Chapter 06-Architectural Consideration

275

Chapter 6

Architectural Considerations

Christopher Arnold FAIA, RIBABuilding systems Development Inc.

Key words: Configuration, Regular Configurations, Irregular Configurations, Proportion, Setbacks, Plan Density,Perimeter Resistance, Redundancy, Symmetry, Asymmetry, Soft-Stories, Weak Stories, Code Provisions,Plan Irregularities, Elevation Irregularities, Architectural Implications.

Abstract: While the provision of earthquake resistance is accomplished through structural means, the architecturaldesign, and the decisions that create it, play a major role in determining the building's seismic performance.The building architecture must permit as effective a seismic design as possible: at the same time thestructure must permit the functional and aesthetic aims of the building to be realized. The three categoriesare: (1) the building configuration, (2) structurally restrictive detailed architectural design, and (3)Hazardous nonstructural components. This chapter discusses one other issue that bears on the architecturaldecisions that affect seismic performance: that of the methods by which mutual architectural andengineering seismic design decisions are made during the building design and construction process. This, inturn, leads to some consideration of the architect/engineer relationship as it affects the seismic designproblem.

276 Chapter 6

6. Architectural Considerations 277

6.1 INTRODUCTION

While the provision of earthquake resistanceis accomplished through structural means, thearchitectural design, and the decisions thatcreate it, play a major role in determining thebuilding's seismic performance. The buildingarchitecture must permit as effective a seismicdesign as possible: at the same time thestructure must permit the functional andaesthetic aims of the building to be realized.

The architectural design decisions thatinfluence the building's seismic performancecan be grouped into three categories. Thesecategories are not exclusive, and each categoryof decision may influence the others, but it isuseful to structure the decisions in this waybecause it clarifies the influences and theirmutual interactions.

The three categories are:

• The building configuration: This is definedas the size, shape and proportions of thethree-dimensional form of the building. Theterms building concept, or conceptual design,are often also loosely used by architects toidentify the configuration, although theseterms also refer to architecturalcharacteristics such as internal planning andbuilding organization. Strictly speaking,configuration refers only to the geometricalproperties of the building form.

• Structurally restrictive detailed architecturaldesign: This refers to the architectural designof building details, such as columns or walls,that may affect the structural detailing inways that are detrimental to good seismicdesign practice.

• Hazardous nonstructural components: Thedesign of many nonstructural components isthe architect's responsibility, and ifinadequately designed against seismic forces,they may present a hazard to life. In addition,they may represent a major cause of propertyloss, and in the case of essential facilities or

other services, their damage may cause lossof building function. Engineering issues inthe design of these components are dealtwith in Chapter 14.

This chapter discusses one other issue thatbears on the architectural decisions that affectseismic performance: that of the methods bywhich mutual architectural and engineeringseismic design decisions are made during thebuilding design and construction process. This,in turn, leads to some consideration of thearchitect/engineer relationship as it affects theseismic design problem.

6.2 CONFIGURATIONCHARACTERISTICS ANDTHEIR EFFECTS

6.2.1 Configuration Defined

For our purposes building configuration canbe defined as building size and shape: the latterincludes the characteristic of proportion. Inaddition, our definition includes the nature, sizeand location of the structural elements, becausethese are often determined by the architecturaldesign of the building, and are a subject ofmutual agreement between architect andengineer. This extended definition ofconfiguration is necessary because of theinteraction of these elements in determining theseismic performance of the building.

In addition, architectural decisions mayinfluence the nature, size and location ofnonstructural components that may affectstructural performance, either by altering thestiffness of structural members or changing themass distribution in the building.. Theseelements are generally part of the initial conceptof the building but they may be added later,when the building is in operation. Thisparticularly applies to in-fill walls, which mayhave a dramatic effect on the effective height,stiffness, and load distribution of columns. Inthis chapter they are discussed later as separate

278 Chapter 6

issues, apart from their relationship toconfiguration. These include such elements aswalls, columns, service cores, and staircases,and also the quantity and type of the exteriorwall elements.

6.2.2 Origins and Determinants ofConfiguration

The building configuration, or concept, isinfluenced by three main factors:

• urban design, business and real estate issues.• planning and functional concerns.• image and style

The selected configuration is the result of adecision process that balances these varyingrequirements and influences and, within abudget, resolves conflicts into an architecturalconcept. In very general terms three basiccategories of architecture can be distinguishedbased on their main objective:

• Economical containers -the "decoratedshed": warehouses, industrial plants, somedepartment stores and commercial buildings

• Problem/solving, functional facility -

hospitals, educational, laboratories,residential,

• Prestigious and/or high-style image -corporate headquarters, some publicbuildings and university buildings, museums,entertainment, and some retail stores.

These categories also bear some relationshipto the architects, or firms, that design them, forthere is much covert specialization inarchitecture. This can cause client confusion:when the client who wants an economicalcontainer goes to a prestige architect, or whenthe client with a difficult planning problem goesto the container architect.

Building function and planning produce ademand for certain settings and kinds of spacedivision, connected by a circulation pattern forthe movement of people, supplies, andequipment. These demands ultimately lead tocertain building arrangements, dimensions anddeterminants of configuration.

Urban design and planning requirementsmay affect the exterior form of the building. Aheight limit may set a certain maximum height;the street pattern may, particularly in a denseurban situation, determine the plan shape of thebuilding, at least for its lower floors. City

Figure 6-1. Set-back regulations, New York


planning requirements sometimes dictate theneed for open first floors, for vertical setbacks,or other characteristics of architectural form.Urban design includes issues such as zoningand planning regulations, which by definingset-backs, height limits and sun-anglerequirements often define the buildingenvelope.

For example, recent studies have arguedconvincingly that early skyscraper form waspredominantly determined by local land-usepatterns, municipal codes and zoning (Figure 6-1). For example, the striking differences in formbetween the skyscrapers of Chicago and NewYork were due to the imposition of a 130 feetheight limit on the former, and no limits on thelatter. Zoning laws in New York, in 1916,spawned the buildings with "wedding-cake"setbacks, while a 1923 law in Chicagopermitted a tower to rise above the old heightlimit, but restricted its total volume(6-1).

Engineers can accept the problems of zoningand building function in determiningconfiguration, because they fit into theengineer's rationalist concept of the world. It isthe third influence, the need for the building topresent an attractive, interesting, unique, oreven sensational image to the outside observer,and often the occupants, that engineers feel thetrouble begins. Here is where the irrationalartist takes over, and the laws of physics andeconomy may be violated.

It is important to understand the need for thearchitect sometimes to provide a distinctiveimage for the building. If this need did not existthe owner might go to an engineer -orcontractor- to obtain a simple economicalbuilding, and indeed, many owners do so..

Up until the early years of the 20th. centuryfor a Western architect the common acceptancerequired a historical style -typically mediaevalor renaissance - even when totally new buildingtypes such as railroad stations or skyscraperswere conceived. In engineering and materialsterms these traditional forms were all derivedfrom masonry structure: the need to keep theblocks of masonry in compression, and thecreation of devices such as arches and vaults, to

enable the masonry to achieve larger spans thanwere possible by using slabs of masonry asbeams or lintels. These masonry determinedforms survived well into the 20th. century, evenwhen buildings were supported by concealedsteel frames, and arches had become astructural anachronism. Moreover, theprevailing historical architectural stylespreferred symmetricalness, and decreed thatbuildings should be massive at the base, withsmaller openings, and their mass shoulddecrease with the upper floors.

Figure 6-2. The International Style

The revolution in architectural aestheticsthat began in the 1920's, and is often called the"International Style" was based on exploitingthe forms that could be created by use of framestructures, combined with a desire to striparchitecture of its decoration and adherence tohistoric styles The International Style inarchitecture was not alone in extoling thevirtues of unadorned structure and absence ofdecoration in its glorification of the beauties ofEuclidean geometry. The same thing was goingon in the world of painting and sculpture, andthese arts were being stripped of theirtraditional content in favor of simplicity,geometry, and new materials.

As architects began to exploit the aestheticsof an architecture based on engineered frames,the seeds of seismic configuration problemwere sown. Load-bearing masonry buildingswere very limited in the extent to whichconfiguration irregularities were possible: withshort spans redundancy was always present: the

280 Chapter 6

extensive use of walls, both in exteriors andinteriors, meant that, even though the masonrywas unreinforced, unit stresses were very low.Large cantilevers and setbacks were notpossible.

But with the steel or concrete frame all theselimitations were unnecessary: the buildingstructure could be unbelievably slender(because now the columns and beams wereanalyzed and sized by engineers), first floorwalls could be omitted, so that the buildingseemed to float in space. Lightness and gracewere sought, rather than ornamented mass.(Figure 6-2) Buildings could even cantileverout safely so that they could become larger asthey rose: the inverted pyramid could be built.These possibilities were eagerly explored by anew generation of architects: with them cameother ideas: the rejection of symmetricalness ofplan in favor of a more exciting and morerational disposition of elements (rationalbecause the building elements were allowed tooccur where planning function was mostefficient, instead of being forced into[sometimes] inefficient symmetry).

Examples of the International Style werelimited to a few avant-garde buildings in allcountries before World War 2, and thenbloomed in the rich economic years that beganin the 50's. The United States , Western Europe,Latin America, the Soviet Union and Japanexploded in a fury of development, almost allconstructed in their regional versions of theInternational Style. These years of intensivedevelopment saw the world's cities grow intohuge metropolises: they were also years inwhich seismic design as it related to the new,spare, framed buildings was inadequatelyunderstood, and it took earthquakes in LatinAmerica, Mexico and the United States (inAlaska, 1964, and San Fernando, 1971) to makeengineers realize that such buildings wereunforgiving and intolerant of the veryirregularities that architects had embraced withsuch enthusiasm.

This architecture of the 50's to the 70's hasleft us with a legacy of poor seismicconfigurations that present a serious problem in

reducing the earthquake threat to our cities. Theproblem is exacerbated when it is allied to theengineering design problem of the use of thenon-ductile reinforced concrete frame structure,which was the norm up to about 1975.

This historical discourse is relevant toseismic design, because it shows that:

• the minimalist structural frame provided thebasis for an architectural aesthetic whichwas in tune with the spirit of the age,aesthetically, economically and politically.

• what we now call discontinuities andirregularities were critical elements of thenew architectural aesthetic.

• these elements were made possible by theuse of the engineered structural frame, andby a new level of architect/engineercollaboration.

It is, however, worth mentioning, that thenew style originated , was promoted anddeveloped in Western Europe, predominantlyFrance and Germany, which, of course, areessentially non seismic zones.

A more complete discussion of the originsand influence of the International Style will befound in Reference 6-2.

6.2.3 Configuration Influences inGeneral

Configuration largely determines the waysin which seismic forces are distributedthroughout the building, and also influences therelative magnitude of those forces. For a givenground motion, the major determinant of thetotal inertial force in the building is , followingNewton's Second Law of Motion, the buildingmass (approximated on the earth's surface by itsweight). While the size and shape of thebuilding (together with the choice of materials),establish its weight the building square footageand volume are determined by the buildingprogram (and the budget) : the listing ofrequired spaces and the activities andequipment that they contain. But for any givenprogram an almost infinite variety of


configurations can provide a solution, and it isthe variables in these configurations that affectthe distribution of inertial forces due to groundshaking .

Thus the discussion of configurationinfluence on seismic performance becomes theidentification of configuration variables thataffect the distribution of forces. These variablesrepresent irregularities, or deviations from a"regular" configuration that is an optimum, orideal, with respect to dealing with lateral forces.

6.2.4 The Optimum SeismicConfiguration.

It is easiest to define a regular building byproviding an example: the design discussedbelow represents an essentially perfectly regularbuilding, which in turn represents an"optimum" seismic design. Its characteristicsare such that deviations from the designprogressively detract from its intrinsic seismiccapabilities: these deviations result in"irregularities" and a familiar list ofconfiguration irregularities can be identified.The discussion of these irregularities from anengineering and architectural viewpoint formthe main body of this section..

Architecture implies occupancy: thus a solidblock of concrete, which might be an optimumseismic design, is sculpture, not architecture.The great pyramid of Gizeh is architecture, andcertainly approaches an optimum seismicdesign, but architecturally it is very uneconomicin its use of space and volume in housing onlytwo small rooms within an enormous volume ofunreinforced masonry (Figure 6-3). Our optimalseismic design is compromised by the need alsoto be reasonably optimal architecturally -that is,in its ability to be a functional andeconomically viable architectural concept.

