code approach required to design for an event, which has only @ small chance of occurring during the life of the facility. If tne designer adopts conservative performance criteria for the facility the client (often society) is faced with costs which may be out of proportion 10 the risks involved. On the other hand, 10 ignore the possibility of a major earthquake could be construed as negligent in these circumstances, To overcome this problem a dual design philosophy has been developed, by which procedure: 1. A_ moderate earthquake, such as may reasonably be ‘expected at the site, is used as a basis for the seismic design. The facility should be proportioned to resist such an earthquake without significant. damage. This “damageability” limit state should ensure safety, limited non-structural damage and the continued performance of facilities and services, particularly those with important post-earthquake functions. The list includes hospitals, police, fire and civil defence facilities, water supply, telecommunications, electricity generation and distribution systems, ete. Almost as important is the maintenance of road and rail communications, particularly for food distribution (including warehouses and their contents). Similarly, the protection of industrial complexes, in their own right, as well as the protection of individual items of equipment in other buildings and facilities, is a necessary consequence of adoption of this limit state. ‘The most severe, credible carthquake that may be expected to occur at the site is used to test safety. In this ultimate limit state, significant structural and non-structural damage is expected but, neither collapse nor loss of life should ‘occur. In Australia, given the levels of seismicity expected, a structure designed for the ultimate limit state (avoiding collapse) will also usually satisfy the damageability limit state. However, in a ly scismically active area, the structure must be designed for both limit states. This is because the loads induced under the damageability limit state earthquake will often exceed the clastic limits of the structure, This would rarely be the case in Australia It is this philosophy which really forms the basis for the future practice of eanhquake engineering in Australia. Figure 5.1 shows flow charts of the design processes for both an earthquake-prone area and a low risk area, such as Australia. Earthquake prone areas 23, Now Zaslang, span, USA ‘Sune Europe, Pads’ Now Guinea lngonesa, Pe Pappnes, Made ease Chas To peso or doa, [7] fro, vind apd Earmquake Peroorance Shea t io} Saageatiy Lint 0] State Tes |_| Earthquake Performance citer 2 io] Ulimate Limit State Te Acceptance Low seismic risk ereas Aunt, Unie King Ehaaca, nermem Curse Design for dead, We and wine badings Yes Eantiquake Performance Grieria 2 No] Unimate’Linit State yes ‘Acceptance Fig. 5.1. Comparative flow charis for design processesAustralian Earthquake Engineering Manual 5.3.2 Analysis options Essentially, an engineer is faced with three possible methods of analysis/design for earthquake loading on a structure located in Australia. In all methods the dual damage criteria discussed above may be applied. The three analysis/design options available are: 1a) to perform a real time dynamic analysis of the structure, b) to perform a dynamic response spectrum analysis of the structure and/or ©) to use the recently released AS11704 - 1993, Minimum. Design Loads on Structures Part 4: Earthquake Loads, to conduct an equivalent static analysis, Both of the procedures (a) and (b) involve analysis probably using a finite element computer mode! of the structure, (It is likely that a suitable model will have been developed in the conventional design process when considering dead, live and wind loadings). Real time dynamic analysis. ‘The method involves subjecting an appropriate finite element ‘computer model of the building, or structural system, to a given, previously recorded, earthquake record and examining its response in real time, Response peaks are generally of most interest. There are certain special circumstances where this procedure is useful but, for general aseismic design, it is of little value as the actual earthquake that the built structure may have to resist cannot be guaranteed to have sufficiently similar characteristics 10 the design earthquake. In particular, the intensity, duration and frequency content of the earthquake may be unsuitable especially if, as often happens, the record comes from another country or continent. Moreover, the method is expensive and time-consuming so that only for special structures can its use be justified. 1f, in addition, inelastic response calculations are involved another level of complexity (and uncertainty) is introduced. Response then becomes dependent, often heavily so, on the nonlinear models chosen and this is in addition to the inherent uncertainties involved in choosing to use one particular record. Response Spectrum approach ‘This procedure involves the development of an appropriate table, or sequence of peak response values, for use with a finite ‘element computer model of the structure. This sequence, or spectrum is developed by analysis and statistical manipulation of an ensemble of existing earthquake records. The procedure to develop a response spectrum has been detailed in Chapter 4, but an outline of the process is as follows: Page 5-4Design philosphy and the Code approach a) A single degree of freedom system with a particular mass, stiffness and damping is selected. The associated natural period is calculated. b) A particular digitised earthquake record is selected and this is used as the input ground motion to the single degree of freedom system. A real time analysis is performed to provide a solution for the variation of acceleration, velocity and displacement of the single degree of freedom system with time. ©) The maximum acceleration (or velocity or displacement) is selected and this, and the natural period of the system, are plotted as the ordinate and abscissa, respectively, of a Tesponse spectrum graph, ) The period of the system is altered, ‘The real time analysis is repeated with the same earthquake input and the maximum acceleration (or velocity or displacement) is again identified. Plotting of the new value of period and the new maximum acceleration provides a second point on the response spectrum. ©) The process is continued across a range of periods thus providing a response spectrum curve associated with a particular earthquake and a panicular level of damping, £) The process may then be repeated for different levels of damping (usually zero 10 10% of critical damping) to produce a family of response spectra associated with @ particular earthquake. ) The complete procedure may then be repeated for another “appropriate” earthquake thus producing another series of curves. h) By averaging, smoothing and statistically manipulating the set of curves associated with a particular level of damping a “design” response spectrum may be obtained. Care must be taken as to which earthquake records are incorporated as “design” response spectra as earthquake records for various pants of the worid have been found to have markedly different characteristics. The first design spectra were produced in the 1950's. A typical response spectrum is shown in Fig. 4.14, References 4.5, 4.6 4.7 and 4.8 may be consulted. ‘The actual response spectrum analysis technique for a structure involves the use of an appropriate finite element model of the structure in conjunction with a selected “design” response spectrum. The essence of the analytical procedure is as follows: a) The first few natural frequencies and modes of the finite element model of the structure are catculated, Usually it is only the first few lowest modes that are of interest. b) At each of the frequencies (or modes) the corresponding, value of peak acceleration is determined from the response spectrum. Page 5-5Australian Earthquake Engineering Manual Page 5-6 ¢) Each of these accelerations is then applied 10 the model. Since the response spectra gives only the maximum esponse, the maximum forces associated with each mode are obtained 4) The total response is then obtained by superimposing the ‘modal responses (Note that the actual time variation of the design earthquake motion is unknown, and therefore it is impossible to compute the time variation of response either for a modal component, or for the total). ) A conservative upper bound for the response may be obtained by performing either a root of the sum of the squares (root mean square, SRSS) of the model maxima or a complete quadratic combination (CQC). The latter procedure is recommended, particularly if the natural frequencies of some of the modes under consideration are similar. Quasi-static Code approach Both the former Australian Earthquake Code AS2121 and the new AS11704 - 1993, in common with almost all other Earthquake Codes around the world, use an equivalent static lateral loading applied to @ structure to represent earthquake loadings. In the case of buildings e quantity usually referred to as the “total base shear” is calculated from the product of the ‘weight of the building and a coefficient. This coefficient takes into account the location and importance of the structure, its ductility or energy absorption capacity, its dynamic characteristics and the local soil conditions and their effect on structural responses, (Once the total base shear has been calculated it is distributed up the structure as a series of horizontal loads at each floor level and the structure is analysed with these equivalent horizontal loads applied. ‘Structures must be designed to resist earthquake induced forces applied in any horizontal direction and these forces may be assumed 10 not act concurrently in the direction of each principal axis, except for certain cases including some irregular structures. In such cases the structural components and footings must be designed for the additive effect of 100% of the horizontal earthquake forces in one direction and 30% in the perpendicular direction, The torsional effects discussed in Chapter 3 Section 3.2 (Structural configurations) are accounted for by applying a torsion, due to a combination of “accidental” eccentricity and a factored static ecceniticity at each floor,Design philosphy and the Code approach 5.4 5.4.1 5.4.2 Australian Earthquake Codes Building Regulations A variety of State and Local Government by-laws contro! building design throughout Australia, Building regulations are uniform throughout the States but there are differences between the States. ‘The Standards Association of Australia issues standards which are of an advisory nature only, unless or until the standard is legally adopted by a Local, State or Commonwealth government, The new “Building Code of Australia” (BCA), issued by the Standards Association, is progressively being adopted by all States and it is intended that this will replace the existing state-based building regulations, The Building Code of Australia has adopted AS11704 - 1993 and as the States progressively adopt the BCA, the Earthquake Loading Standard is also being adopted. It seems unlikely that any State will opt for a variation excluding earthquake loading from consideration in the design process. AS1170.4 Minimum Design Loads on Structures Part 4: Earthquake Loads This Standard provides data and sets out procedures for determining the minimum loads induced in structures by earthquakes. It also sets out minimum detailing requirements for structures. The Standard is intended 1 be applicable to structures, particularly buildings, non-building structures, fixings and non- structural components including building services and architectural elements, So called “special structures”, such as nuclear reactors, dams, transmission towers, piers and wharves are not covered by this Standard. Structures may be classified as special if their failure under earthquake induced loading must be avoided at all costs or they have a particularly non-uniform vertical or horizontal distribution of mass and stifiness, ‘The earthquake induced loads calculated using this Standard are ultimate loads and must be combined with other relevant Joadings expected at the time of an earthquake, Whether or not it is necessary to design a given structure to specifically resist earthquake induced loading depends primarily on the location of the structure, the structural configuration and the importance of the structure. If it is necessary to design the structure against earthquake induced loading, the loadings may be determined either by an equivalent static analysis or a dynamic analysis ‘The design requirements depend on the structure configuration and ductility and the Earthquake Design Category. The Page 5-7Australian Earthquake Engineering Manual Page 5-8 Standard provides detailed guidance as to how to determine whether the configuration of the structure under consideration makes it regular or irregular. It also provides guidance on assessing the ductility or energy absorption capability of the structure, Both the regularity and ductility of the structure directly affect the design requirements. ‘The Earthquake Design Category is established by considering the relevant acceleration coefficient (obtained from a contour map), @ site factor related to the local soil conditions and ‘whether or not the structure is of a domestic or general nature. ‘The Earthquake Design Categories range from A to E and design requirements are specified for each category. The design requirements are also influenced by the regularity and ductility of the structure. The design requirements for domestic structures and the structural detailing requirements for cach Earthquake Design Category are laid out in detail in the Standard. Design requirements for non-structural components are also specified. ‘The design requirements for any given structure may specify an equivalent static or a dynamic analysis. These are discussed below: a) Equivalent static analysis procedure ‘This form of analysis enables a static load t0 be calculated for each floor level in the structure. At the heart of the analysis is the equation to calculate the earthquake induced base shear force V V=KCSIRIG, within the limits V = 0.01 G, and V= 12.5 a/R)G where: 7 = importance factor (either 1.0 or 1.25 depending on structure classification). CS= product of