Our design shows the three basic ways ofachieving seismic resistance, and these are alsopart of the optimization, so the building isseismically optimized architecturally, in itsconfiguration, and also demonstrates the bestarrangement of its seismic resisting elements, incomplete harmony with the architecture (Figure

6-4). For convenience, the building is arbitrarilyshown as three stories: a one story buildingmight be better seismically, all other thingsbeing equal, but with a multi-story building wecan show some necessary attributes of such abuilding.

Figure 6-3. The great pyramid of Gizeh

Considered purely as architecture this littlebuilding is quite acceptable, and would besimple and economical to construct. It is also aprototypical International Style building.Depending on its exterior treatment - itsmaterials, and the care and refinement withwhich they are disposed- it could range from avery economical functional building to anelegant architectural jewel; it is not complete,architecturally, of course, because stairs,elevators etc. must be added, and the building isnot spatially interesting , although its interiorcould be configured with nonstructuralcomponents to provide almost any quality ofroom that was desired with the exception ofinteresting and/or unusual spatial volumes morethan one story in height.

What are the characteristics of this designthat make it regular, and also make it so good -considering only architectural configuration andthe disposition of the seismic resistingelements? Any engineer will recognize them,but it is worth while listing them, because theyare specific attributes whose existence orabsence thereof can be quickly ascertained inany actual design. These attributes, and theireffects, are:

282 Chapter 6

• Low height-to base ratio Minimizes tendency to overturn

• Equal floor heights Equalizes column/wall stiffness

• Symmetrical plan shape Reduces torsion

• Identical resistance on both axes Balanced resistance in all directions

• Uniform section and elevations Eliminates stress concentrations

• Maximum torsional resistance Seismic resisting elements at perimeter

• Short spans Low unit stress in members

• Redundancy Toleration of failure of some members

• Direct load paths, no cantilevers No stress concentrations

6.3 METHODS OF ANALYSIS

6.3.1 Methods of Analysis and theRegular Building

An important aspect of a building's responseto ground motion is the method of analysis usedto establish the seismic forces. The estimate oftotal forces and their distribution is both afunction of and a determinant of the lateralforce-resisting system employed in thebuilding. The great majority of designs estimatelateral forces through use of the staticequivalent lateral force method (ELF)established in typical seismic codes , whichinvolve estimating a base shear and thendistributing the resulting forces through thestructural elements of the building. It is

Figure 6-4. The optimal seismic design


important to recognize that the forces derivedfrom an equivalent force method usedaccording to a typical seismic code and manyother code provisions, assume a regularbuilding, comparable to our ideal formdescribed above. This assumption is noted inthe Commentary to the 1997 NEHRPRecommended Provisions for SeismicRegulations for New Buildings(6-3): "TheProvisions were basically derived for buildingshaving regular configurations. Past earthquakeshave repeatedly shown that buildings havingirregular configurations suffer greater damagethan buildings having regular configurations.This situation prevails even with good designand construction"

The Commentary to the 1990Recommended Lateral Force Requirements ofthe Structural Engineers Association ofCalifornia (Ref.6-4), discusses the design basisfor regular buildings in some detail. Twoimportant concepts apply for regular structures.First, the linearly varying lateral forcedistribution given by the ELF formulas are areasonable and conservative representation ofthe actual response force distribution due toearthquake ground motions. Second, when thedesign of the elements in the lateral forceresisting system is governed by the specifiedseismic load combinations, the cyclic inelasticdeformation demands will be reasonablyuniform in all elements, without largeconcentrations in any part of the system. Theacceptable level of inelastic deformationdemand for the system is therefore reasonablyrepresented by the Rw value for the system.However, "when a structure has irregularities,then these concepts, assumptions andapproximations may not be reasonable or valid,and corrective design factors and proceduresare necessary to meet the design objectives".

It is safe to say, based on studies of buildinginventories, that over half the buildings thathave been designed in the last few decades donot conform to the simple uniform buildingconfiguration upon which the code is based. Fornew designs, the simple equivalent lateral force

analysis of the code must often be augmentedby engineering judgment based on experience.

Progressive evolution of seismic codes hasresulted in increasing force levels and theconsideration of additional parameters inestimating force levels, but the impact ofconfiguration irregularity, which was firstintroduced into the Uniform Building Code in1973, long remained a matter of judgment.However, starting in 1988 the UBC quantifiedsome configuration parameters, to establish thecondition of regularity or irregularity, and laiddown some specific analytical requirements forirregular structures.

6.3.2 Irregular Configurations: CodeDefinitions and Methods ofAnalysis

In the Commentary to the 1980 SEAOCRecommended Lateral Force Requirements andCommentary(6-5), over 20 types of "irregularstructures or framing systems" were noted asexamples of designs that should involve extraanalysis and dynamic consideration rather thanuse of the normal equivalent lateral forcemethod. These types are illustrated in Figure 6-5, which is a graphical interpretation of theSEAOC list. Scrutiny of these conditions showsthat the majority of irregularities areconfigurational issues within the terms of ourdefinition.

This list of irregularities defined theconditions, but provided no quantitative basisfor establishing the relative significance of agiven irregularity. These irregularities vary inthe importance of their effects, and theirinfluence also varies in accord with theparticular geometry or dimensional basis of thecondition. Thus, while in an extreme form thereentrant corner is a serious type of planirregularity, in a lesser form it may have littlesignificance (Figure 6-6). The determination ofthe point at which a given irregularity becomesserious is a matter of judgment.

284

Figure 6-5. Graphic interpretation of "Irregular Structures or Framing Systems" from the commentary to the "SEAOCRecommended Lateral Force Requirements and Commentary" (a) Buildings with Irregular Configuration (b) Buildings with

abrupt changes in lateral resistance (c) Buildings with abrupt changes in lateral stiffness (d) Unusual or novel structuralfeatures.


The SEAOC Commentary explained thedifficulty of going beyond this basic listing asfollows:

Due to the infinite variation of irregularities(in configuration) that can exist, theimpracticality of establishing definiteparameters and rational rules for theapplication of this Section are readilyapparent.

However, in the most recent version of theSEAOC Requirements and Commentary, andstarting in the 1988 revisions to the UniformBuilding Code, (which is based on the SEAOCdocument), an attempt has been made toquantify some critical irregularities, and todefine geometrically or by use of dimensionalratios the points at which the specificirregularity becomes an issue of such concernthat remedial measures must be taken.

Figure 6-6. The reentrant corner plan : a range ofsignificance

The code approach to reducing thedetrimental effect of irregularity is to requiremore advanced methods of analysis where suchconditions occur - more specifically, where theELF analysis method must be augmented orcannot be used. While this may provide a moreaccurate diagnosis, and in some instancesstrengthening of certain members, it does notcorrect the condition: this must still be done bydesign means based on understanding of theeffects of the condition on building response.

The code requirements relating to thedefinition of regularity and irregularity, and thedetermination of the analysis methods requiredhave now become complex, and for designpurposes the relevant sections of the applicablecode should be referred to. The outline thatfollows focuses on identifying the irregularconditions for which the ELF method can be

used, must be augmented or where a morecomplex method is necessary. The irregularitytype references are to the 1997 NEHRPRecommended Provisions for the Developmentof Seismic Regulations for New Buildings asillustrated in Figure 6-7. This figure is a graphicinterpretation of Table 5.2.3.1 and Table 5.2.3.2in the Provisions. The terminology andconfiguration requirements in the UBC and theNEHRP Provisions are essential similar.

The ELF method can be used for thefollowing irregular structural types, with thenoted augmentations:

1. All structures in Seismic Design Category A(in the NEHRP Provisions the SeismicDesign Category is a classification assignedto a structure based on its seismic use group,or occupancy, and the severity of the designearthquake ground motion at the site).

2. Structures with reentrant corners ( planirregularity type 2), diaphragm discontinuity(type 3) out-of-plane offsets (type 4) , inSeismic Design Categories D, E and F, mustprovide for an increase in design forces of25% for connection of diaphragms to verticalelements and to collectors, and connection ofcollectors to vertical elements.

3. Structures with nonparallel systems (planirregularity type 5) in Seismic DesignCategory C,D,E and F, must be analyzed forseismic forces applied in the criticaldirection, or satisfy the followingcombination of loads: 100% of forces in onedirection plus 30% of the forces in theperpendicular direction.

4. Structures with out-of-plane offsets (planirregularity type 4) and in-planediscontinuity in vertical lateral force resistingelements (vertical irregularity type 4) musthave the design strength to resist themaximum axial forces that can develop inaccordance with specially defined loadcombinations.

286

Figure 6-7. Irregularities defined in the 1997 NEHRP Provisions


Other buildings with plan or verticalirregularities as defined in the Tables, that arenot required to use modal analysis as identifiedbelow, may use the ELF procedure with"dynamic characteristics given specialconsideration" : the engineer must use judgmentin computing forces.

Buildings with certain types of verticalirregularity may be analyzed as regularbuildings in accordance with normal ELFprocedures. These buildings are generallyreferred to as setback buildings. The followingprocedure may be used:

1. The base and lower portions of a buildinghaving a setback vertical configuration maybe analyzed as indicated in (2) below if all ofthe following conditions are met:

a.The base portion and the tower portion ,considered as separate buildings, can beclassified as regular and.

b.The stiffness of the top story of the base isat least five times that of the first story ofthe tower.

Where these conditions are not met, thebuilding shall be analyzed using modalanalysis.

2. The base and tower portions of the buildingmay be analyzed as separate buildings inaccordance with the following:

a.The tower may be analyzed in accordancewith the usual ELF procedure with thebase taken at the top of the base portion.

b.The base portion then must be analyzed inaccordance with the ELF procedure usingthe height of the base portion of hn andwith the gravity load and base shear ofseismic forces the tower portion acting atthe top level of the base portion.

Modal Analysis is required in the followinginstances:

1. Buildings which are in Seismic DesignCategory D, E or F, are over 65 feet inheight, and have:

soft stories (vertical irregularity type 1a)extreme soft stories (vertical irregularitytype 1b)mass irregularities (vertical irregularity type2)vertical geometrical irregularity (verticalirregularity type 3)

Exceptions: vertical structuralirregularities of types 1a, 1b or 2 do notapply where no story drift ratio underdesign lateral load is greater than 130percent of the story drift ratio of the nextstory above

2 Buildings , with torsional irregularity (planirregularity type 1a) in Seismic DesignCategory D, E or F and extreme torsionalirregularity ((plan irregularity type 1b) inSeismic Design Category D. In addition anincrease in design forces of 25% is requiredfor connection of diaphragms to verticalelements and to collectors, and connection ofcollectors to vertical elements, and a torsionamplification factor.

3. All structures over 240 feet in height.

The following irregular structures are notpermitted:

Weak story structures (vertical irregularitytype 5) over 2 floors or 30 feet in height witha weak story less than 65% of the strength ofthe story above, in Seismic DesignCategories, B, C, D, E and F.

Extreme soft story structures (verticalirregularity type 1b) and extreme torsionalirregularity structures (plan irregularity

288 Chapter 6

type 1b) in Seismic Design categories E andF.

The Commentary to the NEHRP Provisionsalso provides a procedure which may reduce theneed to perform modal analysis.

"The procedures defined in the Provisionsinclude a simplified modal analysis whichtakes account of irregularity in mass andstiffness distribution over the height of thebuilding. It would be adequate, in general, touse the ELF procedure for buildings whoseseismic resisting system has the sameconfiguration in all stories and all floors, andwhose floor masses and cross sectional areasand moments of inertia of structuralmembers do not differ by more than 30% inadjacent floors and in adjacent stories.

For other buildings, the following criteriashould be applied to decide whether modalanalysis procedures should be used:

1. The story shears should be computed usingthe ELF procedure.

2. On this basis, approximately dimension thestructural members, and then compute thelateral displacement of the floors.

3. Replace the hxk term in the verticaldistribution of seismic forces equation withthese displacements and recompute thelateral forces to obtain new story shears.

4. If at any story the recomputed story sheardiffers from the corresponding value asobtained from the normal ELF procedure bymore than 30%, the building should beanalyzed using the modal analysis procedure.If the difference is less than this value, thebuilding may be designed for the story shearobtained in the application of the presentcriterion and the modal analysis proceduresare not required."

This procedure greatly reduces thelikelihood that the considerably more complexmodal analysis procedure will be required forthe building analysis: this is of majorimportance because building irregularity isquite likely to be present in buildings of modestsize and tight budget, and costly analysisprocedures are not welcome to the owner.

In addition, the 1997 NEHRP Provisionsmake further predominantly nonquantitativecomments about the use of the EquivalentLateral Force procedure for irregular buildings:

"The ELF procedure is likely to beinadequate in the following cases:

1. Buildings with irregular mass and stiffnessproperties in which case the simple formulasfor vertical distribution of lateral forces maylead to erroneous results:

2. Buildings (regular or irregular) in which thelateral motions in two orthogonal directionsand the torsional motions are stronglycoupled, and

3. Buildings with irregular distribution of storystrengths leading to possible concentration ofductility demand in a few stories of thebuilding.

In such cases, a more rigorous procedurewhich considers the dynamic behavior of thestructure should be employed.

The Provisions Commentary points out thatthe ELF procedure, and both versions of themodal analysis procedure (a simple version anda general version with several degrees offreedom per floor which are described in theProvisions) are all likely to err systematicallyon the unsafe side if story strengths aredistributed irregularly over height. This pointsto the importance of eliminating suchirregularities if possible, but often they will bepresent because of detailed architecturalrequirements: if they cannot be eliminated, theengineer must use his judgment to assess theireffects on the analysis


Even if the modal analysis procedure is usedthere are limitations to the information that theanalysis provides. The procedure adequatelyaddresses vertical irregularities of stiffness,mass or geometry. Other irregularities must becarefully considered on a judgmental basis, andso the engineer must rely on his experientialand conceptual knowledge of the building'sresponse in order to effectively accommodateall irregularities.

6.4 GENERAL BUILDINGCHARACTERISTICS

6.4.1 Introduction

These are issues relating to the buildingconfiguration as a whole and apply to allconfigurations.

Irregularity as defined in current seismiccodes , and as discussed above, covers themajority of configuration variables that have asignificant effect on the seismic performance ofthe building. Although definitions vary, there isgeneral agreement on those configurationirregularities that are important.

However, the code listing is not complete:issues of building proportion and size are notincluded, nor are issues such as the buildingplan density or its redundancy the subject ofcode provisions, although the latter is brieflymentioned.. These are discussed below. Theproblem of pounding, which combines the issueof drift with that of building adjacency, and assuch may present an architectural problem, isdiscussed in Section 6.9 below.

6.4.2 Size, Proportion and Symmetry

• Building size:

It is possible to introduce configurationirregularities into a wood frame house thatwould be serious problems in a large building,and yet produce a safe structure with theinclusion of relatively inexpensive andunobtrusive provisions. This is because a small

wood frame structure is light in weight andinertial forces will be low. In addition, spansare short and relative to the floor area, therewill probably be a large number of walls toshare the loads.

For a larger building, the violation of basiclayout and proportion principles exacts anincreasingly severe cost, and as the forcesbecome greater, good performance cannot berelied upon as in an equivalent building ofbetter configuration.

As the absolute size of a structure increases,the number of alternatives for the arrangementof its structure decreases. A bridge span of300ft. may be built as a beam, arch, truss, orsuspension system, but a span of 3000 ft. canonly be designed as a suspension structure. Andas the size increases the structural disciplinebecomes more rigorous: architectural flourishesthat are perfectly acceptable at the size of ahouse become physically impossible at the sizeof a suspension bridge.(Figure 6-8).

Figure 6-8. The designer's suspension bridge

In looking at the influence of building sizeon seismic performance, the influence of boththe dynamic environment and thecharacteristics of ground motion result in morecomplexity than does the influence of size onvertical forces. Increasing the height of abuilding may seem equivalent to increasing thespan of a cantilever beam, and so it is (all otherthings being equal). The problem with theanalogy is that as a building grows taller itsperiod will tend to increase, and a change inperiod means a change in the building response.

The effect of the building period must beconsidered in relation to the period of groundmotion, and if amplification occurs, the effectof an increase in height may be quitedisproportionate to the increase itself. Thus

290 Chapter 6

doubling the building height from 6 to 10stories may, if amplification occurs, result in afour or fivefold increase in seismic forces. Theearthquake in Mexico City in 1985 resulted inmajor response and amplification in buildingsin the 6 to 20 story range, with generallyreduced response in well-built buildings belowand above these heights.

Although a 100-ft. height limit throughoutJapan was enforced until 1964, a 150-ft 13 storylimit was the maximum in Los Angeles until1957, and the limit was 80 ft and later 100 ft onSan Francisco, height is rarely singled out as avariable to be used to reduce the buildingresponse. Two recent exceptions to this may benoted. After the Armenian earthquake of 1988,planners of the reconstruction of the city ofLeninakan limited the height of new buildingsto three stories, because of the groundconditions and the bad experience with tallerbuildings. This decision is especially interestingbecause it required a major shift in planningand architectural thinking: prior to this, almostall Soviet-style housing consisted of medium tohigh-rise blocks. After the Mexico Cityearthquake of 1985 a number of damagedbuildings were "topped" as part of the repairstrategy: a number of floors were removed, thuschanging the building period to something lessin tune with the long period ground motionsthat the city experiences.

The present approach is generally not tolegislate seismic height limits (except insofar asseismic codes impose height limits relating totypes of construction), but to enforce morespecific seismic design and performancecriteria. Generally, urban design, real-estate orprogrammatic factors will be more significant,and earthquake performance must beengineered with the height predetermined bythese factors.

It is easy to visualize the overturning forcesassociated with height as a seismic problem(although the issue is more that of the aspectratio of shear walls rather than the building as awhole), but large plan areas can be detrimentalalso. When the plan becomes extremely large,even if it is symmetrical and of simple shape,

the building can have trouble responding as oneunit to the ground motion. Unless there arenumerous interior lateral-force resistingelements, large-plan buildings imposeunusually severe requirements on theirdiaphragms, which have large lateral spans, andcan build up large forces to be resisted by shearwalls or frames. The solution is to add walls orframes to reduce the span of the diaphragm,although it is recognized that this may introduceproblems in the use of the building. In a verylarge building, seismic separations may benecessary to subdivide the building and keepthe diaphragm forces within bounds, in whichcase the seismic separations may also act asthermal expansion joints.

An interesting example of a correct"intuitive" response to this problem is that ofthe design of the Imperial Hotel, Tokyo, by thearchitect Frank Lloyd Wright in the early1920s. He subdivided this large complexbuilding, with long wings and many reentrantcorners, into small regular boxes, each about 35ft. by 60 ft in plan. In doing this, he appears tohave been concerned about the possibility ofdifferential settlement caused by a travellingwave on the site. In the use of this concept, towhich he attributed in large measure the successof the building in surviving the 1923 Kantoearthquake, Wright was well ahead of his time.The short-pile foundation scheme, whichWright claimed as a major invention, probablyhad much less to do with the building's goodperformance(6-6).

• Building Proportion

In seismic design, the proportions of abuilding may be more important than itsabsolute size. For tall buildings, the slendernessratio (height/least depth) of a building,calculated in the same way as for an individualmember, is a more important consideration thanjust height alone. Dowrick(6-7) suggestsattempting to limit the height/depth ratio to 3 or4, explaining:


"The more slender a building the worse theoverturning effects of an earthquake and thegreater the earthquake stresses in the outercolumns, particularly the overturningcompressive forces which can be verydifficult to deal with."

As urban land becomes more expensive,there is a trend towards designing very slender"sliver" buildings which, although notnecessarily very high, may have a largeheight/depth ratio. Nowhere is this trend moreapparent than in Japanese cities, wheremultistory buildings may be built on sites thatare of the order of 15 to 20 ft wide (Figure 6-9).However, the same economic forces oftendictate that these buildings will be built veryclose together, so that they will tend to respondas a unit rather than as individual free-standingbuildings, although more recent Japanese

buildings have incorporated relatively largeseparations to reduce the risk of pounding.

•••• Building Symmetry

The term symmetry denotes a geometricalproperty of building plan configuration.Structural symmetry means that the center ofmass and center of resistance are located at, orclose to, the same point (unless live loads affectthe actual center of mass). The singleadmonition that appears in all codes and intextbooks that discuss configuration is thatsymmetrical forms are preferred toasymmetrical ones. The two basic reasons arethat eccentricity between the centers of massand resistance will produce torsion and stressconcentrations.

However, a building with reentrant cornersis not necessarily asymmetrical (a cruciform

Figure 6-9. Slender buildings, Tokyo, Japan

292 Chapter 6

building may be symmetrical) but it is irregular,as defined, for example, in current seismiccodes. Thus symmetry is not sufficient on itsown, and only when it is combined withsimplicity is it beneficial.

Nevertheless, it is true that as the buildingbecomes more symmetrical, its tendency tosuffer torsion and stress concentration willreduce, and performance under seismic forceswill tend to be less difficult to analyze. Thissuggests that when good seismic performancemust be achieved with maximum economy ofdesign and construction, the symmetrical,simple shapes are much to be preferred. Butthese tendencies must not be mistaken for anaxiom that a symmetrical building will notsuffer torsion.

The effects of symmetry refer not only tothe overall building shape, but to its details ofdesign and construction. Study of buildingperformance in past earthquakes indicates thatperformance is sensitive to quite smallvariations in symmetry within the overall form..This is particularly true in relation to shear-walldesign and where service cores are designed toact as major lateral resistant elements. It ispossible to have a building which isgeometrically symmetrical in exterior form, buthighly asymmetrical in the arrangement of itsstructural systems. The most common form ofthis condition (sometimes termed "falsesymmetry") is the building with interiorstructural cores that, for planning reasons, areunsymmetrically arranged. This can be a major

source of undesirable torsional response.(Figure 6-10)

Experience in the Mexico City earthquakeof 1985 showed that many buildings that weresymmetrical and simple in overall plan sufferedseverely because of asymmetrical location ofservice cores and escape staircases. Moreover,as soon as a structure begins to suffer damage(cracking in shear walls or columns, forexample), its distribution of resistance elementschanges, so that even the most symmetrical ofstructures becomes dynamically asymmetricaland subject to torsional forces.

Finally, it must be recognized thatarchitectural requirements will often make thesymmetrical design impossible. In thesecircumstances, it may be necessary, dependingon the size of the building and the type ofasymmetry, to subdivide the building intosimple elements.

There is a tendency, as noted above, for thevery tall building to tend towards symmetry andsimplicity. The seismic problems are mostapparent in the low to medium-height building,where considerable choice exists as to planform and the disposition of the major masses ofthe building.

6.4.3 Plan Density, Perimeter Resistance,and Redundancy

The size and density of structural elementsin the buildings of former centuries is strikinglygreater than in today's buildings. Structural

Figure 6-10. False symmetry: offset structural core


technology has allowed us to push this trendcontinually further.

Earthquake forces are generally greater atthe base of the building. The bottom story isrequired to carry its own lateral load in additionto the shear forces of all the stories above,which is analogous to the downward build-upof vertical gravity loads. At this same lowestlevel, programmatic and aesthetic criteria areoften imposed on the building that demand theremoval of as much solid material as possible.This requirement is the opposite of the mostefficient seismic configuration, which wouldprovide the greatest intensity of verticalresistant elements at the base, where they aremost needed.

An interesting statistical measure in thisregard is the ground level vertical plan density,defined as the total area of all vertical structuralelements divided by the gross floor area. Themost striking characteristic of the modernframed building is the tremendous reduction ofstructural plan density compared to historicbuildings.

For instance, a typical 10- to 20- story,moment resistant steel frame building will

touch the ground with its columns over 1% orless of its plan area, and combined frame shear-wall designs will typically reach structural plandensities of only 2%. The densely filled-in"footprints" of buildings of previous eraspresent a striking contrast: the structural plandensity can go as high as 50%, in the case ofthe Tag Mail: the ratio for St. Peter's in Rome isabout 25%, and for Chartres Cathedral 15%.The 16-story Monadnock Building in Chicago,which used exterior bearing walls of brick 6 ft.thick at the ground level, has a ratio of 15%(Figure 6-11).

Analogous to structural plan density is themeasure of the extent of walls in a structure.Surveys of damaged buildings in Japan andTurkey have indicated a clear relationshipbetween the length of walls in a box-typesystem building and the extent of damage. Thisrelationship has been incorporated in theseismic codes of these and other countries toprovide prescriptive guidance for the design ofsimple structures.