earthquake design coefficient (C) and site factor (S) which must not exceed 2.52, where @ is the acceleration coefficient Ry = the structural response factor which refiects the energy absorbing capabilities (ductility) of the selected structural system. Ry varies from 1.5 for unreinforced masonry to 8 for special moment resisting frames. the gravity load which consists of the dead load plus a proportion of the live load that can reasonably be expected at the time of the earthquake, ‘The earthquake induced base shear, V, is distributed up the structure in accordance with the height and vertical mass Eqn. 5.1Design philosphy and the Code approach » distribution, The approximation involved here is quite adequate for structures with reasonably uniform verical distributions of mass and stiffness as such structures respond primarily in their first mode of vibration which may be described by a simple equation. For a very stiff, regular structure the distribution of earthquake induced forces will take the form of an inverted triangle. As discussed in Section 3.2 the resultant of the earthquake forces induced at a particular level is assumed to act through the centre of gravity of the floor. The resultant of the resisting forces provided by the structural elements acts through the centre of stiffness. As these two centres are rarely coincident, because of either static or accidental effects, @ torsional loading is induced at each floor. The Standard provides guidance on the calculation of these torsional moments. The Standard also specifies checks on overall stability and provides guidance for the calculation of storey drifts under the ultimate limit state earthquake event. Dynamic analysis procedure AS discussed in Section 5.3.2 2 dynamic analysis may take the form of a real time analysis or a response spectrum analysis. AS11704 - 1993 recommends the use of the Fesponse spectrum method using either the spectrum provided in the Standard or a spectrum derived specifically for a particular site ‘The normalised response spectrum provided in the Standard has been derived from an ensemble of earthquake records for a number of different soil conditions using the procedure outlined in Section 5.3.2, The basic steps involved in using the nommatised response spectrum provided in the Standard are as follows: Step1 Develop a realistic finite element model of the structure, accurately reflecting the stiffness and ‘mass distribution, Step Scale the normalised response spectrum provided in the Standard by the factor (ah, where @ in the acceleration coefficient and is the importance factor, This procedure effectively produces a linear elastic response spectrum for the ultimate limit state event. Step I] Further scale the linear elastic response spectrum by If t0 account for inelastic behaviour Step]V Analyse the structure using the inelastic spectrum obtained in Step TUL Page 5:9Australian Earthquake Engineering Manual | Step V Compare the horizontal base shear force obtained with the value obtained from an equivalent static analysis. For irregular structures the Standard requires that the results, obtained from the dynamic analysis be scaled such that the base shear force equals that obtained from an equivalent static analysis. Similarly for regular structures the dynamii base shear force must not be less than 90% of the equivalent static analysis base shear force. Conceming Step V, the equivalent static analysis provides an earthquake base shear force (V) that has been developed from both theoretical and experimental data. The dynamic analysis on the other hand provides an accurate distribution of earthquake induced forces for a particular structure and must be scaled, as in Step V, to ensure that the overall base shear force is not compromised. Various other considerations must be taken into account. ‘As with the equivalent static analysis, earthquake directional effects must be accounted for using the 100% and 30% rule described above. Also, in a full three dimensional analysis the dynamic amplifications associated with horizontal effects are automatically accounted for. ‘The finite element model must. however, be adjusted to account for the effects of accidental eccentricities. For a two dimensional dynamic analysis, torsional effects are not included and the equivalent static torsional analysis must be conducted. Finally, the displacements calculated from a dynamic analysis must be multiplied by the deflection amplification factor (kd) to account for inelastic behaviour. Page 5-106 I Design of Buildings 6.1 Introduction Although much of Australia is considered to have very little risk of experiencing @ damaging earthquake, the damage to many older and some modem buildings in the 1989 Newcastle Earthquake begs the question: “Shouldn't the populated areas at least have earthquake design requirements for their historic and ‘new commercial buildings?” ‘This issue is being addressed in the new Australian Earthquake Code with an entirely new seismic map as mentioned in Chapter 5 - Australian Earthquake Codes. Consideration should be given as to how these lateral forces are to be resisted and absorbed by the structure. Connections between structural elements are critical (0 seismic performance; if forces can’t be transferred to ‘other elements partial collapse may eventuate. The basic aim of seismic structural design is to prevent death and injury within and around the structure during and immediately after the earthquake. Where possible structural damage should be minimal, so that the building is operational as soon as possible after the event. Past earthquakes have often shown worse damage for complex building shapes so wherever possible, simple plan and elevation shapes should be used. Figures 6.1 and 6.2 show simple and complex shapes and irregular framing systems. See Ref. 6.27 for a full description. ‘The basic seismic coefficient for a site is dependent on the site's previous seismic history and in Australia's case, based on about 200 years of recorded history, This is a very short time span, geologically speaking and of insufficient length to predict site accelerations with high certainty, In the Australian Earthquake Code (AS 2121, Ref. 6.1) the basic seismic coefficient is known as the Zone Factor, Z, Z has a maximum value of 0.36 in the worst seismic zone 2 while the lowest seismic zone is called “A”, where Z=0.09 for non-ductile buildings but zero for ‘ductile’ buildings, It is recommended that all buildings in Zone A be designed for some basic coefficient, even the low value of 0.09. ‘To this basic seismic coefficient other factors are multiplied to find the base shear coefficient, which when multiptied by the building's ‘seismic’ weight gives the design base shear force ‘The Australian Code includes 4 other factors, namely’ 4) the Occupancy Importance Factor, / b)_ the seismic response factor, C ©). the site-structure resonance factor, § 4) the structure's horizontal force factor, K Page 6-1Austialian Earthquake Engineering Manual PLANS SIMPLE COMPLEX Die) q 25 O b AN vy DQ) Alei[o ELEVATIONS ‘SIMPLE COMPLEX fh i oe = _} f=|[c>}}} |) iP} [el allie C1 Dh LAI DIBIE TIRE Figure 6.1 Simple and complex shapes in plan and elevation. From “Building Configuration and Seismic Design by Arnold and Reitherman. (Ref 627). Page 6-2Design of buildings “IRREGULAR STRUCTURES OR FRAMING SYSTEMS” (SEAOC) ‘A BUILDINGS WITH IRREGULAR CONFIGURATION CPLSS ‘Tabeped ian ‘Lupe pan snaped pian Crvetoem lan (ter compiex shapes & S Quart nllerm spas Sher out nonunforn mes Mute tower Spllievl Uns oh sty Unaaly low toate an [a sunonas with ABnUPT CHANGES WW LATERAL RESISTANCE tote tower ve Incr o elas pening in elshragmt ‘© BUILDINGS WITH ABRUPT CHANGES IN LATERAL STIFFNESS. ‘Shaer was some sore, en canes in memaotreingfameain othary__ntaruplon of vraag sloments__ Abrupt changes Inia ofwambery ruta ale D, UNUSUAL OR NOVEL STRUCTURAL FEATURES Staggered tases Ballon on tildes Cableupporid svete Figure 62 Graphic interpretation of “irregular structures or framing systems” from the commentary 10 the SEAOC Recommended Lateral Force Requirements and Commentary. From “Building Configuration and Seismic Design” by Arnold and Reitherman, (Ref 6.27). Page 6-3Australian Earthquake Engineering Manual Structurally speaking the two most important factors are C and K. The seismic response factor is inversely proportional to an estimate of the structure's natural period, 7. 1 Cae, where Cs 0.12 wT ‘This is because most typical earthquake response spectra show peak response at about 0.3 second period, then reducing response with increasing period. It should be noted that soft alluvial sites usually show peak response at higher periods; both Bucharest (197) and Mexico City (1985) earthquakes had peak responses in the period between 1.8 and 2 seconds, due to very deep alluvium under parts of those cities. Also at softer sites amplification of the base rock accelerations occur and this phenomenon is included jn the site factor, S. Preliminary estimates of a regular structure's natural period can be found by using the formulae in Figure 6.3, €a Eqn. 61 T= 0-09 by vo WALLS Figure 63 Natural period estimates ‘The formula for C is an approximate compromise as it assumes that a damped multi-degree of freedom structure will respond in a similar way to a similarly damped single degree of freedom structure. More data from actual Australian earthquakes could change the formulation of C markedly - this may well be reflected in the new code. ‘The horizontal force factor (K) is a measure of how ductile the type of structure being designed is likely 1o be in an earthquake. ‘All buildings could be designed to remain elastic during the ‘worst earthquake but this would mean much extra expense for larger structural members, so as 10 obiain the level of strength UNIFORM FRAMES Page 6-4Design of buildings required. By detailing ductility into specific zones of the structure, the building is expected to yield (become plastic) at these zones and thus dissipate the inertial energy. These so- called plastic hinge zones must be able to dissipate cnergy for several moment reversing cycles as the building moves forward and back, without losing flexural strength or failing in a brite manner, (Figure 6.4). For these ‘ductile’ structures a lower design base shear force can be specified (lower K) as the building relies on its inherent ductility rather than strength to resist the earthquake. Moment Moment (— yielding occurs LL. eneepy absorbed in fit eycle Elastic response Ductile (non-elastic) response Figure 64 ‘Two empirical concepts have been derived for in-elastic ‘oscillators; the equal displacement and equal energy concept. ‘The structural displacement ductility factor (1) is usually defined as the maximum lateral displacement (at roof level) divided by the roof displacement when yielding first occurs. If the equal maximum displacement is assumed for both elastic and inelastic structures: rm ‘Thus for a ductility factor of 4, the equal-displacement structure needs @ lateral yield force only 25% of the equivalent elastic force and for the equal-energy concept the reduction is to 38% of F, . The equal-energy concep: is most realistic for real structures with a period less than about 0.6 second while the ‘equal-displacement concept is fairly accurate for higher periods. (See Chapier 4, Design curves and inelastic response spectra) Most modem seismic design codes of practice assume a maximum sustainable displacement ductility factor of 6 for the most reliable fully ductile structural type, (for example, ductile Eqn. 62 Eqn. 63 Page 6-5Australian Earthquake Engineering Manual 6.14 moment resis if Space frames). In AS 2121 these fully ductile structures have been given a value of 0.67, while fully elastic responding structures had a K of 4 initially, but this was reduced to 3.2 in a later amendment K “ Horizontal Force Factor Displacement Ductility Factor Fully ductile 067 6 0.80 5 1.00 4 Limited ductility 133 3 1.67 24 2.00 2 Elastic 42) 1 Table 6.2 ‘Table 6.1 shows the range of displacement ductility factors and their corresponding current AS 2121 horizontal force factors. Fully ductile structures are really only applicable in areas where major earthquakes are likely to occur. No such areas seem to exist in Australia, Also to obtain full ductile capability stringent detailing requirements are needed for the plastic hinges to ensure they remain energy dissipating after many reversed cycles and to preclude brittle shear type failures at or near plastic hinge zones. To obtain this reliability a ‘capacity’ design approach is preferred in which everything away from the plastic hinges is designed to over strength forces and moments, in the hope that these areas never fail before the series of hinges yield. ‘Such an approach seems unrealistic for the Australian level of seismic risk. It is therefore recommended that Australian buildings be designed using a ‘limised ductility’ approach where displacement ductility factors between 2 and 4 are specified. Limited ductile buildings can be designed for relaxed detailing requirements in their plastic hinge zones and the recommendations in this manual are largely based on the limited ductile requirements specified in the New Zealand materials’ codes of practice. Distribution of lateral forces ‘The base shear force V is determined in both of the structure's main plan directions using the base shear coefficients multiplied by the seismic weight, W, which is an estimate of the gravity load likely to be present at the time of the earthquake, Usually this is the deat load of the building plus some proportion of the design Jive load, For normal office type occupancies one thirdDesign of ouilgings of the floor’s live load is used (NZ Code NZS 4203 Rel. 6.2), while for floors used for storage (libraries, cool stores, etc) two thirds of the design live load is commonly used in the seismic ‘weight calculation. For roofs with access only for maintenance, only the dead load usually contributes to W, Note that AS 2121 (1979) (6.1) specifies only the dead load in all situations except for the addition of 25% of the live load when the area is used for storage. It is usually assumed that earthquake forces do not act in both orthogonal directions at the same time and therefore mic analyses are performed for the structural system in both major directions independently. Complex structural forms may require earthquake force analyses in different horizontal directions. Hy F, > T Hy [= = a WH OE ST Base shear V = (ZIKCS)W v= SR +H, w= Dw R=p-_ WA) H.= 007TV < 025V = 0 when T < 07sec Figure 65 Once the base shear force has been derived this force must be distributed up the structure, as a lateral force at each floor level, where the lumped masses ane assumed to be. A triangular force distribution, increasing up the structure is usually specified for uniform structures, as this approximates the forces required 10 give a good representation of the structure's first mode shape. This is the lowest mode in which the structure will tend to vibrate during the ‘quake. In this way the complex dynamic problem is approximated by a pseudo-static one with the constanlly varying inertial forces being replaced with a set of static horizontal forces. To take account of second and higher Ist mode 2nd 3rd Shapes Page 6-7Australian Earthquake Engineering Manual 6.1.2 mode ‘whip lash’ effects, in slender structures, a fraction of the base shear, H, is often added to the top of the structure while the remaining base shear is distributed as a top heavy triangle up the building, (Figure 6.5) eg= 018 - | Dwisting Floor level x e = distance between centre of rigidity (CR) and centre of mass (CM), Structural Eccentricity ‘Torsional design moment M, = F(é # 0.18) From AS2121:1979; Accidental Boventricity eg 1.Te, ~ if, + 0.10 Figure 6.6 Torsional lever arm ‘The force at any level, F, is distributed to the horizontal force resisting elements present at that floor in proportion to their rigidities (stif ness), as long as @ suitably stiff floor diaphragm exists to distribute the forces in this way. Horizontal torsion ‘The horizontal seismic forces at any level need to be increased to allow for horizontal torsion; the twisting of the structure horizontally about its centre of rigidity. (See Chapier 5 - Australian earthquake codes.) ‘The torsional moment at any floor level should be equal 10 the horizontal force at that level, H; times a lever arm equal 10 the distance from the floor’s centre of mass to its centre of rigidity and an accidental torsional lever arm of 0.1b, where 6 is the width of the building measured perpendicular to the seismic force, (Figure 6.6). This is a simplified approach and iregular structures must be carefully torsionally analysed. As discussed in Chapter 5 - Australian Earthquake Codes, these provisions have been found wanting under certain circumstances and will almost certainly be revised in the new Code. Page 6-8Design of builings 6.2 Building structure types 6.24 Frames ‘Two way ductile frames are commonly used to resist seismic actions. For these structures beams frame in both major directions, both intemally and extemally meeting at common columns, (Figure 6.7). feces cece este aeeeeeeeeetadt rea fod 3 SSS or Pr ID Ot ht ot ft Figure 6.7 Plan of a two-way frame ‘These ductile moment resisting space frames have the lowest K value (0.67 - corresponding to an assured highest structural ductility factor (= 6) in the current Australian Earthquake code). Diaphragm action is provided by precast or in-situ, reinforced or prestressed concrete slabs. Precast units don't usually provide satisfactory diaphragm action in themselves and require a structural topping of reinforced concrete, tying all the components together. A minimum topping thickness, containing light mesh reinforcing, is usually 5mm but 63mm is preferable. A variation of the 2-way frame is the tube framed structure, where the seismic frames are positioned around the building's periphery, (Figure 6.8). Intemal columns carry only gravity loads and beams connected to them are usually designed with simple pin supponss so as not to attract seismic moments and shears. ONE -WAY SLABS INTERNAL ‘BEAM - COLUMN ‘CONNECTION Page 6-9Australian Earthquake Engineering Manual SEISMIC FRAMES Ne GRAVITY BEAMS GRAVITY COLUMNS (PROPS) Figure 68 Peripheral (tube) frame In seismic frames the beams should normally be deeper than their width and preferably they should be narrower than the columns they intersect with, Wide-shallow beams have low moment and energy dissipation capability. This requirement produces higher interstorey heights than flat slab or beam structures but often air-conditioning and other ducting can be ‘accommodated in the beam depth. B $e oe] Pt a aes 4245p 4. PLAN FRAME A-A FRAME 8-8 (SEISMIC ONLY) (SEISMIC AND GRAVITY) Figure 69 Seismic bear bending momenis Often beams in one direction ate deeper than those in the other direction due to one-way slabs contributing gravity loads to Page 6-10Design of buildings 6.2.2 beams in one-direction only, (see Figure 6.9). This allows ducting to pass under the shallow beams without requiring lowering of the ceiling. Peripheral frames require more columns to counteract the lack of internal seismic frames, but have the major advantage of no deep beams intemally conflicting with service ducts and other services. For pure frame action all service cores, heavy partitions and stair flights need to be seismically separated so that the extra stiffening is not provided by these elements during the earthquake, If unseparated these elements destroy the frame action and usually fail themselves in a brittle fashion, due to their inherent lack of ductility (see Section 6.6}, ‘The major disadvantage with pure seismic frames is their usual high flexibility causing large lateral seismic deflections, ‘Thus curtain wall glazing systems and precast concrete wall panels require carcful detailing of their fixings and seismic separation ‘gaps to allow these large drifts to occur without binding and pounding of these non-structural elements, (Figure 6.10). Moment resisting space frames with reduced ductility requirements are specified in the Australian code with K=1, which is equivalent to H=4. One-way frame structures are possible when shear walls are used in one orthogonal direction with frames only in the other direction, (Figure 6.11). Therefore it is often possible to have seismic beams in one direction only, allowing freer positioning of services. In situations where extra headroom is needed for services, seismic beams can be haunched near their ends allowing extra space in the mid-span region. Beam plastic hinge positions are often moved away from the critical joint region with haunched beams. Secondary beams are usually required around slab penetrations for stair and lift wells to help distribute the diaphragm forces 10 the seismic beams and columns. Walls Most shear wall structures have single or multiple vertical cantilever walls assumed fixed at ground or foundation level where the plastic hinge is required to form in the moderate to severe earthquake. The disposition of the shear walls in plan is very important when it comes to the building's seismic response. Plans (a) and (b) of Figure 6.12 although simple and symmetrical have little torsional stiffness and would require the exterior columns 10 resist the torsional actions. Plan (©) with a similar amount of wall as (b) has much higher torsional stiffness due to the large lever ann from the floor plan's centre of mass 10 cach wall’s centre of stiffness. Page 6-11Austraian Earthquake Engineering Manual COLUMN SEPARATED. infu GLAZING WALL PANEL \ N FRAMES SEPARATED \ PRECAST AUD res Figure 6.10 Separated wall panels in ductile frame SEISMIC FRAME / > FRAME ACTION | K= 0:57 oF SEISMIC. WALL. SEISMIC WALL — oe i il th i GRAVITY BEAMS: i if | ‘SHEAR WALL ACTION ! =H ‘Sate an ce kang ‘SEISMIC FRAME PLAN Figure 6.11 Frames in one direction, shear walls in the other Page 612Design of buildings i © © $ @ e 2 2 ; | i a 4 | ‘ | : © & © ® =O fab tb) (ch PLAN Figure 6.12. Different positions of seismic walls (K=1.0 or 1.33) cR SS N N N N N N Ny N N N Ni PLANS (M= CENTRE OF MASS CR = CENTRE OF RIGIDITY Figure 6.13 Torsionally unbalanced plans ‘Tall structures with asymmetrical shear wall positions also behave badly in earthquakes due to torsion, caused by the large distance between centre of mass and centre of rigidity, (Figure 6.13). Placing shear walls near the building's perimeter and in @ symmetrical pattem should be the main planning aim. Single and 2 storey shops with substantial walls on their back and sides but only light framing and glazing on the front face have behaved poorly in earthquakes due to the inherent torsional imbalance, (Figure 6.14). Page 6-13Australian Earthquake Engineering Manual waus — ce Hey —- ) TORSION sd LARGE TORSIONAL, I DEFLECTIONS, Figure 6.14 Asymmetrical wall pattern 6.2.3 Coupled shear walls ‘These are slender cantilever walls linked by ductile coupling | beams at each floor level. In a moderate to large earthquake the coupling beams, as well as the base of each cantilever wall (pier) will form plastic hinges to dissipate seismic energy. Coupled shear walls ate often used for lift wells where the ‘coupling beams are the lintels over the lift door openings, (Figure 6.15). FLOORSLAB Z 3 | COUPLING BEAMS =F peels = PIERS: COUPLING BEAMS Figure 6.15 Coupled shear wall structure Page 6-14Design of buildings 6.2.4 6.2.5 Coupled shear walls are very efficient seismic resisting elements as they have large energy dissipating potential and are stiffer than ductile moment-resisting frames, thus reducing the lateral seismic drifts Combined walls and frames (K = 0.8) Combinations of ductile frames and shear walls are possible (hybrid or dual systems) in the same direction but careful consideration is required about the interaction between the wo structural types, especially near the top of the structure where dissimilar deflected shapes occur, (Figure 6.16). Paulay and Goodsir (6.3) have developed a seismic design procedure for hybrid frame wall structures and have shown them to have good energy dissipation potential and reduced lateral drifts when compared with pure frames. The chances of column hinges forming is virually eliminated with hybrid structures, SZ Sw WD 7 Lateral Frame element Wall element Coupled Frame-wall building load (Shear model) (Bending model) Figure 6.16 Deformation patterns of laterally loaded frame, walls and coupled wall-elements. From Paulay and Goodsir (6.3) Braced frames Braced frames with diagonel braces from beam-column to beam-column can be made seismically resistant with ductile capability in tension and/or compression in the braces. However, such structures are less efficient in dissipating energy than pure frames. Thus braced frames are usually designed for higher lateral forces, (higher K factor). V-braced frames (Figure 6.17) usually give the lightest design but where the slendemess ratio of the braces is greater than about 40, the compressive strength of the brace may fall to less than 30% of the initial strength, due to buckling and subsequentwstralian Earthquake Engineering Manual tensile yielding. Studies of light cross braced frames have shown reduced lateral deflections and inelastic behaviour in both beams and columns. Also substantially increased column axial loads were noted when compared with unbraced frames, (Ref. 6.4), K-BRACED Figure 6.17 Concenirically braced frames (K=1.0) Eccentrically braced frames, as shown in Figure 6.18, are designed to exploit the inelastic behaviour of steel yielding in shear. INELASTIC ne iat INELASTIC — DEFORMATIO! ir ZONE AT ZONE EACH FLOOR ECCENTRIC Z Figure 6.18 Becentrically braced frames ‘This form of structure has reasonably high ductility potential falling between ductile moment-resisting and slender cantilever shear walls, (equivalent to K factor of about 1.0 in AS 2121 with a structural ductility factor of 4). The K-brace system has a more predictable failure mechanism and has advantages over Z- races due to the gravity shears being less in the link ECCENTRIC Y= BRACED kKDesign of buildings 6.2.6 6.2.7 6.2.8 6.3 6.3.1 Bearing wall systems Bearing walls carry most of the building's gravity loads together with shear walls or braced frames and the bearing walls! resistance to seismic forces requires consideration. A significant problem exists with bearing walls, in that they try to act as shear walls in their plane during the earthquake by attracting significant shear forces due to their in plane stiffness. Because they have little if any horizontal shear reinforcement they usually fail in a brittle fashion, often suddenly at low seismic force levels and leave nothing to support the gravity loads. ‘Therefore bearing walls not designed 10 resist seismic forces are not a suitable structural type in areas of medium to high seismicity. Tilt up structures Such structures are prevalent in Australia in low rise ‘commercial buildings (3 or less storeys). The detailing of the tlt up panel connections to their foundations and other panels are often inadequate for any seismic loading. Brittle, non ductile details are often used and therefore care should be exercised in the seismic design and detailing, Vertical accelerations Jn most recorded earthquakes the peak vertical accelerations have approached the peak horizontal accelerations and it is usually accepted that the vertical accelerations will be at least ‘equal to 60% of the corresponding horizontal accelerations. It is thus necessary to consider the effects of these vertical accelerations on Jong span structures and especially horizontal cantilevers. The N.Z. Code (Ref. 6.2) specifies an upward or downward seismic coefficient of 0.9 for horizontal cantilevers with the gravity loads being ignored when considering the upward design forces. Energy dissipation techniques Beam and column plastic hinges Ductile plastic hinges dissipate energy by the cyclic yielding in tension and compression of prin: a) reinforcing steel or structural steel in reinforced concrete and steel. structures respectively. In reinforced concrete sections the longitudinal reinforcing must be restrained from buckling by having closely spaced transverse closed ties which confine the core concrete. as well as carry much of the high shear forces. Confining the core concrete also inereases the maximum compressive strain possible in the concrete which also increases the ductility capability. After several revers: cycles during a severe Page 6-17Australian Earthquake Engineering Manual earthquake the cover concrete will usually have spalled. The amount of compression longitudinal reinforcement in a section is proportional to the section's ductility potential and the NZ Concrete Code (NZS 3101) (Ref. 6.5) specifies a minimum ratio