In Figure 6-12, although both configurationsare symmetrical and contain the same amountof shear wall, the location of walls is

Figure 6-11. Structural plan density

294 Chapter 6

significantly different. The walls on the rightform greater lever arms for resistingoverturning and torsional moments. In resistingtorsion, with the center of twist of asymmetrical building located at or near thegeometrical center, the further the resistingmaterial is placed from the center, the greaterthe lever arm through which it acts, and hencethe greater the resisting moment that can begenerated. Placing resisting members on theperimeter whenever possible is alwaysdesirable, whether the members are walls,frames, or braced frames, and whether theyhave to resist direct lateral forces, torsion, orboth.

Figure 6-12. Location of lateral resistance elements

The design characteristic of redundancyplays an important role in seismic performance,and is significant in several aspects, mostespecially because the redundant design willalmost certainly offer direct load paths and inthis it tends to result in higher plan density asdiscussed above. In addition, historic buildingstended to be highly redundant, because shortspans required many points of support, and thuseach supporting member incurs much lowerstresses, often even within the capability ofunreinforced masonry. Thus, the verylimitations of traditional materials forced thedesigners into good design practices such asredundancy, direct load paths and high plandensity.

The detailing of connections is often cited asa key factor in seismic performance, since themore integrated and interconnected a structureis, the more load distribution possibilities thereare.

6.5 SEISMIC SIGNIFICANCEOF TYPICALCONFIGURATIONIRREGULARITIES

6.5.1 Introduction

The discussion of configuration issues thatfollows incorporates all the code-defined issuesbut, in going back to our original definition ofconfiguration, categorizes configurationproblems in ways that relates the seismicimplications to those of their architecturalorigins as decisions made at the conceptualstages of the design.

For each configuration issue, five issues areoutlined: definition of the condition, its seismiceffects, its architectural implications, historicalperformance in past earthquakes, and solutions.The notes on architectural effects discuss theorigin and purpose of the condition inarchitectural terms: the discussion of solutionsdeals with conceptual design approaches, and ismost relevant for the consideration of existingbuildings.

6.6 PLAN CONFIGURATIONPROBLEMS

6.6.1 Reentrant Corners

•••• Definition

The reentrant , or "inside" corner is thecommon characteristic of overall buildingconfigurations that, in plan, assume the shape ofan L, T, H, +, or combination of these shapes.

•••• Seismic EffectsThere are two related problems created by

these shapes. The first is that they tend toproduce variations of rigidity, and hencedifferential motions, between different parts ofthe building, resulting in a local stressconcentration at the "notch" of the reentrant


corner. In Figure 6-13, if the ground motionoccurs with a north-south emphasis at the L-shaped building shown, the wing orientednorth-south will, for geometrical reasons, tendto be stiffer than the wing oriented east-west.The north-south wing, if it were a separatebuilding, would tend to deflect less than theeast-west wing, but the two wings are tiedtogether and attempt to move differentially attheir notch, pulling and pushing eachother.(Figure 6-14). For ground motions alongthe other axis, the wings reverse roles, but thedifferential problem remains.

Figure 6-13. Separated buildings

Figure 6-14. The L-shaped building

The second problem is torsion. This isbecause the center of mass and center ofrigidity in this form cannot geometricallycoincide for all possible earthquake directions.

The result is rotation, which tends to distort theform in ways that will vary in nature andmagnitude depending on the nature anddirection of the ground motion, and result inforces that are very difficult to analyze andpredict.

The stress concentration at the notch and thetorsional effects are interrelated. The magnitudeof the forces and the seriousness of the problemwill be dependent on:

• the mass of the building• the structural systems• the length of the wings and their aspect ratios• the height of the wings and their height/depth

ratios

In addition, it is not uncommon for wings ofa reentrant corner building to be of differentheight, so that the vertical discontinuity of asetback in elevation is combined with thehorizontal discontinuity of the reentrant corner,resulting in an even more serious problem.

The reentrant corner is perhaps the majorirregularity that will be found in olderbuildings, including unreinforced masonry. Inaddition, in such buildings it is rare to findseismic separations at the intersections of thewings, so the prospects for torsion and stressconcentration are high, when the wings are longand tall.

•••• Architectural Implications

Reentrant corners create a useful set ofbuilding shapes, enabling large plan areas to beaccommodated in compact form, while stillproviding a high percentage of perimeter roomswith access to light and air. Thus suchconfigurations are common for high-densityhousing and hotel projects, in which habitablerooms must be provided with windows.

Concerns for daylighting and naturalventilation that were prevalent during theenergy crisis of the 1970's resulted insomething of a revival of interest in theincreased use of narrow buildings and thetraditional set of reentrant corner

296 Chapter 6

configurations. The courtyard form, mostappropriate for hotels and apartment houses intight urban sites, has always remained useful. Inits contemporary form the courtyard oftenbecomes a glass-covered atrium, but thestructural form is the same.

•••• Historical Performance

Examples of damage to reentrant cornerbuildings are common, and this problem wasone of the first to be identified by observers. Ithad been identified before the turn of thecentury, and by the 1920s was generallyacknowledged by the experts of the day. Naito(6-8) attributed significant damage in the 1923Kanto earthquake to this factor. The samedamage phenomena were reported for the 1925Santa Barbara and 1964 Alaska earthquakes(Figure 6-15), and for the 1985 Mexico Cityearthquake Large wood frame apartment houseswith many reentrant corners are common in LosAngeles and suffered badly in the Northridgeearthquake of 1994.

•••• Solutions

There are two basic alternative solutions tothis problem: to separate the buildingstructurally into simple shapes, or to tie thebuilding together strongly at lines of stressconcentration and locate resistance elements toreduce torsion.

If a decision is made to use separationjoints, they must be designed and constructedcorrectly to achieve the intent. Structurallyseparated entities of a building must be fullycapable of resisting vertical and lateral forceson their own. To design a separation joint, themaximum drift (or some reasonable criterion)of the two units must be calculated by thestructural engineer. The worst case is when thetwo units would lean towards one anothersimultaneously, and hence the dimension of theseparation space must allow for the sum of thedeflections. In a tall building the relativemotion between portions of the building willbecome very large, and create major problemsof architectural detailing.

Figure 6-15. Damage concentrated at the intersection of two wings of an L-shaped school, Anchorage, Alaska, 1964


One of these is to preserve integrity againstfire and smoke spread. The MGM Grand Hotelin Las Vegas is a T-shaped building in plan,with seismic joints approximately 12 in. indimension. In the fire of 1983 these jointsallowed smoke to propagate to the upper floors,resulting in many deaths.

Several considerations arise if it is decidedto dispense with separation joints and tie thebuilding together. Collectors at the intersectioncan transfer forces across the intersection areas,but only if the design allows for these beam likemembers to extend straight across withoutinterruption. Walls in this same location areeven more efficient than collectors. (Figure 6-16).

Figure 6-17. Solutions to the L-shaped building

Since the free end of the wing tends todistort most under tension, it is desirable toplace resisting members at this location.

The use of splayed rather than right-anglereentrant corners lessens the stressconcentration at the notch, which is analogousto the way a rounded hole in a steel beamcreates less stress concentration problems thana rectangular hole, or the way a taperedcantilever beam is more desirable than one thatis abruptly notched (Figure 6-17).

6.6.2 Variations in Perimeter Strengthand Stiffness


This section discusses the detrimentaleffects of wide variations in strength andstiffness in building elements that provideseismic resistance and are located on thebuilding perimeter

•••• Seismic Effects

If arranged to provide balanced resistanceperimeter resistance elements are particularlyeffective in reducing torsional effects becauseof their long lever arm relative to the center ofresistance. If the resistance is not balanced, thedetrimental effects can be extreme.

Figure 6-16. Splay in plan relieves reentrant corner problem: analogies to beam

298 Chapter 6

This problem may occur in buildings whoseconfiguration is geometrically symmetrical andsimple, but nonetheless irregular for seismicdesign purposes. If there is wide variation instrength and stiffness around the perimeter, thecenters of mass and resistance will not coincide,and torsional forces will tend to cause thebuilding to rotate around the center ofresistance. This effect is illustrated in Figure 6-18.

Figure 6-18. Torsional response

A common instance of this problem is that

of the open-front building. The weaknesses ofopen-front designs have been discussed byDegenkolb(6-9):

Figure 6-19 shows the plans of three similarbuildings , each with three shear walls soarranged that there is an open end and thereforemajor torsions in the building. If the buildingsare similar, with uniform shear elements(uniform distribution of stiffness) andconsidering only shear deformations, it canrather simply be proved that the torsionaldeflection of the open end varies as the squareof the length of the building.


A common example of this condition occursin store front design, particularly on corner lots,and in free-standing commercial and industrialbuildings with varied openings around theperimeter. A special case is that of fire stationsthat require large doors for the movement ofequipment. In these buildings it is particularlyimportant to avoid major distortion of the frontopening, for example if the doors jam andcannot be opened, the fire station is out ofaction at a time when its equipment is mostneeded.

Tilt-up concrete industrial and warehousebuildings, in which lateral resistance isprovided by the perimeter walls, often alsorequire a variety of openings for entrances,loading docks, and office windows, with a

Figure 6-19. Open front design: torsional deflection varies as the square f the length


consequent variation in seismic resistancearound the perimeter.


A classical instance of this problemoccurred in the J.C.Penney Department Store inAnchorage, Alaska, in the 1964 earthquake.The building was so badly damaged that it hadto be demolished. The store was a five-storybuilding of reinforced-concrete construction.The exterior walls were a combination ofpoured-in-place concrete, concrete block, andprecast concrete nonstructural panels whichwere heavy, but unable to take large stresses.The first story had shear walls on all fourelevations. The upper stories, however, had astructurally open north wall, resulting in U-shaped shear wall bracing system (similar to atypical open-front store) which, when subjectedto east-west lateral forces, would result in largetorsional forces (Figure 6-20).

A special case is also that of apartmenthouse and hotels that are oriented to a view,such as a beach. which implies the need forlarge openings on the view elevation. The ElFaro building was a small apartment houselocated facing the beach in the Chilean resorttown of Vina del Mar. In order to exploit theview, two elevations are open: the stairs andelevator shaft are concentrated to the rear of thebuilding and their walls provide the seismicresistance. The result is a wide eccentricitybetween the centers of mass and resistance. Inthe Chilean earthquake of 1985, this buildingrotated and very nearly collapsed: it wassubsequently demolished. (Figure 6-21)


The objective of any solution to thisproblem is to reduce the possibility of torsion,and to balance the resistance around the

Figure 6-20. J.C.Penney department store, Anchorage, Alaska, 1964Note: unbalanced location of perimeter walls,particularly on third, forth and fifth floors, leading to severe torsional forces and near collapse.

300 Chapter 6

perimeter. Four alternative strategies can beemployed, and are shown in Figure 6-22.

Figure 6-21. El Faro apartments, Vina del Mar, Chile,1985

Figure 6-22. Solutions to open front buildings

The first approach is to design a framestructure with approximately equal strength andstiffness for the entire perimeter. The opaqueportions of the perimeter can be constructed ofnonstructural cladding material that will notaffect the seismic performance of the frame.This can be done either by using lightweightcladding, or by ensuring that heavy materials(such as concrete or masonry) are isolated fromthe frame.

A second approach is to increase thestiffness of the open facades by adding shearwalls at or near the open face. This solution is,of course, dependent on a design which permitsthis solution.

A third solution is to use a very strongmoment-resisting or braced frame at the openfront, which approaches the solid walls instiffness. The ability to do this will bedependent on the size of the facades: along steelframe can never approach a long concrete wallin stiffness. This is, however, a good solutionfor wood frame structures, such as apartmenthouses with a ground floor garage space,because even a rather long steel frame can bemade to approach plywood walls in stiffness.

Finally, the possibility of torsion may beaccepted and the structure designed to resist it.This solution will only apply to small structureswith stiff diaphragms, which can be designed toact as a unit.

6.6.3 Nonparallel Systems


The vertical load resisting elements are notparallel or symmetric about the majororthogonal axes of the lateral-force resistingsystem.


This condition results in a high probabilityof torsional forces under a ground motion ,because the centers of mass and resistancecannot coincide for all directions of groundmotion. Moreover, the narrower portions of the


building will tend to be more flexible than thewider ones, which will increase the tendency totorsion.