of 0.5 for compression steel to tension stee! in any fully ductile plastic hinge region. In experimental testing, beam section rotational duciilities 6, /¢, (not to be confused with the overall structural ductility factor (jt, Figure 6.19), greater than 40 have ‘been possible for several full reversals of moment. dy by > ey Pre. / + PLASTIC 7 HINGES HINGE ABOUT To FORM FIRST _YIELO CONDITION ULTIMATE CONDITION (BOTH EXAGGERATED! STRUCTURAL OUCTILITY FACTOR eke Ay ROTATIONAL SECTION OUCTILITY = Oy 9; y Figure 6.19 Definitions of dueciility factors 6.3.2. Beam hinge versus column hinge mechanisms Beam hinge mechanisms (Figure 6.20) are more suitable for dissipating seismic energy than column hinge mechanisms (Figure 6.21) because: a) A greater number of plastic hinges form before @ mechanism forms - thus smaller section rotational Guciilities are required for each hinge, b) Beam hinges are more ductile than column hinges due to ow axial loads, Page 6-18Design of buildings ©) Beam hinges do not have 10 carry large axial loads, as columns do, during and after the earthquake. i Mp (-ve) (tg Mplevel BEAM HINGES yo COLUMN HINGES. Figure 620 Beam hinge mechanism iq ‘«— Column hinges aw Figure 6.21 Colwnn hinge mechanism (Soft storey) In past earthquakes well detailed (confined) columns have deflected laterally by up 10 & metre in a single storey in ‘soft storey’ buildings without collapse, while others have coltapsed due to excessive ductility demands and column shear failures caused by the plastic moments at the column ends, (Figure 6.22 and 6.23), It is more difficult to detail for the required ductility in a column than in the corresponding beams. Page 6-19Australian Earthquake Engineering Manual a Figure 622 Poor detailing of tied reinforced concrete column, Olive View Hospital. (Ref. 6.28) Page 6-20vesign of vanaings Figure 623 Good detailing, poor structural form: spiral reinforced concrete colunn, Olive View Hospital Page 6-21Australian Earthquake Engineering Manual Page 6-22 ‘This is due to the axial load effects and the much higher section duciilities required in column hinge mechanism, compared with the beams in a beam hinge mechanism, for the same structural ductility factor. For example @ 10 storey reinforced concrete frame forming a column hinge mechanism would require section ductilities (9,,/6, for columns) approaching 120, whereas the same building dewiled as a beam-hinging mechanism, 6/6, for beams of about 20 is required for 1 =4 in both cases. (The column base hinge section ductilities approach 12 only in the beam hinging case). See Ref. 6.6 for more details. For tall frame structures the ‘weak beam - strong column’ approach should be followed so that a beam hinge mechanism will form with high probability. Note that some column hinges may form for short periods of time due to dynamic effects. Colump-hinge or ‘soft storey’ structures should only be designed in seismic zones for buildings up to 2 oF 3 storeys, with special attention being paid to the column design and detailing. For Jong span beams and/or when the gravity loads are high the positive bending moment hinge will form some distance out from the column face, (Figure 6.24). This will usually mean greater rotational section ductilities are required but the positive hinge has a longer length in which to form. Note that these hinges are unidirectional, that is they only yield in one direction, negative hinges from at column faces and positive hinges form within the span. The actual maximum rotational section ductilities are usually greater for these unidirectional hinges, see Figure 6.25 and see Ref. 6.7 for more details. The limits on shape of flexural members specified in AS 2121 (Appendix 5B), (Ref. 6.1), namely width to depth ratio not less than 0.3 with the width, 6 not less than 250 mm should apply for all beams and columns which are likely to form plastic hinges, However the beam width extreme of 1.5D +b, (where D is the total beam depth and b, is the column width) is considered too large for members forming plastic hinges and @ limiting beam ‘width of half the column depth + the column width i.e. [F-] but less than 2d, is recommended, (Figure 6.26). The aim is to keep the longitudinal beam reinforcing reasonably close to the column core and a minimum of 75% of the effective bean reinforcing should pass through or be anchored within the column core. Eqn. 64Design of buiicings NEGATIVE HINGE atbeiii dy YY ER Eee — POSITIVE HINGE COLUMN ea Ha level FACE craviey anf + SEISMIC SEISMIC ‘eve HINGE POSITION WHEN STRUCTURE MOVES —— Figure 6.24 Unidirectional beam hinges when gravity loads exist BENDING MOMENT M M 2 = %y 6 ff [ors e a BI-DIRECTIONAL PLASTIC HINGE UNI-OIRECTIONAL PLASTIC _HINGE Figure 6.25 Moment rotation curves for bivdirectional and uni= directional plastic hinges Page 6-23Australian Earthquake Engineering Manual 6.3.3 Figure 626 Beam dimension restrictions Wall hinges Slender reinforced concrete seismic walls can be detailed in such & way that they exhibit excellent energy dissipating behaviour. Cantilever seismic walls are usually designed to form a single plastic hinge at the base; this hinge needs to form over a height at least equal to the walls’ width. This length could amount to 2 or 3 storeys for high walls. Large vertical strains associated with high rotational ductilities will be reached at each end of the wall. Horizontal ties will be required not only to carry the high horizontal shears but 10 confine the compression concrete and restrain the principal reinforcement from buckling if compressive strains greater than 0,003 are required in reaching the desired structural ductility. For low axially loaded walls no contribution of the concrete in carrying shear force is usually assumed for fully ductile walls. However, for limited ductile wails some shear may usually be assumed carried by the concrete. Higher maximum curvatures (larger section ductilities) are possible when the majority of the principal vertical reinforcement is positioned near the end of the wall, rather than spreading it uniformly across the wall. Minimum vertical reinforcing (A,, /bd = 0,25%) is recommended for the inner 80% of the wall width with the remainder placed in the outer 10% of the width zones. With reinforced concrete block masonry this isn't usually possible and a more uniform spread of venical reinforcing is more practical. The NZ Code (Ref. 6.5) recommends principal reinforcement ratios between 0.7/f, and Page 6-24| | | i i | oF ounarigs 6.3.4 6.3.5 L6/f, for limited ductile end regions of walls, with the bar diameter not exceeding one eighth of the wall thickness. (/, is the yield stress of the longitudinal reinforcement). To preciude a brite type shear failure at the wall’s base occurring before a ductile flexural hinge forms, some over estimate of the wall's shear must be used to design the horizontal reinforcing. See Section 6.4 for details. Instability of the compression end of the wall should be considered at the high strains reached and some form of retum will often be necessary. To prevent buckling of compression reinforcement near the ends of limited ductile walls the spacing of the horizontal ties should not exceed 10 times the diameter of the longitudinal bar being restrained, one half the wall thickness nor 150mm. The potential yielding region of the wall should be taken from the critical section as the greater of the wall width or 1/6 of the wall’s total height. (21,, maximum). Due to the impracticality of confining walls with only a single layer of vertical reinforcing, these should not be used as limited ductile walls. If a ductile mechanism is required to form at wall’s base the total shear stress on any wall should not exceed 0.5Yf MPa Squat shear walls Walls whose length, 1, is greater than or equal to its height, h,, are termed squat and these usually fail in shear or by sliding shear, The flexural steel demand is usually small and for squat walls it is beter to spread the vertical steel evenly over the walls length (p = 0.25%, minimum reinforcing percentage). No significant arch action occurs, as demonstrated in deep beams, due to the seismic forces being applied evenly along the wall at each floor level, In squat walls shear dominates seismic behaviour, and larger values of K should be used in their design. K = 1.67 being suggested, (4 = 2.4), AS 2121 should be used to determine the ratio of vertical shear reinforcement, (p, = 0.0025 +0.5 (2.5 ~ Fy Li.)(P, 0.0025), where p, is the ratio of vertical shear reinforcement area to gross concrete area of horizontal section, Px is the ratio of vertical shear reinforcement are to gross concrete area of vertical section, A, is the height of wall and J, is the horizontal wall length). Coupled shear walls Cantilever shear walts pierced with a regular pattem of holes for lif door openings or stair well door openings can be designed as coupled shear walls, where the short beams between the walls transfer shears and axial loads during an earthquake. Eqn. 65 Page 6-25Australian Earthquake Engineering Manual The axial force in the walls result from the accumulation of the shear forces in the coupling beams, (Figure 6,27). ace BEAMS Oo 7 oO j te | | te 4 | L CentRotoaL j i |— CENTROM ' oO FT he oO he loOe logs he | a - ote Vd va NGay Ve Nba Me ea FREE BOOY WALL 1 { _ 1 t beet 1 EXTERNAL MOMENT My = M,+Mz+ TL Figure 6.27 Coupled shear wall actions STIRRUPS, BEAM PRINCIPAL J REINFORCEMENT —\ CONVENTIONAL BASKETTING STEEL TO PREVENT COVER LOSS cs Sy SECTION DIAGONALLY SPIRALS AROUND REINFORCED PRINCIPAL REINFORCEMENT Figure 628 Coupled beam reinforcing Page 6-266.3.6 ‘The intended seismic behaviour is that the coupling beams should form plastic hinges up the structure before plastic hinge zones form at the base of each walt, thus forming the ‘collapse’ mechanism, Because the shear strength of the coupling beams will often dictate their flexural capacity as well as their rotational ductility potential, special detailing may be required if large section ductilities need to be sustained. This could include diagonal reinforcing in the coupling beams with transverse rectangular ties or spirals around each diagonal strat to prevent buckling when the strut is yielding in compression, (Figure 6.28). ‘The concrete in conventionally reinforced coupling beams becomes more and more ineffective in transferring shear to the walls as the yield cycles progress and diagonal cracking increases. Thus stirrups need to be provided to carry all the shear force produced when the coupling beams reach their maximum flexural strength. Sliding shear ‘failure’ at the wall face can often occur with squat beams (length/depth < 1.5). ‘The N.Z. Code (NZS 3101) recommends that diagonal reinforcing be used in coupling beams when the seismically induced shear stress exceeds: OLA where /, is the clear span of coupling beams and h is the overall depth of beam, The total shear stress in any coupling beam should never exceed: 0.9% MPa. Limited ductility requirements can be satisfied with conventionally reinforced coupling beams. In moderate ‘earthquakes, requiring low values of section ductility, the only damage (yielding and cover loss) should occur in the coupling beams and these could then be easily replaced or repaired after the event. Seismic isolation The seismic induced forces and accelerations reaching the superstructure of @ building or bridge can be considerably reduced by isolating the structure from its foundations. The isolation can be effected by using rollers, rubber bearings, or sliding plates, The isolator's high lateral flexibility, however, increase the displacements considerably between the structure and its foundations. Therefore some sort of energy dissipator or damper is usually used with the isolators to reduce the lateral seismic displacements and dissipate seismic induced energy in the form of heat. These devices include a lead plug within elastomeric (sandwich of rubber and thin steel plates) bearings, Design of buileings Eqn. 6.6 Eqn. 6.7 Page 6-27Australian Earthquake Engineering Manual lead extrusion devices, yielding steel devices, friction devices, etc. loading arms I anchors ‘Torsional beam device Lead extrusion device XN Flexural plate device Flexural beam device supersiructae anchor BBD WO elastomeric bosning Flexural beam device Lead-rubber device Figure 6.29 Steel and lead mechanical enersy dissipating devices Page 6-28Design of buiioings 6.3.7 6.3.8 Figure 6.29 shows some of the deviees developed for use under buildings and bridges. The main advantage of seismic isolation, often known as base isolation, is that the response of structures with low natural periods (< 0.5 sec) is substantially reduced by effectively increasing the isolated structure's period up to about 2 seconds and thus moving the response out of the critical Period range of most earthquakes. Sites situated on deep alluvium, where maximum response periods are 1.5 sec or larger are unsuitable for base isolation, as the structure's Tesponse will probably increase, There is little point in isolating flexible tall structures as they already have high natural periods. ‘The reduced seismic response of isolated structures also means reduced interstorey drifts which allow much simpler detailing of Ron-siructural elements, including glazing and cladding systems, stairways, partitions and suspended ceilings. The savings in non-structural detailing should at least equal the extra foundation beams and isolators. A structure isolated on a site where major earthquakes are almost never expected would usually be designed to remain elastic above the isolators during the ‘design earthquake’. For more detail on seismic isolation see references 6.8, 6.9, and 6.10. Braced dampers It is possible to design frames with energy dissipating bracing which incorporates devices in the bracing to dissipate energy during the earthquake, without permanent damage to the bracing. These devices could dissipate energy by yielding of mild steel sections or the extrusion of lead, for example. They would not normally require replacement after the event as laboratory testing has shown that these devices and those used to isolate structures, can withstand many yield cycles equivalent to several major earthquakes before replacement is necessary. Active control ‘The use of controlled bracing tendons in which a dynamic set of forces continuously resist the structural response during an earthquake has