The problem is often exacerbated byperimeters with variations of strength andstiffness (Figure 6-23). A characteristic form ofthis condition is the triangular or wedge-shapedbuilding that results from street intersections atan acute angle. These forms often employ asolid, stiff party wall in combination with moreopen flexible facing the street. The result is aform that is very prone to torsion.

Figure 6-23. Wedge shaped plan: invitation to torsion


Non-rectiliner forms have becomeincreasingly fashionable in the last few years asa reaction against the rectangular "box". Formsthat are triangular, polygonal, or curved havebecome commonplace, even in very largebuildings. However, in some instances thedesired forms can be achieved by nonstructuralelements attached to a structure which may beessentially regular and rectilinear. (Figure 6-24)Extreme forms of non -rectilinearity are afeature of "deconstructionist" architecture,which is discussed in Section 6.11.

The traditional , trapezoidal or "flatiron"form resulting from the street-layout constraintsis still common in high-density urban locations.


This form has been fairly recently identifiedas a problem configuration. The form was notidentified as irregular in the 1890 SEAOC

Commentary , but it is identified as irregular inthe 1988 UBC, the 1990 SEAOC Commentary,and subsequent codes and provisions.

Figure 6-24. Form achieved by nonstructural attachmentsto main

Many buildings of this type wereconstructed in Mexico City, resulting from thehigh density and street layout of the city, andinstances of poor performance were observed inthe 1985 earthquake. Many buildings sufferedsevere distortion, particularly wedge-shapedbuildings with stiff party walls opposite theapex of the triangular form (Figure 6-25). Inmany cases the condition was exacerbated byother irregularities such as a soft story.


Since 1988 the UBC and the NEHRPProvisions place some special requirements onthe design of these types of configuration.Particular care must be exercised to reduce theeffects of torsion. In general, opaque wallsshould be designed as frames clad in

302 Chapter 6

lightweight materials, to reduce the stiffnessdiscrepancy between these walls and the rest ofthe structure.

Alternatively, special design solutions maybe introduced to increase the torsionalresistance of the narrow parts of the building,although this may be difficult to achieve whilestill retaining open facades or internal areas.

6.6.4 Diaphragm Configuration


The diaphragm configuration is the shapeand arrangement of horizontal resistanceelements that transfer forces between verticalresistance elements.


Diaphragms perform a crucial role indistributing forces to the vertical seismic-resisting elements. The diaphragm acts as ahorizontal beam, and its edges act as flanges.Diaphragm penetration and geometricalirregularities are analogous to suchirregularities in other building elements, leadingto torsion and stress concentration.

The size and location of these penetrationsis critical to the effectiveness of the diaphragm.The reason for this is not hard to see when thediaphragm is visualized as a beam: it is obviousthat openings cut in the tension flange of abeam will seriously weaken its load-carryingcapacity. In a vertical load system. a penetrationin a beam flange would occur in either a tensionor a compression area: in a lateral load system,the hole will be in a region of both tension andcompression, since the loading alternates indirection.

Figure 6-25. Distortion in wedge-shaped building, Mexico City, 1985


When diaphragms form part of a resistantsystem, they may act in either a flexible or stiffmanner. This depends partly on the size of thediaphragm (its area between enclosingresistance members or stiffening beams), andalso on its material. The flexibility of adiaphragm, relative to the shear walls whoseforces it transmits, also has a major influenceon the nature and magnitude of those forces.


Diaphragms are generally floors or roofs,and so have major architectural functions asidefrom their seismic role. The shape of thediaphragm is dependent on the overall planform of the building, and how it can besubdivided by walls or collectors.

In addition, however, architecturalrequirements such as staircases, elevators andduct shafts, skylights, and atria result in varietyof diaphragm penetrations. In some cases, as inthe need for elevators in an L-shaped building,the logical planning location for elevators (atthe hinge of the L) is also the area of greatestseismic stress.


Failures specifically due to diaphragmdesign are difficult to identify, but there isgeneral agreement that poor diaphragm layoutis a potential contributor to failure.


Diaphragm penetrations are a form ofirregularity specifically called out in the 1990SEAOC Commentary that requires engineeringjudgment. In addition, current codes andprovisions specifically define such penetrations,and impose some additional requirements onthe diaphragm design in such cases.

The general approach to the design ofpenetrations in diaphragms is to:

• Ensure that penetrations do not interfere withdiaphragm attachment to walls or frames.

• Ensure that multiple penetrations are spacedsufficiently far from one another to allowreinforcing elements to develop theirrequired capacity

• Ensure that collectors and drag struts areuninterrupted by openings

6.7 Vertical ConfigurationProblems

6.7.1 Soft and Weak Stories


A soft story is one that shows a significantdecrease in lateral stiffness from thatimmediately above. A weak story is one inwhich there is a significant reduction in strengthcompared to that above.


The condition may occur at any floor, but ismost critical when it occurs at the first story,because the forces are generally greatest at thislevel.

The essential characteristics of a weak orsoft first story consist of a discontinuity ofstrength or stiffness, which occurs at thesecond-story connections. This discontinuity iscaused because lesser strength, or increasedflexibility, in the first story structure results inextreme deflections in the first story, which, inturn, result in a concentration of forces at thesecond story connections.

If all the stories are approximately equal instrength and stiffness, the entire buildingdeflection under earthquake forces is distributedapproximately equally to each story. If the firststory is significantly less strong or moreflexible, a large portion of the total buildingdeflection tends to concentrate there, withconsequent concentration of forces at thesecond-story connections .(Figure 6-26)

304 Chapter 6

Figure 6-26. The soft-story effect

In more detail, the soft-story problem mayresult from four basic conditions. These arediagrammed in Figure 6-27 and are:

• A first-story structure significantly taller thanupper floors, resulting in less stiffness andmore deflection in the first story.

• An abrupt change of stiffness at the secondstory, though the story heights remainapproximately equal. This is causedprimarily by material choice: the use, forinstance, of heavy precast concrete elementsabove an open first story.

• The use of a discontinuous shear wall, inwhich shear forces are resisted by walls thatdo not continue to the foundations, but stopat second floor level, thus creating a similarcondition to that of the second item above.

• Discontinuous load paths, created by achange of vertical and horizontal structure atthe second story.

The above characteristics, individually or incombination are readily identifiable in existingbuildings provided that the building structurecan be studied in its entirety, either in the fieldor by reference to accurate as-built constructiondocuments.


A taller first story often has strongprogrammatic justification, when large spaces,such as meeting rooms or a banking hall, mustbe provided at ground level. Similarly, an openground floor often meets urban design needs byproviding both real and symbolic access to aplaza or street, or by providing space at the baseof a building. The changes in proportionprovided by a high story, or the "floating box"concept (now somewhat outdated), are very realaesthetic tools for the architect, althoughengineers may find such concepts hard torationalize in their terms.

Figure 6-27. Types of soft story


Engineers must accept that some form ofvariation in the first story will remain adesirable architectural characteristic for theforeseeable future: whether it is "soft" or"weak" in seismic terms is a matter for thearchitect and engineer to resolve.


The general type of soft first storyconfiguration was early identified as a problem.Failures in masonry buildings in the 1925 SantaBarbara earthquake were identified by Dewelland Willis(6-10) as soft-first-story failures.

In more recent times, with extensive use offrame structures, damage to reinforced-concretebuildings in Caracas (1967) clearly identifiedthe risk to tall buildings with this condition. Inthe Mexico City earthquake of 1985,researchers determined that soft first storieswere a major contributor to 8% of seriousfailures, and the actual percentage is probablygreater because many of the total collapseswere precipitated by this condition.

The particular case of the discontinuousshear wall has led to clearly diagnosed failuresin United States buildings. Olive View hospital,a new structure that was badly damaged in the1971 San Fernando earthquake, represents aclassic case of the problem.

The vertical configuration of the mainbuilding was a two-story layer of rigid frameson which was supported a four-story shear wall-frame structure (Figure 6-28). The second floorextended out to form a large plaza.

Figure 6-28. Olive View hospital, San Fernando, 1971 (a)elevation of stair towers (b) section through main building

The severe damage occurred in the soft-story portion: the upper floors moved so muchas a unit that the columns at ground level couldnot accommodate such a huge displacementbetween their bases and tops and failed. Thelargest amount by which a column was leftpermanently out of plumb was 2 1/2 feet.

Though not widely identified, the stairtowers at Olive View also show a clear andseparate example of a discontinuous shear-wallfailure. These seven-story towers wereindependent structures, and proved incapable ofstanding up on their own: three stair towersoverturned completely, while the fourth leanedoutwards 10 degrees. The six upper stories wererigid reinforced -concrete walls, but the bottomstory was composed of six free-standingreinforced-concrete frames, which failed. Theexception was the north tower, whose wallscame down to the foundation directly withoutany discontinuity; this was the only tower toremain standing. Olive View hospital wasdemolished after the earthquake, and a newhospital built on the same site.

The performance of the Imperial CountyServices Building, El Centro, in the ImperialValley Earthquake of 1979, provides anotherexample of the effects of architecturalcharacteristics on seismic resistance. Thebuilding was a reinforced-concrete structurebuilt in 1969. In this mild earthquake thebuilding suffered a major structural failure,resulting in column fracture and shortening (bycompression) at one end-the east-of thebuilding. (Figure 6-29). The origin of thisfailure lies in the discontinuous shear wall atthat end of the building.

The fact that this failure originated in theconfiguration is made clear by the architecturaldifference between the east and west ends: thisis an example of the large effect on seismicperformance of a relatively small designvariation between the two ends of the building..The difference in location of the small ground-floor shear walls was sufficient to create amajor difference in response to the rotationalforces on the large end shear walls (Figure 6-30).

306 Chapter 6

Figure 6-29. Imperial County Services Building, ElCentro, California. failure of end bay at discontinuous

shear wall, (Imperial valley earthquake of 1979)

Figure 6-30. Imperial County Service Building, plan andelevations

A more recent instance is that of a medicaloffice building in the Northridge earthquake of1994, constructed at about the same time as theprevious two buildings discussed. The simplerectangular building had discontinuous shearwalls at each end. These proved inadequate todeal with the forces, with consequent severetorsional damage at each end of the building,(Figure 6-31) This building also had a structuraldiscontinuity at the second floor that caused the"pancaking" of the second floor.


If a high first story is desired, either:

• Introduce bracing that stiffens the columnsup to a level comparable to thesuperstructure.

• Add columns at the first story to increasestiffness, or

• Change the design of the first-story columnsto increase stiffness.

If a large opaque wall is required in alocation that could create a soft first story:

• Insure that such a wall is not part of thelateral load resisting system

• Reduce the mass of the wall by use of lightmaterial and hollow construction

• If a heavy wall is necessary, then insure thatthe wall is detached in such a way that thesuperstructure is free to deflect in acomparable way to the first floor

If the architect insists on such material anddesign constraints that a major discontinuousshear wall is the only solution, the engineershould refuse to do it. The liabilities involved inusing such a proven failure mechanism are toogreat.

If the lateral resistance system is based onthe use of an interior core (for a high-rise officebuilding, for example), the perimeter columnsmay be tall, but there is no soft first story,provide the core is brought down to the ground.In such a building it is not difficult, if the core-plan dimensions are sufficient, to insure that thestiffness of a tall first story is adequate toprevent structural discontinuity at the secondfloor. One condominium building a goodexample of architect- engineer collaboration.That building achieved an elegant exteriorappearance which appeared to be a soft firstfloor. However, the seismic resistance wasprovided by a strong interior box shear wallstructure that enabled the taller first floor to beaccommodated with ease. The building suffered


virtually no damage in the strong Chileanearthquake of 1985.

It should be noted that in the 1997 NEHRPProvisions structures with a weak-storydiscontinuity in capacity that is less than 65%of the story above are not permitted over 2stories or 30 feet in height in Seismic DesignCategories B,C,D,E and F.

6.7.2 Columns: Variations in Stiffness,Short Columns, and WeakColumn/Strong Beam.


This section considers the use of columns ofvarying stiffness, by reason of either differencesin length or deliberate or inadvertent bracing:the use of columns that are significantly weakerthan connecting beams: and the use of columns

in one floor that are significantly shorter thanthose on other floors.