been investigated. Some forms of active control could have a future in reducing the response of structures during frequent small intensity tremors but are unlikely 10 be economic or practical for the rare medium to major earthquakes. The requirement of a continual power supply to activate the control system during the earthquake is one of the more practical problems. (Ref. 6.11) Page 6-29Australian Earthquake Engineering Manual 6.4 6.4.1 D n (ie uattstit tt} bp {p el WALLS BEAMS M Sty Shy lp=h Detailing for ‘limited’ ductility Reinforced concrete Limited ductility structures should have designated end regions in beams of beam-sway mechanisms, in columns of low rise column-sway mechanisms and at the bases of cantilever walls where plastic hinges are likely to form during the earthquake. ‘The minimum length of these end regions is shown in Figure 6.30. Figure 6.30 Minimum end region lengths (Lp) “The required shear strength should have a suitable margin over the required flexural strengths, so as to preclude brittle shear failures such that: Ji V_ is the shear force due to the Strength Method load combination including earthquake (U=D+13L,+E, for example, where D = Dead load, Lp = reduced live load and E = earthquake forces) and V., is the shear force due to the earthquake forces, E alone, (9, the strength reduction factor = 0.85 for shear). The required flexural strength outside the designed and detailed end regions should be such that: a ae oan om, me +[25-1] Ma K d where the subscripts mean the same as previously and 6 = 0.90 for pure flexure situations. This requirement usually means that no extra reinforcement is needed at critical sections (compared with non-ductile designs) but the point of terminating tension flexural reinforcement will be increased slightly, Eqn. 68 Eqn. 69 Page 6-30Design of buileings 6.4.2 Confinement of end regions ‘The required confining reinforcement around the longitudinal bars should be calculated from the following expressions: MEtO3NA S Eqn. 6.10 0.66 £Azh (See Figure 6.31). M, is the mid-depth moment associated with D+ 1.3L, +E for example. When y> 1.0 , single layer reinforcement is not allowed and the area A," must be confined as follows: Ag, (transverse reinforcement) = a(ow Sy me Eqn. 6.11 in where OSR¢ 1 but <1 1+ pm 8, = spacing of confinement hoops or cross-ties of area Az, Of yield stress fy, ratio of all Jongitudinal reinforcement within Ag =Ay /A,", and m = f,/0.85f/ All longitudinal reinforcement within A,* should be tied even when 7< 1 and also in walls where there are 2 layers of steel and in all positions where the longitudinal steel ratio exceeds alt, Figure 6.31 Confinement of wall end region ‘The Australian Code (Ref. 6.1) Clause SB4, specifies a minimum ratio of compression to tension steel of only 0.25 at column faces for a timited ductility reinforced conerete moment resisting space frame (K=1, L=4), whereas most concrete codes specify at least a 0.5 ratio, as does the new Australian concrete code (AS 3600:1988) (Ref. 6.12). The lower ratio is a Page 6-31Australian Earthquake Engineering Manual realisation that if larger seismic forces are used in the design (K=1 compared with K =0.67 for fully ductile), then lower potential section ductilities can be detailed at beam plastic hinges, and a lower ductility factor of 4 (rather than 6) is assumed. For example, a confined beam with an effective depth of 800mm and mild steel reinforcing with A,’/A,=0.25 shows a theoretical decrease in section ductility to 65% of the ductility for the same beam with twice the amount of compression steel; A,‘/4,=0.50. Also the Australian Code contains no maximum spacing for transverse confining steel to restrict compression steel buckling. The NZ Code (Ref. 6.5) specifies a maximum spacing of 6 main bar diameters or 150mm (whichever is Teast) where full plastic hinge potential is required. ‘The limited ductility (K = 1) spacing of stirrup - ties is restricted to d/4 for the seven hoops closest 10 the column face in AS 2121 and AS 3600, With a minimum requirement of two 12mm bars at the top or bottom of beams it is unlikely that section ductilities necessary for w=4 overall will be sustainable due to the compression reinforcement buckling Reduction in tangent modulus of elasticity (softening) of steel at low levels of stress caused by the Bauschinger effect (Figure 6.32) can cause compression bars to buckle at low stress levels after a few reverse yielding cycles. Confining steel maximum spacing requirements for limited ductility end regions should be amended to the smallest of 10 times the longitudinal bar diameter or 150mm. The minimum Yongitudinal reinforcement in beams should be two 16mm diameter bars both top and bottom throughout the beam length. At least one-quarter of larger top flexural reinforcement at beam ends should continue throughout the beam's length, as specified in AS 3600. STRESS fs FIRST YIELOING a Es STRAIN BAUSCHINGER EFFECT Figure 6.32 Bauschinger effect in steel In the end regions of walls the ratio of longitudinal reinforcement over any part of the cross-section should not be less than 0.7/ and not greater than 16/f,, Bar diameters used in walls should not exceed one-eighth of the wall thickness. ‘The shear stress able to be carried by the concrete in end regions should be found from: Page 6-32Ve = (0.035 + 5p, }ffe where Peat and OOF sv, s01F bd but also v,20.4.[N,/A, where N, = seismic load combination axial load. For walls or columns, v, may be multiplied by a coefficient 3N, 1A, it 4 for axial compression and 1+ se 4 “| - AL ] for axial tension, where N, is negative. Outside end regions the value of v. may be doubled. (N, is the load factored axial load). Shear reinforcing should be spaced at a maximum of d/4 throughout beam and column end regions (100mm maximum for columns) and transverse shear steel can be used as part or all of the confining steel requirements, The maximum spacing limit may be doubled to d/2 or 600mm, whichever is least, outside end regions, (Figure 6.33). MAX PY th Design oF butioings Eqn. 6.12 Eqn. 6.13 ob somm SSLOR 10 MAIN BAR DIAM OF 150mm 60 MIN, SECTION 4.3 fc 0.008 Ag Act < 0-06 Ag ale 200m 5' te or 400mm i & ‘BEND AROUND MAIN Bars BEAM SECTION NOTE cuoseo Sas For eeans, STitkues eNO REGION Grow. Ke Figure 6.33 Minimum transverse reinforcing spacings PART. FRAME ELEVATION C Page 6-33Australian Earthquake Engineering Manual 6.4.3. Beam - column joints Page 6-34 At interior beam - column joints where beams may enter the column on all four sides, large horizontal and vertical shear forces exist when beam hinging occurs near the column face, (Figure 6.34) Horizontal joint shear force, Vn= Cy + T2— Vea =n +Aadfy— Veot Vertical joint shear force, Vj, = Vix. Dh, No requirement for horizontal joint ties in interior joints is specified in AS 3600: 1988 because it is assumed that the beams in the other direction will confine the joint concrete. However, transverse beam hinging may have left full depth cracks at the column faces, thus eliminating any confining effect. Careful consideration of how V,, and Vjy are carried across the joint is required otherwise brittle joint shear failures will occur, rather than the desired flexural beam hinges. Where several yield moment reversals are expected during the earthquake, the joint conerete's shear carrying capacity will diminish considerably and a reduction in the amount of joint shear force carried by the concrete should be used. NEGATIVE POSITIVE, HINGE Figure 6.34 Internal beam -column joint forces ‘The requirements of C4B4.1.2.2 ‘Confinement of Joints’ (AS 2121 and AS 3600) for the transverse reinforcement in exterior columns (same confining as for column ends with maximum spacing of 100mm) should be treated as minimal shear/confining stee! and consideration of the actual joint shear forces with beam hinging occurring at the column face should be considered, (Figure 6.35).Design of buildings Horizontal joint shear force, Vy =A,f,— Vertical joint shear force, Vj = Vjq. D/he. Both interior and exterior beam column joints should have transverse hoops designed using the following equations: Ven=Vin~ Vex» Where Vey is the shear force carried by the concrete and is equal to: 2 Xe (0, 4.) for limited ductile situations BY Note that Vi, =0 when columns are in tension, y, Ma Ss design axial load on column (compression due to gravity and seismic loads, Ag = gross area of column, j= joint width, taken as column width usually, Aj, = total area of effective horizontal joint shear reinforcement, and Jyx = yield stress of hoops. All the required shear hoops need to be positioned between the fop and bottom layers of beam reinforcing. That is, hoops placed above and below the top and bottom beam steel Tespectively are not effective in carrying @ proportion of the horizontal joint shear. Intermediate longitudinal column bars are usually needed to carry a proportion of the vertical joint shear. ‘These bars often can be used in estimating the flexural strength of the column as well, as they will not usually all be yielding at the design seismic column moment. ‘The shear force carried by the vertical shear steel: when N, is tensile, Vay should be intempolated berween 0.6V, when NV, =0 and zero when N, = 0.2f-” (tensile), The total area of vertical joint shear reinforcement within a width bj should not be less than Aj, = Vy Af, With shallow beams with large principal reinforcement ratios and correspondingly large joint shears there will often be considerable difficulty in placing all the required horizontal Joint hoops. In fully ductile frames the joint hoops would usually be placed one hoop set directly on top of the next set with no clear space between. Tt is therefore recommended for limited ductile frames that the principal beam reinforcement Eqn. 6.14 Eqn. 6.15 Page 6-35Australian Earthquake Engineering Manual ratios (4,/b,) be kept 10 0.02 or Hess, so as to reduce the problems of placing the beam - column joint hoops, Ts As fy be/ by EXTERNAL BEAM ~ COLUMN JOINT FORCES Figure 6.35 External beam -column joint forces and joint width restrictions 6.4.4 Diaphragms Diaphragms which transfer earthquake induced horizontal floor forces to primary lateral force resisting elements should be designed to resist the shear force requirement described previously, namely: Diaphragms should be designed as seismic walls with their full length detailed as a potential plastic hinge (end zone) except when the diaphragm can resist forces corresponding to elastic response (t= 1, K=4) without yielding. In this case the confinement requirements for seismic walls need not apply. Page 6-36 bj JOINT WIDTH LESSER OF by or beh 2 A Re ;, JOINT WIOTH LESSER OF bw or babe, 2 Eqn. 6.16Design of buildings Diaphragms which must transfer horizontal seismic shear forces from one vertical primary lateral force resisting element to another should remain elastic and therefore should be designed for the larger seismic forces (& = 4), (Figure 6.36). TRANSFER OIAPHRAGM SEISMIC SHEAR WALL SEISMIC WALL COLUMNS, Figure 6.36 All diaphragms should be reinforced as for two-way slabs with reinforcing in both directions; the amount at least satisfying the shrinkage and creep requirements. Where precast flooring elements are used, a cast-in-place ‘reinforced concrete topping of at least 50mm thickness may be used to wansfer the seismic shear forces through diaphragm action. Requirements for the topping include reinforcement in {wo directions at least to minimum shrinkage levels, proper bonding of topping to precast elements if composite action under gravity load is to be relied upon and if composite action is not required then the surface of precast elements must be clean and intentionally roughened. If the shear stress due 10 diaphragm shear transfer by the topping alone exceeds O.3,if.° MPa, ties with effective area of 4mm2/m? of floor area should be used to connect the topping to the precast elements, ‘The spacing of these connectors $1500mm with tributary area Of topping/connector not greater than 2.25m? and the connectors ‘must engage horizontal reinforcing or be suitably anchored in both topping and precast elements, (Taken from NZS 3101 (Ref. 6.5), Section 14) Page 6-37Australian Earthquake Engineering Manual 6.4.5 Structural steel ‘The recommendations in the NZ Steel Code (NZS 3404:1989, (Ref. 6.13) relating to limited ductile structures assumes structural ductility factors between 1.25 and 2. (K between 3.2 and 2). However, for Australian seismicity levels limited ductility conditions should apply for K factors down to 1 (1=4). Moment redistribution may be used up to @ limit of 15% reduction in the peak bending moments with equilibrium between intemal and extemal forces being maintained. The energy dissipating elements (plastic hinges) should be carefully chosen and detailed while the other structural elements, including connections, should be designed with an over strength factor of at least 1.35, 10 allow for material strength variations and strain-hardening. For satisfactory yield behaviour the following limits on the steel properties are recommended: a) Maximum specified yield stress of 450 MPa, b) Maximum ratio of yield stress to ultimate stress = 0.9 and c) Minimum length of yield plateau to be 3 times the yield strain ‘Maximum width to thickness ratios for limited ductile sections are based on Clauses 10.8 of AS 1250:1981 (Ref. 6.14) (also see Ref, 6.15). Maximum strains up to 10 times the yield strain are expected before local buckling occurs, (Table 6.2). Pans of members | Parts of members requiring limited requiring elastic, ductility behaviour Flanges and plates in compression with ba 136 136 one stiffened edge T (eg. Lor [ flanges) Flanges of welded box sections in bay 512 560 compression T Flanges of rectangular hollow bah 420 635 | sections T | Webs under flexural aie 1120 1600 | compression T Webs under uniform af, siz 635 compression | T Table 6.2 Maximum widih to thickness ratios. Page 6-38Sle Design of builgings Clause 10.9 (AS 1250) specifies requirements as to the spacing Of lateral restraints that should be used for sections of I ductility Pants of members yielding requiring limited ductility Flanges length where M >0.85Mpo > 6407, ei Spacing of braces within length where M > 0.85 Mpc one brace required Spacing to brace adjacent to length where M > 0.85 Mpc under uniform compression Table 6.3 Spacing of lateral restrainis Note: Pars of member in elastically