Seismic forces are distributed in proportionto the stiffness of the resisting members. Hence,if the stiffness of the supporting columns (orwalls) varies, those that are stiffer (usuallyshorter) will "attract" the most forces. Theeffect of this phenomenon is explained inFigure 6-32. The important point is thatstiffness (and hence forces) variesapproximately as the cube of the column length.

Similarly, a uniform arrangement of shortcolumns supporting a floor will attract greaterforces to that floor, with a correspondingpossibility of failure. Typically such anarrangement may also involve deep and stiffspandrel beams, making the columnssignificantly weaker than the beams.

Figure 6-31. Discontinuous shear wall failure, office building, Northridge

308 Chapter 6

Figure 6-32. Effect of variations of column stiffness

Such a design is in conflict with a basicprinciple of seismic design, which is to design astructure in such a way that under severeseismic forces, beams will deform plasticallybefore columns. This is based on the reasoningthat as beams progress from elastic to inelasticbehavior they start to deform permanently. Thisaction will dissipate and absorb some of theseismic energy. Conversely, if the column failsfirst and begins to deform and buckle, majorvertical compressive loads may quickly lead tototal collapse.

Mixing of columns of varying stiffness ondifferent facades may also lead to torsionaleffects, since the building assumes the attributesof varying perimeter resistance discussedabove.


The origin of variations in column stiffnessgenerally lies in architectural considerations.Hillside sites, infilling of portions of frameswith nonstructural but stiff material to createhigh strip windows, desire to raise a portion ofthe building of the ground on tall "pilotis",while leaving other areas on shorter columns, orstiffening some columns with a mezzanine or a

loft, while leaving others at their full, unbracedheight.

These issues are important because theireffects may be counterintuitive. For example,infilling may be done as a remodel activity laterin the building life for which the engineer is notconsulted, because intuition may suggest to thedesigner that he is strengthening it in the act ofshortening it rather than introducing a seriousstress concentration for which the structure wasnot designed. For vertical forces a reduction inthe effective length of a column is beneficialbecause it reduces the likelihood of buckling,but the effect under lateral forces is quitedifferent.

Variations in openings in different facadesare often required from a daylighting or energy-conservation requirement. Where openings arecreated by variations in structural arrangement,rather than by variations in cladding, some ofthese conditions may well arise.


Significant column failures, sometimesleading to collapse, have been attributed tothese conditions in a number of recentearthquakes, particularly in Japan, LatinAmerica, and Algeria.

Many Japanese schools, employing shortcolumns on one side of an elevation, or using aweak- column, strong-beam configuration,suffered severe damage in the Tokaichi-okiearthquake in 1968 and the 1978 Miyagi-ken-oki earthquake. (Figure 6-33)

In Latin America, the problem hasfrequently been caused by inadvertent stiffeningof columns through nonstructural infill which,when combined with high glazing, creates shortcolumns.

In the El Asnam (Algeria) earthquake of1980, many apartment structure failures werecaused by short columns used at ground level toprovide a ventilated open space (called a "videsanitaire") in a semi-basement location . Thesignificant failure of a large condominium andhotel structure in the Guam earthquake of 1993has been ascribed in part to the creation of a


short column condition by the introduction ofnonstructural stiffening elements(6-11) (Figure 6-34)


The general solution is to match the detailedseismic design carefully to the architecturalrequirements. The weak-column, strong-beamcondition can be avoided by insuring that deepspandrels are isolated from the columns; in thesame way the lengths of columns around afacade can be kept approximately equal.

Horizontal bracing can be inserted toequalize the stiffness of a set of columns ofvarying height (Figure 6-35). Heavynonstructural walls must be isolated fromcolumns to insure that a short-column conditionis not created. (Figure 6-36).

6.7.3 Vertical Setbacks


A vertical setback is a horizontal, or nearhorizontal, offset in the plane of an exteriorfacade of a structure.


The problem with this shape lies in thegeneral problem of discontinuity: the abruptchange of strength and stiffness. In the case ofthis complex configuration, it is most likely tooccur at the line of the setback, or "notch".

The seriousness of the setback effectdepends on the relative proportions andabsolute size of the separate parts of thebuilding. In addition, the symmetry or

Figure 6-33. Short column failure, school, Japan: Miyagi-ken -oki, 1978

310 Chapter 6

Figure 6-34. Short column failure, Guam, 1993

Figure 6-35. Horizontal bracing to stiffen a high open end entrance


B u ild in g S tru c tu re

G ap

S tee l A n g les

C o n cre te B lo ck

Figure 6-36. Heavy nonstructural wall isolated fromstructure at top and side

asymmetry in plan of the tower and base affectthe nature of the forces. If the tower or base orboth are dynamically asymmetrical, thentorsional forces will be introduced into thestructure, resulting in great complexity ofanalysis and behavior.

The setback configuration can also bevisualized as a vertical reentrant corner.Stresses must go around a corner , because anotch has been cut out, preventing a more directroute. Hence, the smaller the steps or notches ina setback, the smaller the problem. A smoothtaper avoids the notch problem altogether. Atapering beam will not experience stressconcentrations, whereas a notched beam will.

Setbacks with shear walls in the towerportion that are not continued to the ground arehighly undesirable. Besides the change ofstiffness where the shear wall enters the basestructure, the shear wall will transmit largeforces to the top diaphragm of the base.

Although, typically, setbacks occur in asingle building, the condition can also becreated by adjoining buildings of differentheights which have inadequate or nonexistentseismic separations.


Setbacks may be introduced for severalreasons. The three most common are zoningrequirements that require upper floors to be setback to admit light and air to adjoining sites,program requirements that require smaller

floors at the upper levels, or stylisticrequirements relating to building form.

Setbacks relating to zoning were common afew decades ago when daylighting was a majorconcern, and resulted in characteristic shapes ofolder high-rise buildings in New York and otherlarge cities. Stylistic fashions replaced theseforms with those of simple rectangular solids,made possible by advances in artificial lightingand air-conditioning. Now, there is a renewedinterest in set-back shapes for stylistic reasons,while at the same time energy conservationrequirements have reinstated a functionalinterest in setbacks for daylighting reasons.

An interesting example of this stylistic trendis that of the new planning code for SanFrancisco, which specifically mandatessetbacks for large buildings in the downtownarea. These represent relatively minorvariations in the vertical plane of the facade,rather than the abrupt rising tower on a base,which is of more serious seismic consequence.The trend is, however, away from verticalstructural continuity at the perimeter and thusintroduces complexity and cost into thestructural solution.

A type of setback configuration only madepossible by modern framed construction is thatof the building that grows larger with height.This type is termed inverted setback or invertedpyramid depending on its form. Its geometricaldefinition is the same as that of the setback, but,because of the problems of overturning, itsextremes of shape are less. Nevertheless. somesurprising demonstrations of this shape haveappeared, and it appears to be one whose imagehas a powerful design appeal (Figure 6-37).


Although commonly identified as aconfiguration problem, severe failures ofmodern buildings attributed to this conditionare few. While traditional towers, primarilychurches, have suffered their share of failures,the number of those that have survived severe

312

damage is remarkable. An example from theKobe earthquake of 1995 shows a failure in asetback building at the plane of weaknesscreated by a combination of the setbacks andadjoining openings in the wall (Figure 6-38)

While there have been recorded failures ofinverted-setback buildings, notably in theAgadir (Morocco) earthquake of 1960, some ofthe more striking examples have performedwell. This is probably because the appearanceof instability inherent in this form results inspecial attention being paid to its structuraldesign. Typically, such buildings devote amuch larger percentage of their constructioncost to structure than more conventionalbuildings.


Setbacks have long been recognized as aproblem, and so the Uniform Building Code hasattempted to mandate special provisions forthem currently, the earthquake regulations ofthe Code refer to setback configurations asfollows:

Buildings having setbacks wherein the plandimensions of the tower in each direction isat least 75% of the corresponding plandimension of the lower part may beconsidered as uniform buildings withoutsetbacks, provided other irregularities asdefined in this section do not exist.

An appendix to the 1990 SEAOCCommentary to this section includes a lengthydiscussion of the setback problem and anapproach to its analysis .:

In general, conceptual solutions to thesetback problem are analogous to those for itshorizontal counterpart, the reentrant cornerplan. The first type of solution consists of acomplete seismic separation in plan, so thatportions of the building are free to reactindependently. For this solution, the guidelinesfor seismic separation, discussed elsewhere,should be followed.

Figure 6-37. Dallas City Hall : an inverted pyramid

313

When the building is not separated, theanalysis proposed in the appendix to the 1990SEAOC Commentary provides the bestguidelines, with some necessary interpretationsto fit the particular case. Particular attentionshould be paid to avoiding vertical columndiscontinuity, so that setbacks should bearranged to coincide with normal bay sizes(which may result in a series of small bays).

Any large building with major setbackconditions should be subjected to specialanalysis, or at least to careful investigation ofprobable dynamic behavior. Finally, theinverted setback configuration of any extremeform and size should be avoided in seismicareas, unless the owner is willing to assume theconsiderable additional structural costs that willbe incurred.

The 1997 NEHRP Provisions, as notedearlier, permit vertical setback configurations tobe analyzed using the simple ELF method if thestiffness on the top story of the base is at least

five times that of the first story of the tower.The UBC permits use of the standard ELFmethod for a two-stage analysis of tower andbase if the average story stiffness of the base isat least 10 times greater than the average storystiffness of the tower.

6.8 STRUCTURALLYRESTRICTIVEARCHITECTURALDETAILING

6.8.1 Components and Connections


By structurally restrictive detailing wemean detailed architectural design of acomponent that prevents good seismic designpractice in the structural design.

Figure 6-38. Failure of set-back building along a plane of weakness

314 Chapter 6


This problem represents a micro version oftypical overall building configuration problems.Architectural detailing may place dimensionalor location constraints on structural designresulting in weakness or eccentricity of forceactions that can lead to stress concentration orlocal torsion. The problem is most critical atbeam-column connections, which are highlystressed, but often represent a critical elementin the aesthetic scheme of the building.

Structural detailing ideally provides fordirect load transfer and minimum localeccentricity, with forces resolved at a point.Architectural detailing may result in inadequatesize and eccentric or discontinuous load paths(Figure 6-39). The problem is particularlycritical for reinforced-concrete structures,where constraints may provide inadequate roomfor proper placing of reinforcing.

Figure 6-39. Eccentric load paths created by architecturaldetailing of structural connection


Detailed design is an important element inarchitectural expression. As an example, thedesign of the perimeter beam-conlumnconnection can provide the building with apredominantly horizontal, vertical, or neutralemphasis. (Figure 6-40). But the structuralimplications of these variations may not beunderstood by the architect. Another example isthe use of taper or the insertion of recesses incolumns. Tapered columns may be a correct

expression of structural forces, and be easy toaccommodate, or they may directly contradictstructural action and lead to weakness. Recessesare often designed by architects to accentuatethe line at which materials meet one another,particularly when the materials are different ormeet at right angles, as in a column-slabjunction.

Figure 6-40. Facades: differences in architecturalemphasis

•••• Historical Performance:

Specific performance attributable to thiscondition is difficult to document but theproblem is generally recognized by engineers.Two well documented cases do exist wherearchitectural detailing contributed to failure.

The first is that of the column design of theOlive View Hospital, damaged in the 1971 SanFernando earthquake (discussed previously asan example of soft-first-story failure.). Asignificant difference in performance wasobserved between corner and internal columnsin this building. The twelve L-shaped cornercolumns were completely shattered and theirload-carrying capacity reduced almost to zero.The interior columns, of square section, hadspiral ties, and although they lost most of theirconcrete cover, they retained load-carryingcapacity and probably saved the building fromcollapse. Because of their architectural form, itwas not possible for the corner columns to usespiral ties (Figure 6-41). Higher stress andtorsion in the corner columns may also havecontributed to their poor performance.


Figure 6-41. Exterior column sections at Olive ViewHospital, San Fernando, California. Due to their shape,

corner columns could not be spirally reinforced

The Imperial County Service Building at ElCentro, California, suffered severe damage inthe 1979 Imperial Valley earthquake, and fourcolumns at one end of the building were badlyshattered. Detailed study of these columnsshowed that an architectural recess had beenplaced at the line where the columns met theground. (Figure 6-42). This recess caused areduction in sectional area of the column and areduction in axial load-carrying capacity.Analytical and experimental studies haveshown that this change in column sectionaccentuated the undesirable performance ofthese columns(6-7).

Figure 6-42. Column detail, Imperial County ServicesBuilding, El Centro, California. Note architectural recess

affecting reinforcement continuity

•••• Solutions:

Close coordination between architect andengineer is necessary to insure that architecturaldetailing does not result in undesirablestructural design constraints.

6.9 PROBLEMS OFADJACENCY

6.9.1 Pounding

•••• Definition:

Pounding is damage caused by twobuildings, or different parts of a building,hitting one another.

•••• Seismic Effects:

Pounding as characterized in Codes andGuidelines and in most analytical researchstudies takes the form of in plane displacementsof two adjacent buildings, as in theinvestigation of a row of adjacent buildings byAthanassiadou et al(6-12). Empirical observationshows that building separations are complex intheir basic conditions and in their effects, andlack of separation is not necessarily detrimental.

Observation has shown that the endbuildings of a row of adjacent buildings tend tosuffer more damage than interior buildings.Analytical pounding studies consider regularbuildings in elevation. In fact, the swaycharacteristics of buildings are much influencedby irregularities, particular that of soft firststories, that can lead to extreme displacementsor even collapse. Some of these characteristicsare shown in Figure 6-43. Similarly, analyticalstudies have always assumed regular buildingsin plan. Since adjacent buildings with little orno separation will generally be found in theolder sections of down town, building plans areoften very irregular, leading to torsional effectsunder ground motion. These characteristics areshown in Figure 6-44.

316 Chapter 6

Study of pounding damage in Kobe in 1995showed that very large deflections were oftencaused by design flaws (such as a soft firststory) or near source extreme shakingvelocities, or a combination of the two. Inaddition, many instances of large buildingdeflections (or "leaning") related toground/foundation failures. These effects arenot accounted for in code type separationrequirements, which assume a uniformdeflection for the height of the building, relatedonly to ground motion.

Observation has also shown that, in somecases, the close proximity of buildings may actas a support, particularly for buildings in mid-block, and increasing the space betweenbuildings might serve, in some cases, toincrease deflections and damage rather thanreducing them. A probable instance of thiswas observed by the author in Mexico City, in1985. In this instance, a tall slender buildingwith an apparent serious soft first storyproblem, appeared to be restrained by low, stiffbuildings on either side. (Figure 3 ). Severalinstances of this phenomenon were observed inKobe.

This point is very difficult to assess. Theresponse to shaking of a number of adjacentbuildings with essentially no separationbetween them must be equivalent to theresponse of a large building with a variety ofstrengths, stiffness and other structuralcharacteristics which would be very difficult toanalyze.

The possibility of pounding is a function ofthe vertical deflection or drift of adjoiningbuildings (or parts of a building). Drift iscalculated by applying the code design forces tothe building and then observing the deflectionsthat result. Since these estimated forces will beless than what we know can occur, calculateddeflections must be corrected to obtain a morerealistic estimate of how much the building mayactually move. Alternatively, an accurateestimate of drift may be made that accounts forall foreseeable factors.

Potential pounding presents some particularproblems of a socio-economic nature whereexisting buildings are concerned. The socio-economic problems consist of how to involvethe adjoining building owner in possibly costlystudies, design and construction work that theowner may not wish to participate in or may

Figure 6-43. Irregularities may create extreme displacements or collapse


even actively oppose them. The problems areparticularly critical in the case of commonstructure, because rehabilitation is verydifficult, if not impossible, without theneighbor's involvement and probably somedegree of rehabilitation to his property. In thecase of falling hazards, it would be desirable forthe neighbor to mitigate them, but the extent towhich the federal owner can require this are notclear. The problem of pounding is traditionallydealt with by requiring a large gap betweenbuildings. This can, in theory, be achievedwithout impacting the adjoining owner.

While in engineering terms it may seemobvious that it is in the adjoining owners bestinterest to cooperate in evaluation andmitigation, in socio-economic terms there maybe many reasons, valid or otherwise, forreluctance. The owner may have real economicconstraints in incurring any costs of evaluationor mitigation, and be quite ready to accept thepossible risks of inaction. In addition, the ownermay have short term intentions of redevelopingor selling his property, and so not wish to incur

expenses that will be of no conceivable benefitto him. Thus, the possibility of cooperativerehabilitation will be much conditioned by howthe adjoining owner sees his economic futureand views unsolicited action by a neighbor thatmight impact it.


Pounding is included in this discussion ofconfiguration issues because it is a matter ofwhere buildings are located relative to otherstructures, which is an early architecturaldecision. The problem has considerablearchitectural implications for the constructionof buildings on constricted urban sites, becauseto make provision for the worst case conditioncould result in large building separations andsignificant loss of usable space.

While building codes place modest limits ondrift (for example, 0.005 time story height)based on static analysis, actual experience withdrift and calculations of realistic figures providesome startling numbers. Freeman(6-13) calculated

Figure 6-44. Irregularities in plan may create additional torsional effects that impact adjoining buildings

318 Chapter 6

the actual drift on flexible buildings up to 20stories under 0.4 g acceleration as being 0.020 -0.055 times the story height. For a 12 -storybuilding this translates into 40-110 in. for a 14-ft story height. A separation that couldaccommodate two such buildings vibrating outof phase would have to be 18 ft. 4 in. wide.

Clearly compromise is necessary, butnonetheless, loss of usable space measured inlineal feet becomes serious. In addition, the ideaof urban buildings with spaces of 2 - 3 feetbetween them suggests a very difficultmaintenance problem.

•••• Historical Performance:

Problems of adjacency have been routinelynoted by earthquake investigators over the pastseveral decades. In the 1972 Managuaearthquake, the five-story Grant Hotel suffereda complete collapse of its third floor whenbattered by the roof level of the adjacent two-story building.

In the 1964 Alaska earthquake, the 14-storyAnchorage Westward Hotel pounded against itslow rise ball room and an adjoining six-storywing, although separated by a 4-in, gap. Thepounding was severe enough in the high rise todislocate some of the metal floor decking fromits steel supports.

In recent earthquakes, pounding hascontinued to be a serious issue. The earthquakethat struck Mexico City in 1985 has revealedthe fact that pounding was present in over 40%of 330 collapsed or severely damaged buildingssurveyed, and in 15% of all cases it led tocollapse. Many instances of pounding wereobserved in the Kobe earthquake of 1995.

•••• Solutions:

Perhaps due to the high incidence ofpounding damage observed in the 1985 MexicoCity earthquake a number of researchers havestudied pounding problems in recent years. Tworecent studies, by Jeng et al,(6-14) and the studyby Athanassiadou et al,(6-12) are representative,and both contain a full set of references to other

studies of the problem. Jeng et al. present a newmethod for estimating the likely minimumbuilding separation necessary to precludeseismic pounding: two 10 story concrete framebuildings are analyzed by way of example.

Athanassiadou et al. studied the seismicresponse of adjacent buildings in series, withsimilar or different dynamic characteristics,using SDOF systems subjected to base motions.

These, and other studies, confirm the resultsof empirical surveys, and to providequantitative information that is necessary forcode and design practice development, althoughas yet the quantitative data is not readilytransferable to code values.

To assume that code limits on drift providean accurate estimate of possible drift isunrealistic, but accurate estimates may providevery large worst case figures. Blume, Corningand Newmark suggest an alternative method(6-

15):

Compute the required separation as the sumof the deflections computed for each buildingseparately on the basis of an increment indeflection for each story equal to the yield-point deflection for that story, arbitrarilyincreasing the yield deflections of the twolowest stories by multiplying them by afactor of 2.

An earlier edition of the Uniform BuildingCode contained a rule of thumb intended for therelatively stiff structures of that day(6-16):separations should be " one inch plus one halfinch for each ten feet of height above twentyfeet".

It should be noted that, notwithstanding thehigh cost of land in Japanese cities, newstructures in Kobe seem to be providing agenerous allowance for differential drift.

A possible alternative approach is to placean energy-absorbing material between thebuildings; this obvious simple approach seemsto have been little studied.

Many buildings in Mexico City were, infact, protected from collapse because they wereerected hard up against adjoining buildings on


both sides, so that whole blocks of buildingsacted as a unit, and the group was stronger thanthe individual structures. As evidence of this,Mexican studies showed that 42% of severelydamaged buildings were corner buildings,lacking the protection of adjoining structures.This finding suggests the need for seriousresearch on the subject of allowable drift,pounding, and the design and construction ofclosely spaced buildings.

6.9.2 Other Adjacency Problems

Two other problems of adjacency give causefor concern: one is that of damage caused to abuilding by falling portions of an adjoiningbuilding: in the 1989 Loma Prieta earthquake adeath was caused in downtown Santa Cruzwhen a portion of unreinforced masonry wallfell through the roof of a lower adjoiningbuilding, and six deaths were caused in SanFrancisco when part of a masonry wall fell onsome parked cars. The other adjacency problemis that created by structural elements - generallywalls or columns - that are common toadjoining buildings: while instances of damagecaused by this condition are not specificallyidentified, there is a clear problem when anowner wishes to rehabilitate a building whichhas structural elements common to an adjoiningbuilding that is not undergoing relatedrehabilitation.

6.10 THEARCHITECT/ENGINEERRELATIONSHIP

6.10.1 Architect-Engineer Interaction

In the United States the architect/engineerrelationship is delicate because typically theengineer is employed by the architect, and if hecomplains too much about the architect's designhe may be replaced. An architect who finds hisdesign criticized by his engineer can generallyfind an alternative engineer who will

accommodate him. It is extremely hard toascertain whether this second engineer reachesthis accommodation because he is moreignorant than his colleague, more of a gambler,or more inventive and clever.

There are, of course, many instances wherearchitects and engineers have built up closerelationships and communicate fruitfully, withthe engineer participating at an early stage ofdesign. However, even in these instances thepressure of business often means that, forfinancial reasons, the engineer is not employeduntil the building schematic design is complete.This applies particularly to private work, wherethe developer must have a design- perhaps onlya three dimensional sketch -in order to procurefinancing, and he does not want to incuradditional consultant costs until the financing issecured.

The following description is of thepreliminary design process of a large U.S.architectural for a client in the Pacific Rim:

".. we developed a method whereby wewould send a team of three people for aweek, working in the client's office, or froma hotel room, but having client input intodaily charettes, lots of alternatives in sketchform, not spending many hours ofpresentation, but spending the hours ondesign. At the end of the week we wouldgenerally have a viable concept that theclient had signed off."(6-17)

Thus the schematic design for a multi-million dollar project is completed in a week:presumably the design is then brought to theengineer for him to insert a structure.

Obviously, in this instance, much dependson the knowledge and experience of this threeperson team to ensure that the design isstructurally reasonable. More risky is whenanalogous processes are conducted by a singlearchitect with a desire to produce a design thatwill amaze his client.

In seeking improved architect/engineerinteraction a number of conditions must apply:

320 Chapter 6

• The engineer must communicate directlywith the architectural design person or team

• The architect must take seriously his sharedresponsibility with the engineer for theseismic performance of the building. Recentexperiences, such as Northridge and Kobe,should encourage this attitude.

• Mutual respect and cooperation: anadversarial relationship will not beproductive.

• Common language and understanding:

The architect must have someunderstanding of seismic engineeringterms -such as acceleration, amplification,base shear, brittle failure, damping (and soon through the engineering glossary). Atthe same time the architect should have ageneral understanding of thecharacteristics of typical seismic structuralconcepts: shear walls, bracing, momentframes, diaphragms, base-isolation etc.The new concepts of performance -basedseismic design should also be understood.In turn the engineer must understand thearchitect's functional needs andaspirations.

• Collaboration must occur at the onset of aproject: before architectural concepts aredeveloped or very early on in theirconception

• Business conditions that restrict earlyarchitect/engineering interaction must bealleviated (by the use of a general consultingretainer fee, for example, recovered fromthose projects that are achieved).

• If the architect does not want to interact withhis engineer, or if for some reason isprevented from doing so, then he shouldwork with simple regular forms, close to theoptimal seismic design

While it is reasonable for engineers to askthat architects become better informed aboutseismic design and the consequences of theirconfiguration decisions, the engineer mustunderstand that while for them seismic design is

of paramount importance, for architects andtheir clients it takes very low priority as far astheir own interests. For the architect, seismicdesign and safety is taken care of by theengineer: it is no more a subject of concern thanprovision for vertical forces, which nevercomes up for discussion between owner andarchitect, and seldom between architect andengineer. The architect is preoccupied withissues of codes and regulations relating toplanning and design far removed from seismicproblems, but of great importance and interestto his client. Similarly, the architect iscontinuously evaluating planning options,materials issues and both functional andaesthetic concerns upon which his client isconstantly questioning him. Above all, the workmust be done on time and on budget, and thearchitect would also like the job to beprofitable.