responding structures should be braced in accordance with allowable stress rules. For columns, when P/P, > 0.15, 0.85 factor becomes 0.75, ‘The spacing limits are given in Table 6.3 assuming a Ratio R=10 (rotational section ductility = plastic hinge Totation/elastic rotation) giving a value of 1 for a in the AS 1250 formula: Sl + RB) Frames with weak beams - strong columns, should be designed ior multistory situations where ever possible. Column plastic hinges will usually be required to form in iow rise, long span portal frames, Steel I or box sections used as columns can dissipate energy by forming hinges if local buckling is prevented and low levels of axial load are present, A maximum axial force ratio of 0.7 for PIP, is recommended for limited ductility columns, calculated from the Clause 1053.1. AS 1250 formula, P/P,=(1+B-2)/(1 + B-%) where B is the ratio of end moments, largest over smallest, Positive for double curvature and negative for single curvature 4. is the normalised slendemess ratio LE wii Identical maximum width to thickness ratios as for beams (Table 6.2) are recommended for column sections, as are the Spacing of lateral restraints given in Table 6.3 Eqn. 6.17 Eqn. 6.18 Page 6.39‘Australian Earthquake Engineering Manual Connections should remain elastic wherever possible during the earthquake and should therefore be designed for the appropriate over strength design actions, (over strength factors = 1.35 for G250 steel and 1.50 for G350 steel). The recommended minimum design actions for all connections are a) 50% of the maximum in plane strength of the member in tension or compression, O54,F,); b) 30% of the maximum in plane elastic flexural strength of the member (0.30 Z F,); and ©) 15% of the strength of the member in shear (0.15 Ay F,). ‘These actions need not be applied simultaneously but rather (a) above or (p) and (c) together, The NZ Code (Ref. 6.13) recommends 67% higher values for (b) and (c) design actions above when considering column splices in the lower half of a multistory building, but these are considered conservative and will probably be reduced when further testing is completed. Although structural steel is a very ductile material this inherent benefit can be greatly reduced by stress raisers caused by sharp reentrant comer and notches which greatly increase local strains leading {0 premature failure. Care must be taken in the detailing of components, welds and fasteners so that smooth stress flows can take place between them. ‘Some common ductile beam-column joints of moment resisting frames, welded ané bolted, are shown by figures 6.37 — 6.40. Lamellar tearing, (the fracture cracking or separation of parent or base metal primarily in planes parallel to the plane of rolling, due to high through thickness strains, most often as 2 result of weld contraction, (Ref. 6.16), can occur in joints which are thick and relatively short and stiff. Such joints are common in moment resisting frames. Most occurrences of tearing happen when butt or fillet welds greater than 20mm are made on the surface of plates or sections (> 30mm thick) when under high restraint, (Figure 6.41). Figure 6.41 also shows some basic joint weld designs which should be used 10 avoid lamellar tearing during an earthquake. Bolted connections where yield strains are likely to be exceeded should have drilled rather than punched holes. This is due to the micro-cracking which occurs at the exit side of punched holes, which can cause brittle behaviour and is @ source of stress raisers. Reaming of punched holes has been shown not 10 considerably reduce this problem. Page 6-40 he.dour wo ples | SECTION A.A cq far Hetydiath ome | ae v Ee Nae pee sulfener ———f Mie column web sioner ah, te Figure 6.37 Structural stel beam-column joint, ductile moment-resisting frame SECTION A.A, ‘column flange doubler plates often ‘welded to YY flange doubler web plate column web. siifener ~ HSFG bots end plate Figure 6.38 Bolted beam-colun joint, ductile moment-resisting frame Design of puitdings T G Mv tT c c joint forces Joint forces Page 6-41Australian Earthquake Engineering Manual ange plates may extend ver column lengor sifener 4 ELEVATION horizontal —__ stiffeners ag: veces NY in other direction SECTION 4-A Figure 6.39 Fully welded 3-D beam-cotuman joint oe ifthe y wv Ww Lb ee vr ws al ws web doubler [PP Figure 6.40 Ductile moment-resisting space frame beam-columas Page 6-42Design of buildings Wildes jcnt o avoid detects Figure 641 Basic joint design to avoid Lamellar tearing. From Ref6.16 Page 6-43Australian Earthquake Engineering Manual 6.4.6 Page 6-44 Reinforced masonry Duc to the restrictions in placing transverse confining reinforcing within concrete block masonry, as well as the variable strengths of the components; block, grout and mortar, the ductility potential of reinforced concrete cannot be matched by reinforced masonry. Therefore higher K values should be ‘used for reinforced masonry construction when compared with similar reinforced concrete structures. A muliplier factor of 1.25 is suggested. For example, for moment resisting space frames, K,, = 1.25 x 1 = 1.25. Again, as for reinforced concrete, limited ductility requirements should suffice for Australian seismic conditions. Thus the Kp, values (K for reinforced masonry) should be in the range of 1.25 103.2. ‘The NZ Reinforced Masonry Code (NZS 4230:1990, (Ref, 6.17) only allows masonry structures to a height of 4 storeys when a light roof is used (3 storeys for heavy roof situation). ‘These limits could probably be relaxed in Australia’s seismic areas. In plastic hinge zones, a maximum total shear stress of 1.20 MPa is recommended for engineer supervised construction with the compressive stress of concrete masonry, jin’ =8 MPa. A shear stress of 0.15 MPa is the maximum to be provided by the masonry. (1.60 MPa and 0.3 MPa for general non-seismic conditions, total shear stress and shear stress carried by masonry respectively). The shear strengths provided should have a ‘margin over the required flexural strength in plastic hinge end regions ((o preclude brittle shear failures) such that 4V, = Vp + Vin + 2Veqs $= 0.85 for shear, Vp is shear due to dead oad, Vig = shear due to reduced live loads and V,, is shear due to lateral earthquake, It should be noted that the ultimate strength design method should be used for reinforced masonry, rather than the traditional allowable (working) siress method, with a maximum concrete strain of 0.0025 for non-ductile design Flexural strengths provided outside plastic hinge regions should be such that OM; 2 Mp + Mig + 1.5 M,q.=0.9 for flexure where Mp , Mix and M,, are the bending moments corresponding to the loads above. Figure 6.30 shows potential plastic hinge (end) regions plastic hinge length, ‘The maximum spacing of shear reinforcement within end regions shall not exceed one quarter of the member's depth for limited ductility design. it should be noted that concrete masonry can be confined by the use of steel plates in the mortarcourse between blocks at the ends of walls or columns with the Subsequent increase in the maximum concrete strain to 0.008 (See NZS 4230, Ref. 6. 17) for more details). These confining. Plates should be unnecessary for limited ductile structures, Walls Minimum reinforcement Walls in seismic areas need to be reinforced both vertically and horizontally. ‘The minimum area of reinforcement in each direction should be 0.07% of the gross cross-sectional area of the wall Perpendicular to the steel being considered. However, the sum Of the horizontal and vertical reinforcement ratios should be at Teast 02% of the gross cross-sectional areas. Stack bonded walls should have their minimum horizontal reinforcement increased to 0.15%, The minimum reinforcement used should be 10mm diameter. Maximum reinforcement In a flue or cavity the max. area should not exceed 8f, times, ‘area of flue or cavity and at laps the total area should not exceed 13f, Maximum spacing of wall reinforcing, horizontally and vertically should be 1000mm and 600mm for running and stack bond respectively. Lateral ties should be used around vertical reinforcement when its percentage exceeds 1% of the gross wall cross-sectional area. Where shear reinforcement is required when y, exceeds v,,) the minimum area of shear reinforcement should be found from A,=0.15 6, s/f, . where 6, (mm) is the effective web width and s (mm) is the spacing of the shear stcel, Spacing of this reinforcement should not exceed 05h, 0.5 D, ‘nor 600 mm for beams and columns nor 0.5 by for walls. Beam - column joints Dimension limitations Where limited ductility plastic hinges may form at column faces the horizontal joint dimension in the plane of loading: he = 60 dyy, when f, = 300 MPa 5 dyy when f, See Figure 6.42, Minimum vertical dimension of joint D =50 dy. , f= 300 MPa = 80 dy... f =430 MPa Beam bar anchorage lengths should be measured from h,/2 from the beam entering column face oF 10 dyy, . whichever is the lesser, (Figure 6.43). Design of buildings Page 6-45Australian Earthquake Engineering Manual Horizontal joint shear reinforcement ‘The horizontal shear to be resisted by horizontal shear reinforcement, in Eqn. 6.19 y, Y= GV © = 0.8 for masonry in shese Where Vy = 0.52 BV where ¥, = 0.15 MPa for limited ductile situations Total Area of horizontal joint reinforcement, Aj, = Vsa/fy, ‘Vertical joint shear reinforcement Eqn 6: Vertical design shear, Vj, oaeae Yop where Vy ‘The total area of vertical joint reinforcement, Ay = Vie/f, ‘The spacing of horizontal and vertical joint reinforcement should not exceed 200mm. abc dob OF 50 dpc fy = 300HPa UA Peet > 60 dpb Figure 6.32 Reinforced masonry minimum interior beam-column dimensions Page 6-46Design of buildings ANCHORAGE LENGTH 6.5 dhe | ob © Figure 6.43 Reinforced masonry minimum exterior beam-column dimensions Strengthening of risk b dings Most buildings in Australia will not have been designed or detailed to withstand any seismic forces and many may never ‘experience any seismic shaking in their lives. However, after the experience of the December 1989 Newcastle earthquake the city authorities or the owners may decide some seismic ‘upgrading or strengthening is necessary to reduce the tisk of Joss of life in important buildings or buildings where occupancy levels are high. Many Earthquake Risk Buildings are commercial buildings built before the Second World War and constructed of Uunreinforced brick or stone masonry, materials known to behave very poorly in even low to medium intensity seismic shaking Unreinforced masonry exhibits negligible ductility and usually fails suddenly in a britle shear mode. Masonry buildings often have considerable lateral strength in the plane of their walls due to the large mass of material used and the small proportion of ‘area of openings to total wall area. However, these short, stiff structures usually have natural periods close to the predominant period of the carthqual (< 0.53) and. therefore experience high lateral accelerations, ‘The other major deficiency with unreinforced masonry buildings is that the structural parts are often poorly connected; the floors, roof and walls are seldom attached to each other in Such a way that seismic induced forves can be distributed from the floors and roof into the surrounding walls and thence into the foundations. Masonry walls lack any of the tensile bending 0250 dpe Page 6-47Australian Earthquake Engineering Manual strength required to sustain out of plane forces and tend to fall ‘outwards by failing near their bases and thus allowing the floors and roof to collapse. See Figure 6.44 for a diagrammatic representation of common hazards. If the floors and roof can be attached satisfactorily to the walls, seismic forces are able to be transferred to transverse walls which may have sufficient shear strength in their plane to carry these forces. Some typical problems associated with the performance of masonry buildings in carthguakes arc: 1. Gables not tied to roof, thereby falling outwards. 2. Walls not tied to floors or transverse walls and no bending strength to resist out of plane forces. Parapets: no strength to resist out of plane forces. Large openings for shop frontages - lack of stiffness below Ist floor causes torsional response of building, comers fall ‘out, oF total collapse, 5. Timber floors and roofs lack strength in plane, unable to act as diaphragms to transmit forces back to walls, 6. Masonry walls lack shear strength in wall plane - britle shear failures occur. ‘Many ‘non-structural’ parts on or within earthquake risk buildings constitute seismic hazards and risks to life. These include parapets, chimneys, water tanks, omamentation, insecure fire escapes, balconies, verandas, insecure heavy equipment and heavy light fittings, In small to medium earthquakes it is usually these items which fail first causing injury and death in and outside the structure, often blocking exits and means of escape. If it is felt these items are sufficiently dangerous they should either be secured, strengthened or removed. In New Zealand, where the seismic risk is much greater than in Australia, commercial buildings built before 1935 have been classified into categories of risk and the worst cases either demolished or strengthened to half the 1965 seismic code levels, of force. Time limits for this work to be completed by have been set by several local authorities. In the most seismically active areas of Australia it may be appropriate practice to follow 4 similar procedure for old unreinforced masonry structures and more modern taller buildings lacking any ductile detailing and satisfactory connections between structural elements. ‘The timber floors often used in unreinforced masonry buildings are usually only capable of acting as horizontal diaphragms for low force levels. During strengthening these floors may require the addition of plywood sheathing or diagonal boarding with close nail spacing to produce the required diaphragm strength (Under floor horizontal steel trusses could also be employed). Such methods probably still won't provide ‘infinite’ stiffness to the diaphragm so that every wall clement will not receive an Page 6-48applied shear force proponional to its stiffness. Thus Strengthened walls would be designed for inertia forces from respective tributary floor areas, forces from the wall's own inertia and tributary wall face forces, The NZ Earthquake Society recommendations for earthquake Fisk buildings (Ref. 6.18) includes a chan for structural type factors, SE for differing types of sirengthening system. Thie vs reproduced in Table 6.4 with the conversion made to the equivalent Australian code horizontal force factor Kp. Design of buildings CHIMNEYS CRASH THROUGH ROOF Figure 644 Seismic weaknesses in unreinforced buildings. (See text ‘for explanation of numbers) Page 6-49Australian Earthquake Engineering Manual em Description of Strengthening System Kg (Note 2) 1 Ductile frames - (Note 1) 08 2 Cast in situ concrete shear walls and single 038K storey columns detailed for limited ductility 3 ‘Sprayed concrete or gunited shear wall detailed for ductility in accordance with NZS 3101: a) Two or more elements linked by a 1.67 diaphragm b) Single elements 1.67 4 Cast in situ, sprayed concrete of concrete 1.67 block shear walls detailed for limited ductility 5 1.33 6 Buildings with diagonal stee! bracing - 08K (Note 3) 7 Single storey cantilevered building supported by: a) Double reinforced face loaded wall either 1.67 sprayed or cast in situ b) Single reinforced face loaded walls 20 8 | All timber strengthening systems 9 One and two storey buildings in Occupancy 08 Classifications 3 & 4 of Table 3A with Iateral force resisting systems from items 2 10 8 above but with total gravity Joad support in addition to load-bearing masonry walls. 10 Elastically responding systems 2) Reinforced concrete or reinforced 33 masonry b) Steel or pre-stressed concrete 42 nl Unreinforced masonry walls, 42 Table 6.4 ‘rucuural Type Factors for Earthquake Risk Buildings Ky Where Kp= Structural type factor from AS 2121:1979 Page 6-50Notes 1. Structural type 1 ductile frames should not be used for buildings of OC.1 and OC.2 of Table 6.5 Unless the frame stiffness approximates the stiffness of the remnant structure, This form of structure is not recommended for structures over two storeys of OC.3 and OC.4 where the frame stiffness is not within 50% of the remnant structure, 2 Where reinforcing or prestressing cables with J,> 380 MPa are used the above factors should be multiplied by 1.1, 3. Where a mixed system of diagonal steel bracing and unreinforced masonry is used the Kg for the steel bracing should be 1.0K. In such cases the following criteria should also be satisfied: i Not more than 20% of total seismic load shall be carried by existing masonry. i Loads should be distributed according 10 relative stifiness and Kg value should be weighted average applicable 10 the combination of new strengthening system, iii Gi) above need not apply where assumption is made that all loads are taken by new strengthening, iv Compatibility of deformations must be achieved, 4. Inall cases, particularly Item 9, structural integrity ‘must be maintained under vertical load, ‘The occupancy classification (OC) should be determined by Considering both the occupant Toad and the intensity of ‘occupation Occupant = The maximum number of people exposed to Load Oy isk during the normal functioning of the building, including those immediately outside the building. Weekly hours of Occupant _ Occupant load | normal’ occupancy Intensity 0; * Gross floor area 40 Where floor area is measured in 100's of m2, ‘The Occupancy Classification is determined as follows: For essential buildings, OC = 1 For all other buildings, OC is determined from Figure 6.45, Design of builtings Page 6-51Australian Earthquake Engineering Manual 200) 150 Ocoupare load (Oy) 100. 0) 3 0 5 30 Occupant intensity (Oy) ( persons J 10m? ) Figure 6.45 Occupancy classification (non-essential buildings) (Ref.68) ‘The Risk to certain types of buildings is related to their Occupancy Classification as given in Table 6.5 from Ref. 6.18. It should be noted that the Risk Factor, R, is equivalent to the Importance Factor, J, in AS 2121 but R has a larger range of values. 1 has only two values, one for essential facilities and one for all other buildings Page 6-52Design of buildings pee ee EE Category” Description Occupancy Classification (OC) | as EEE 4 2a Buildings which are intended to remain 1 functional in the emergency period of major earthquakes 2 Buildings whose failure could cause high 1 Joss of life in the surrounding area 1 Structures containing highly hazardous 1 contents 3a Buildings which should be functioning in the 2 Restoration Period for major earthquakes 3b Buildings whose contents have a high value 2 to the community Buildings containing many people, the 2 Collapse of which could cause a high loss of life 4 Buildings with nonmal occupancy of usage 3or3 Buildings with a less than normal occupancy 4 Table 6.5 Relationships Between Category, Class and Occupancy Classification (OC), * Risks Factor of 2 down to 1 (From Table 4, NZS 4203:1984, (Modified]). In NZ it has been common practice to strengthen earthquake risk buildings to force levels of between one half and two thirds of the 1965 loading code requirements. One possibility i that Australian Earthquake Risk Buildings should be strengthened {or at least the full AS 2121 lateral force levels, asthe existing ‘lesign seismic coefficients are so small. Limited ductility Fequirements should be met wherever possible in the strengthening undertaken ‘There are a multitude of ways in which an Earthquake Risk Building can be strengthened and every case must be Considered carcfully before work begins. The future use and expected life of the building will be paramount in deciding where and how strengthening should be attempted, Some of the ‘more common strengthening methods are 1. Sprayed conerete (up to 150 mm thick) onto existing masonry walls Grouting up the cavity between brick skins with expansive admixtures to increase bond Page 6-53Australian Earthquake Engineering Manual 3. Structural steel braced frames for walls, floors and roof diaphragms 4, Reinforced concrete walls and new slabs cast over old timber floors and adequately attached to existing walls using exposed bars 5, Reinforced concrete bond ring beams around building attached to roof trusses and timber floors 6. Ground anchors to resist overturning under higher lateral forces 7. Prestressing or Macalloy bars to induce compression stresses in walls in an attempt to keep the wall in one piece, increase the shear strength of the wall and reduce the need for tensile bending strength. A problem with filling the cavity between brick skins or between veneer and main walls with concrete or grout is that water seeps through the brickwork releasing calcium salts which produce efflorescence on the faces of the wall. Brick walls need to be in sound condition with relatively good mortar joints for drilling and placement of vertical reinforcing or restressing cable prior to grouting, Lime mortar was predominantly used before the First World War and deteriorates badly over time to such an extent that it has negligible shear strength. Cement mortars retain their strength much better and are less affected by weathering. Figures 6.46, 647, 6.48, 6.49, and 6.50 show actual strengthening deiails used in NZ earthquake risk buildings (References 6.19, 6.20, 6.21, 6.22). All of these buildings are single or two storey built prior to 1935. (The NZ Codes contained no seismic design requirements till 1935). It is often difficult to ascertain the reliable strength of masonry walls and horizontal diaphragms (floors and roofs) to be used in the strengthening calculations. Cores can be drilled in masonry walls but these only give an estimate of the shear strength, as the whole wall’s strength is dependent on the combined effects of often weathered mortar joints, cracking and other damage, much of which may not be visible. A portion of the wall may be cut out and tested off-site but this is expensive and often impractical, especially with historic buildings. Page 6-54Design of buildings [ 9 omreo 400 al cern — enna 3 Wt 8 3 es op onan ayo bron Atrettarac, fous FES eR Figure 6.46 Strengthening masonry walls. (Ref, 6.19) Page 6-55Australian Earthquake Engineering Manual pe es mats rahe ate etoteae +s as ator, conan © alia LGiasssanns = | i 5 = Sonat ar pas: for 0x 50 seus ssoxsa oe sco ks Betwebe sts caso AcA #688675 UG earn tka ES Hees PGP Mice CLs PICA TAILS =r etis or an rc srRUCEX == exSTHG TRE t choo FcR FIGURE & ss aie spe CUP EAD eedhodo smn BI an zon es. my es e/a . es sors 4 sm SE [ee THREADED 3$ea———_| ‘Beno OL | EoD tom 25056 : ‘hac stew ExcH as eee TRUBOLT- | Figure 6.47 Floor strengthening and floor to wall connections. (Ref. 6.19) Page 6-56Design of builaings La” New reintorce — tarecete eat ae Goenice dowelted FE ip ee "fe eenerele. Hard plaster Tie bock on 1-5m. sure grid (Gwer stressing bers. 75mm. reintocced FLOOR PLAN a ae ee concrete footing eee ~ existing lee Timber post cast into “new slab. CT rsisting wet _ ‘ondereat Ye ——— od form ‘shear key, Existing concrete Feeting. new POST SEATING DETAIL WALL STRENGTHENING DETAILS Figure 6.48 Wall strengthening from Ref. 6.20 Page 657Australian Earthquake Engineering Manual sexrco9- 17 ai ootsg) voted steve 29 aL aster pegne: Gol 4HS How 238 hyping Yassrid YOM BeeneeZAO a a 3 Wisd NY 234 c09¢:s290) sincavnsd L253 __| ova—I|- srsiee yrange erusua) Fai raster wstaa9 oh d3reMt ebtoat stxe0) PNuLstia BSeecarocs: $5ce ab areO~ ly ‘CBhe suid oo co8: anamey | cones “eh any “ool Yad SWHIS *ONO9 No-AeLES MOG] Ed MANNS FETO! sr aNisd SN ASSN sae voor “ussH 01. 30734 WOVL Smita 3woss mraces-s's92 cid sesart weee | oben Figure 649 Strengthening details from Ref. 6.21 Page 6-58Design of buildings SNOWD3NNOD” ONS F FOOT —WoIGAL 2 voypas 1 uopes SNOIIDSNNOD BOOTS jpyooaden ws GuLcoy pote meus wodun aye WaidaL Figure 6.50 Floor and roof to wall connections from Ref, 6.22 Page 659Australian Earthquake Engineering Manual Recommended allowable strength values for existing unreinforced masonry, timber diaphragms, timber walls. concrete footings, cic are reproduced from Ref. 6.18 in Table i 66 Item Materials Strength values 1 Unreinforced masonry Seismic in-plane shear stress based on gross area when: a) Maximum ratio of unsupported height or 130 kPa” length to thickness not more tian 10 b) Maximum ratio of unsupported height or 100 kPa* length to thickness not more than 16 ©) Maximum ratio of unsupported height or notallowed Jengih to thickness greater than 16 2 Horizontal Diaphragms 8) Roofs with straight sheathing and roofing 2.2.kNim for applied directly to the sheathing seismic shear ) Roofs with diagonal sheathing and roofing 5.8 kN/m for applied directly over sheathing. seismic shear ©) Floors with straight vongue and groove 2.2 kN/m for sheathing seismic shear 4) Floors with straight sheathing and groove 44 kNim for sheathing seismic shear ©) Floors with diagonal sheathing and 5.8 kNimn for finished wood flooring seismic shear 1) Floors or roofs with sheathing and plaster ‘Add 0.7 kN/m tothe re-nailed to the joints or rafters to the allowable values for approval of the Local Authority items 1(@) and 1(0) 3 Wood walls a) Wood stud walls with wood lath and 0.7 kN/m each side for plaster seisinic shear ) Wood stud walls with plaster and lath other | 1.4KN/m each side for than wood lath seismic shear 4 Plain concrete footings Je= 10 MPa unless ‘otherwise shown by st 5 Reinforcing steel f,= 200 MPa (Gntess tested) 3 Structurat steel J, 200 MPa } For all other stresses, bearing, shear bolts, (unless tested) | compression etc, reduce NZS 3404 for grade 250 | by 2007250 Table 6.6 Recommended allowable strength values Page 6-60Design of buildings * Values may be increased by the addition of 30% of the computed axial compressive stress (under 0.9D) due 10 loading from above. A supporting element shall have a stiffness of at least twice that of the tributary wall, based on gross section properties. Notes 1. Values in this table are based on the Los Angeles ‘Ordinances for earthquake risk buildings. 2. Materials to be sound and in good condition. 3. The wood lath and plaster must be re-nailed 10 existing joists or rafters in an approved manner. 4, Design engineers must satisfy themselves that values quoted above can safely be applied in particular cases. Domestic housing ‘Timber houses usually behave well even in major earthquakes, due to their high levels of redundancy and the good energy absorbing capability of nailed timber connections. However, items such as brick chimneys, brick veneers and unbraced underfloor foundations often cause damage and inconvenience during and after small earthquakes. Brick chimneys which are usually unreinforced and inadequately held at floor and roof levels sway out of phase with the flexible timber house, causing pounding at roof level which causes failure and collapse of the chimney above that level. If the roofing material is masonry or ceramic tiles the broken chimney often crashes through the roof; corrugated stecl roofing is more resilient to such an impact. Unreinforced brick or concrete block veneers which are not adequately tied to their timber stud walls often fail in diagonal tension (shear) with major cracking occurring between adjacent diagonal comers of door or window openings. Untied unreinforced veneer has negligible out of plane resistance and fends to fall outwards, thus blocking exit ways and also allowing subsequent water damage. Piles should be tied to foundation walls used to support the veneer. Inadequately braced foundations allow the house to bodily move off its foundations causing the house to be uninhabitable Gue to smashed flooring, broken services, broken windows and non-closing doors. The most common causes are unbraced jack studs (the vertical extension of the pile under the floors joists). ‘These jack studs act as a soft storey when inadequate diagonal bracing is provided. When the house's bearers are not tied adequately to the piles the house will also move laterally Simple wiring out of the pile stapled to the bearer or floor joists corrode away over time leaving no lateral restraint. At least 12mm galvanised bolis should be used at this connection. More Page 6-61Australian Earthquake Engineering Manual 6.6 6.6.1 roof bracing in the roof plane may be necessary 10 stop heavy tiled roofs from flexing and dislodging tiles, Wire netting may be placed over the roof rafters or the ceiling joists to stop tiles falling through the ceiling, Modem houses with large windows and doorways require diagonal bracing in their walls using steel strap or, structural plywood sheathing (7.5mm) or glass fibre reinforced board. The recommended nailing of sheet bracing is 50 x 2.5 mm galvanised fat head nails @ 150mm centres at all edges. ‘The most important item in a house after a large earthquake is, the hot water tank, which if secure will provide drinking water in an emergency, as the reticulated supply will inevitably be cut off for some time, either by pipe damage or for checking of the system. Hot water tanks are usually very unstable but can be stabilised by simply fitting steel straps around the tank and securing them to surrounding wall studs. A useful practical publication on how to strengthen houses against earthquakes is available from N.Z., Ref. 6.23. Non-structural elements and connections Infill panels Non-structural elements can modify a structure's response by stiffening it and lowering the natural period and thus usually increasing the seismic induced accelerations experienced by the building. Such non-structural elements include infill wall panels, lift and service core walls and stair flights, In areas of high seismicity infill walls should be seismically separated where they butt against columns and along their top, under beams, (Figure 6.51). no ~~ seismic GAPS —_| Teste SEPARATED INFILL] Aeceres ered WALL PANEL Tee Yt Page 6-62 1 \ SEISMIC FRAME Figure 651 Separated infill masonry wall tae sll Fa Sanaa ee ata caA MAI ms———— aeDesign of buildings In Australian seismic zones this requirement is probably only Recessary in limited ductility frames over a certain height, Perhaps $ storeys? However in all frame structures with lunseparated infill masonry wall panels, the panels should be checked to see that they can withstand the extra in-plane seismic forces without failing in a brite manner. Preferably they should be designed elastically with K=4 for in-plane as well as face loads, or for regular constant storey height structures design the frame and panels for: 3.33n nti K= where n= number of storeys (Taken from NZS 4230, 1990 (Ref. 6.17). In most cases it would be better to design the structure as a shear wall form rather than a pure frame with a corresponding langer K’ value Special care needs to be taken where the infill panels are ing from one or two floors in a frame, as this sudden change in stiffness will usually promote the formation of Column sway (soft storey) mechanisms, The affected columns will not have been detailed for plastic hinging and collapse of a storey may occur. Unseparated infill panels on 3 sides of a building, as a requirement to leave unrestricted glazing on the front facade of shops, will often promote severe torsional response and inflict damage to the ends and comers of the walls as shown earlier. Unseparated infill and secondary concrete walls should always be reinforced and the recommended maximum spacing in concrete masonry in buildings higher than 2 storeys are 1000mm, centres both horizontally and vertically for running bond and 600mm centres both ways for stack bond walls, wiNoow HEIGHT NON sepaRaTED IN~ FILL PANEL SEISMIC FRAME Figure 652 Non-separated infill panel with window opening at top Infill walls which don't extend to the full column height (Figure 6.52) have often induced brittle column shear failures, due to Eqn. 6.21 {LARGE sHoRT COLUMN SHEARS Page 6-63Australian Earthquake Engineering Manual 6.6.2 6.6.3 the much larger column shear forces induced when column plastic hinges attempt to form over the short clear height. It should be noted that separated infill walls will considerably stiffen the beam they sit on, which will concentrate the beam plastic hinges over a short length, or worse, force the hinges into the columns. Unseparated infill panels with openings require special study to ascertain if diagonal bracing action is achieved. ‘Separated masonry infill walls (partitions or non bearing walls) should be not less than 90mm thick and should have maximum reinforcing spacing of either 1000mm each way or 600mm vertical and 2800mm horizontally. Reinforced galvanised wire Lattice laid in the horizontal mortar joints is allowed in the NZ code (NZS 4230, Ref. 6.17) for wall thicknesses less than 110 mm. Veneers Unreinforced veneers have performed particularly badly in small to moderate earthquakes and they usually fall onto foot paths, break through light verandas or block exit ways. Veneers need to be tied back at regular centres to the walls behind. The NZ Code recommended maximum height for unreinforced veneers in the lowest seismic zone (C) in general locations is 6 m for timber support walls and 8 m for concrete or masonry support. Where the veneer is adjacent to expressways and public places the maximum height from the ground to the highest part of the veneer is only 3 metres, which means a maximum of about S00 mm of unreinforced veneer above an exit way. ‘These limits may be too extreme for Australian seismic risk levels but masonry veneer is brittle and falls to pieces at low ground accelerations. Thus high veneers should either be reinforced or the public below must be protected with suitably designed canopies. Stairs Flights of stairs can act as diagonal braces between floors in flexible framed buildings. These unintentional braces. will modify the structure's response and if they fail will cut off the means of escape in multistory buildings. In flexible structures stairs should only be fixed at one level and sliding joints or hangers be provided at other levels, (Figure 6.53) Page 6-64 aves ' t i 'Design of builgings STAIRS FIXED ONLY AT 0p HANGERS SUIOING JOINT, ELEVATION RUBBER OR STEEL PLATES TI | | beth sevaration—f[ ae U it . iit Se/et err HANGERS PLAN Figure 653 Separated stairs 6.6.4 Glazing Glass curtain walls and glazing panels can be dangerous during earthquakes if movement between the glazing and its frame is not possible in the plane of the glazing. Glass can only carry small shear stresses before breaking. These shears are forced on the glass by the frames attempting to follow the lateral building deflections. Flexible gaskets between the glass and surrounding transoms and mullions will usually provide enough movement (15-20mm/interstorey) for moderate earthquakes. Problems will usually occur at the comers of the building when the structure moves diagonally. Rounded glazed comer details will prove almost impossible to detail for the required seismic drifts. Seismic detailing for glazed cunain wall systems has been developed in the U.S. and NZ, (refer to “Architectural Design for Earthquake”, 6.24). In the recent Californian Loma Prieta Earthquake (Oct. 17-1989) the majority of non-structural damage occurred in medium rise steel-framed buildings (ERI Report, 6.25), where the structure's Mlexibility caused excessive storey drifts, even in this medium (M = 7.1) earthquake. Page 6-65Australian Earthquake Engineering Manual 6.6.5 Precast panels and their connections Heavy precast concrete cladding panels need to be carefully attached to the building in such a way that they do not add to the lateral stiffness of the structure, This usually implies fixing in all directions at top of beam level and flexible fixings allowing movement only in the plane of the panel at or near the bottom of the beam, (Figure 6.54). cous, PANEL FIXED IN ry — (‘ ALL DIRECTIONS CO TT L Abxeo oH Zz LE = | FREE TO | HOVE IN 4 PANELS. PLANE 1 i owe ‘SEPARATION ALLOWANCE FOR VERTICAL AND BETWEEN PANELS HORIZONTAL HOVENENT IN PaNet'S PLANE BUT NOT J TO PANEL: (OUCTILE BOLT FOR EXAMPLE) PANELS SPANNING BETWEEN FLOORS Figure 654 Panels spanning between floors ‘The connections should be near the ends of the beams rather than near mid-span and should not be fixed within a column's clear height. Any connection system which allows the panels to rotate in their plane during the earthquake will cause conflict between panels with subsequent life-hazardous damage. Precast cladding panels, like other non-structural elements, should be designed for seismic forces whose magnitude depends on where the panel is positioned on the building, whether the panel is ductile or brittle, the panel's material, the seismic risk of the panel falling and hitting someone below, and the seismic coefficient of the building it is attached to, Non- structural elements positioned near the top of a building will experience much larger accelerations than elements fixed near the base, so therefore should be designed for larger lateral forces, How the building itself responds to the “quake dictates the displacements and accelerations applied to the panel. Thus the building's own seismic coefficient is important. ‘The heavy Page 6-86Design of builaings 6.6.6 panel's seismic design coefficient in the NZ Code (NZS 4203) when positioned near the top of a multistory structure is about 4 times the building's seismic coefficient for a reasonably ductile ‘concrete panel. The panel must therefore be designed for a force equal 10 this part's seismic coefficient times the panel's sclf- ‘weight applied nommal to the panel's face. ‘The connections of precast panels to the main structure are even ‘more critical than the panel design itself because if the connections lack ductility and/or strength and they fail the panel could fall. For the example quoted the design seismic coefficient would be about 30 times greater than the building's Coefficient in the NZ context. The connection design forces specified for panels over pedestrian areas (high risk) often work Out to be between I and 4 times the proportion of the panel's weight carried by each fixing acting in any horizontal direction. ‘This is due to structural deformations causing large forces on the panels, as well as the panel's inertial forces. The connections are designed for strength rather than ductility as one often can't rely on the ductility of bolts and cast-in- sockets fixed into britue thin concrete sections. It is doubtful whether panels and connections in Australia should be designed to such force levels but they should be designed to seismic coefficients greater than that used to design the building they are fixed to. This is especially important in tall flexible tower buildings. A value of 2 for C, for the connection of exterior panels is given in Tabie 71 of AS 2121 (Ref. 6.1), (C, is the seismic response factor of the part). Suspended ceilings Unbraced suspended ceilings have often partially or entirely failed in eanhquakes. Heavy acoustic ceiling panels can jump out of their supporting rails, due mainly to vertical accelerations which can be as large as the horizontal acceleration components Of the earthquake. Unbraced ceilings (pendulum type) will vibrate independently of the structure around them causing damage to walls and partitions and distortions to the ceiling grid, often causing the release of ceiling tiles or the collapse of large portions of the ceiling. Heavy light fittings incorporated into the suspended ceiling grid are particularly vulnerable. Fire sprinkler systems can rupture if not connected 10 the structure securely. Quoting from the NZ loadings Code (4203) “Holding down devices for elements should allow their casy removal and Feplacement for access into the ceiling space. Where the suspended system depends on Isteral support from the structure, the members transmitting these loads need to be adequately spliced to transmit the specified forces without buckling and/or deforming at splices in @ manner that will result in exce cumulative deformations. Page 6-57Australian Earthquake Engineering Manual ‘rocing form A oor of building tecsion ond wie woe fe wire brace vqvore bk “tivng save ct o WE Soe “Grol bs lection SF romner fined to ceiling rnver by screwing theough Hh cseas bolton pate AY | ete to Hor Figure 655 Suspended ceiling details from Ref. 6.26 Page 6-68Design of buildings 6.6.7 Hangers need (0 be designed to withstand compressive loads, but they do need to be connected to the ceiling and structure in a manner that allows relative upward movement without disengagement.” The horizontal scismic coefficient for suspended ceilings in NZ's least zone of risk ranges from 0.2 to 0.8 dependent on height of ceiling above ground and the building's risk factor. The equivalent Australian seismic coefficient should probably be about 50-67% of those values. (The present non-essential building's suspended ceiling force has a maximum of 0.108 W, in the current AS 2121 code, where W, is the weight of the ceiling under consideration.). Some details of suspended ceiling bracing and details at the edge and at partitions are shown in Figure 6.55. (Ref. 6.26) Interstorey deflections Over recent years there has been vigorous discussion about what seismic interstorey drifts should apply and what should be the limiting drift for buildings with non-separated, non- structural elements. The present NZ Code, NZS 4203, (Ref. 6.2), specifies an interstorey drift limit of 0.0006 of the storey height for non-separated elements. The absolute interstorey drift limit is 0.01 of the storey height where elements are separated. Both of these limits are multiplied by a factor, equal to 2/3 for the least risk seismic zone C. The actual interstorey drifts are found by linear multiptication of the elastic lateral deflections under code lateral forces. The multiplication factor usually lies between | and 3; unity for limited ductile structures and higher values for fully ductile reinforced concrete and structural steel frames. However, the code points out that these interstorey drift levels only represent about 1/3 of the deflections likely to occur in a moderate earthquake (El-Centro 1940, for example’ substantial damage to separated panels, glazing, etc is likely to ‘occur due to elements hitting each other or binding up in their glazing frames, in moderate shakes. In the Austr imterstorey drift limits of about 0.0006 (unseparated) and 0.010 of the storey height (separated) may be appropriate realistic levels to eliminate most non-structural damage 10 building cladding in the highest seismic risk areas. (The proposal in NZ is to increase the interstorey drift limit to 0.025 and change the multiplying factor to substantially increase the calculated inelastic drifts when the structure forms hinges. This will have the effect of either forcing the design of stiffer strucures or having to increase the septrations required between non- structural elements and the supporting structure). The current Australian code specifies a separation found by multiplying the deflections under code forces by 3/K for buildings in Zones 1 and 2. Page 6-69