Architects vary greatly in their interests: thestereotype of the architect as an unworldlyaesthete is seldom true. Some architects arebrilliant salespeople and business managers:some are very close to engineers, and interestedin how the building is engineered andconstructed: some are excellent projectmanagers and will ensure that budgets andschedules are kept: some are inspiringmanagers of people and will run an exciting andenjoyable office: some are brilliant at the designof details, the behavior of materials and thedevelopment of construction documents: andsome are thoughtful and inventive designers.The large, well-run office will have a mix of theabove in its staff. The small office must try andfind a few people that combine the above roles.As the profession of architecture becomes morecomplex, specialization is becoming morecommon: even large firms cannot play all roles,and the small office must specialize in a limitedtype of design. The advent of CAD and otherinformation systems has extended the range forthe small practitioner, but these systems needlarge capital investments that produce their ownforms of limitation.


6.11 Future Images

6.11.1 Beyond the International Style

The tenets of the International Style beganto be seriously questioned in the mid-1970's,both in print by architectural critics andhistorians and in practice by architectsbeginning to bring new design approaches tothe drawing board and to construction. Thisquestioning finally bore fruit in an architecturalstyle known broadly as "Post-modern".Although this term was criticized by critics andthe architects who were seen to be designing inthis style, the term became a useful mark ofidentity. In general, post-modernism meant:

• the revival of surface decoration on buildings• a return to symmetry in overall form• the use of classical forms, such as arches,

decorative columns, pitched roofs, innonstructural ways, and generally insimplified variations of the original elements.

• a revival of exterior color as an element, witha palette of characteristic colors (e.g. darkgreen, pink, Chinese red, bright yellow, buffetc)

Developments of post-modernism alsoinvolved both the revival of full, scholarly,classical revival as a style., and also verypersonal images by a few prominent architectsin terms of scale and forms, which were derivedfrom a variety of sources, such as Victorianengineering, ancient Egyptian architecture andnon-Euclidean geometry.

In seismic terms, this change in stylisticacceptance was, if anything, beneficial. Thereturn to classical forms and symmetry washelpful to the structure as a whole, and almostall of the decorative elements werenonstructural. Inspection of an early icon ofpost-modernism, the Portland office buildingdesigned by Michael Graves, (Figure 6-45).shows an extremely simple and ordinarystructure. Indeed, the Portland building, whichcreated a sensation when completed, has a form

that approximates our optimal structure: thesensation is all in the nonstructural surfacetreatments. Designed as an economicaldesign/build project the building has recentlyundergone seismic retrofit unrelated to itsconfigurational characteristics.

Figure 6-45. Office building, Portland, Oregon. MichaelGraves, architect, 1979

It should be noted, however, that an interestin seismic design had no influence on thedevelopment of post-modernism - it is, and wasa strictly aesthetic and cultural movement.

At the same time that post-modernism wasmaking historical architectural style legitimateagain, another style evolved in parallel:. Thisstyle, originally christened "hi-tech" (the termhas not stuck) returned to the celebration ofengineering and new industrial techniques andmaterials as the stuff of architecture. This styledeveloped primarily in Europe, notably inEngland and France, and was exemplified in afew seminal works, such as the PompidouCenter in Paris, the Lloyds building in London,and the Hong Kong and Shanghai bank in HongKong. These buildings proclaimed a newversion of the functionalism of the thirties,updated to provide flexibility, adaptability andadvanced servicing for an uncertain future,using exposed structure with beautiful castingsas connections. In truth, these buildings are as

322 Chapter 6

aesthetically and stylistically conceived as anypost-modern or classical revival building.

The rise of post-modernism releasedarchitects from the strait-jacketed moralities ofthe International Style. As a result, at present akind of aesthetic bedlam reigns, and severalcompeting private styles co-exist, competingfor clients - and finding them. The leadingexponents of the new styles form anarchitectural jet-set, cruising the worlddropping off their stylistic gems to clients andcountries that can afford them.

The importance of well-publicized designsby fashionable architects is that they create newaccepted styles. Architects are very responsiveto form and design and once a form gainscredence practicing architects the world overbegin to reproduce it. Today's New Yorkcorporate headquarters high-rise becomestomorrow's suburban Savings and Loan Office,as became clear in the adoption of the metal andglass curtain for building exteriors. The firsttwo highly publicized curtain walls were that ofthe United Nations building and the LeverBrothers building, both in New York city in theearly 50's: by the mid 60's every town inAmerica had its stock of blue-green glazedcommercial buildings.

So, to predict the design vernacular of thefuture it is necessary to look at what is beingdone in high-style architecture, and inparticular, to try and guess which forms seem tocatching the imagination of architects andstarting to be reproduced at a more modest level. Amid the bedlam of design voices, threeinfluential trends can be discerned..

6.11.2 Influential Trends

•••• The bridge building:

The bridge building form is that of twinhigh-rise buildings connected at the roof withhorizontal occupied space that acts as a bridge.The concept is that of a single building. Theprototypical form of this, that has seizedarchitect's imaginations, is that of the GrandArch of the Defense (Figure 6-46), in Paris, one

of the late President Mitterand's "grandprojects". This is a single office building, some34 stories tall, designed as a cubical arch,framing the end of the Defense development onthe perimeter of Paris. The arch is in line withthe main axis through Paris to the Louvre, onwhich lies the Arch De Triomphe. Thehorizontal bridge structure provides exhibitionand meeting spaces.

Figure 6-46. Grande Arche of the Defense, Paris, JohanOtto von Spreckelsen, architect

A similar form is that of the Umeda SkyBuilding (Figure 6-47) in Osaka, Japan. Thisbuilding incorporates a mid-air garden , midairescalators and a mid-air bridge to connect thetwo parts of the building. The architect, HiroshiHara, sees this form as the beginning of anapproach to a three-dimensional network to ourcongested cities. This building is in a fairlysevere seismic zone and is carefully designedfor earthquake resistance.


Figure 6-47. Umeda Sky building, Osaka, Japan, HiroshiHara, architect, 1988

The bridge or twin tower forms haveimmense drama and appeal, and so we canexpect to see five story versions of themappearing in our shopping malls and suburbancenters.

•••• The warped building:

A strong design trend is that of buildingsthat use warped forms, often combined withnon vertical walls and irregular warped exteriorsurfaces. The most prominent exponent of theseforms is the American architect Frank Gehry,who is now building these forms all over theworld. His Guggenhein Museum in Bilbao,Spain, completed in 1997 is typical of his style,and has been hailed as a masterpiece byarchitectural critics world-wide (Figure 6-48).His tower for the Rapid Transport Headquarters

in Los Angeles, (Figure 6-49) shows his warpedand non vertical forms applied to a skyscraper.Despite its flourishes, the building is essentiallyrectilinear with the warped elements achievedby nonstructural add-ons to the main structure.

•••• The Deconstructed Building

Deconstruction is a term applied to the workof a number of architects presently workingaround the world: the term is derived from thelanguage and literary movement of the samename that originated in literary criticism. Theprinciples of deconstruction were firstformulated by the French philosopher and criticJacques Derrida, in the early 1970's and havesince revolutionized literary criticism and thestudy of language and meaning.

Because deconstructed buildings essentiallyignore the limitations of constructability, fewhave yet been built. One of the architects mostcommonly associated with deconstruction is theIraqi, Zaha Hadid, who works in London.Figure 6-50 shows her design for a normallyprosaic building - a fire station completed in1993 in Germany.

6.12 CONCLUSION

These examples of new trends inarchitecture have been selected becauseexperience has shown the force of imagescreated by architectural innovators, howeverstrange they may at first appear. The architectsillustrated are those -among many- who arehaving great influence in the schools ofarchitecture and among younger professionals.Engineers may expect to be confronted by thesekinds of configurations in the coming years.

Engineering rationality, and evenbuildability, appears to have little influence onthese forms. There is controversy in theprofession about this, and many critics view thenew architecture as akin to theater set design, inwhich image is everything and its method ofconstruction and longevity is irrelevant. Be thatas it may, the zeitgeist is changing, andarchitects will perforce have to obey it.

324 Chapter 6

Successful engineers will understand theseimperatives, enjoy the experimentation that thiswork represents, and assist the architects inrealizing their ambitions. New methods ofanalysis will help, but engineers must alsocontinue to develop their own innate feeling forhow buildings perform, and be able to visualizethe interaction of configuration elements thatare quite unfamiliar.

Meanwhile, the residue of configurationproblems left by the architecture/engineering ofthe 50's to 70's must be dealt with. Some willdisappear as aging buildings are replaced: thisshould be encouraged, as it is the onlyguaranteed way of removing the earthquakethreat. For other buildings, engineers must usetheir ingenuity and imagination to findaffordable methods of retrofit. And there needbe no recriminations: these problems are thejoint product of architect/engineer interactionthat, in its time, was fruitful: nature always hasthe last word in reminding us of our collectiveignorance. Figure 6-49. Rapid Transport District Headquarters, Los

Angeles, Frank Gehry, architect, 1995

Figure 6-48. Guggenheim Museum, Bilbao, Spain, Frank Gehry, architect, 1996


Simple, economical buildings will continueto be built, and our optimal seismic design willcontinue to be viable. It may form the basis ofperformance based design which, if it is to besuccessful, will have to be free of the kinds ofirregularities that make performance predictiondifficult or impossible. We may expect designto develop in ways analogous to the poetry andprose of written communication. Mostdiscourse is carried out in prose: the serviceablelanguage of business and news reporting. At thelevel of literature, prose approaches an art form,in which the subtleties of language and humanbehavior are explored. Out in advance, oftenalmost unintelligible, are the poets using wordsand language in new and unexpected ways: butover time they reveal insights in language socompelling that our speech and even ourbehavior is changed. Thus the language ofShakespeare shows up in the newspaper andeven the office E-mail.

REFERENCES

6-1 Willis, C., Form Follows Finance, PrincetonArchitectural Press, New York, 1995

6-2 Arnold, C., "Architectural Aspects of SeismicResistant Design" : Proceedings, Eleventh WorldConference on Earthquake Engineering, Acapulco,1996

6-3 Building Seismic Safety Council, NEHRPRecommended Provisions for Seismic Regulations forNew Buildings, Building Seismic Safety Council,Washington, DC (1997)

6-4 Seismology Committee, Structural EngineersAssociation of California, Recommended LateralForce Requirements and Commentary, StructuralEngineers Association of California, 1990

6-5 Seismology Committee, Structural EngineersAssociation of California, Recommended LateralForce Requirements and Commentary, StructuralEngineers Association of California, 1980

6-6 Reitherman, R.K., "Frank Lloyd Wright's ImperialHotel: a Seismic Re- evaluation" ,Proceedings,Seventh World Conference on EarthquakeEngineering. Istanbul, 1980

6-7 Dowrick, D.J., Earthquake Resistant Design, JohnWiley and Sons, London, 1989

6-8 Naito, T., "Earthquake-proof Construction", Bulletinof Seismic Society of America, Vol. 17, No. 2, June1977.

6-9 Degenkolb, H.J., "Seismic Design: StructuralConcepts" Summer Seismic Institute forArchitectural faculty, AIA Research Corporation,Washington, DC, 1977

Figure 6-50. Fire Station, Vitra factory complex, Weil-am-Rhein, Germany Zaha Hadid, architect, 1995

326 Chapter 6

6-10 Dewell, H and Willis, B, "Earthquake Damage toBuildings," Bulletin of. Seismic Society of America,Volume 15, No. 4 , Dec. 1925

6-11 Comartin, C.D., ed., "Guam Earthquake of August 8,1993: Reconnaissance Report", Earthquake Spectra,11, Supplement B, Earthquake Engineering ResearchInstitute, Oakland, CA, 1993

6-12 Athanassiadou, C.J., Penelis, G.C and Kappos, A.J.:Seismic Response of Adjacent Buildings with Similaror Different Dynamic Characteristics, EarthquakeSpectra, Volume 10, Number 2, EarthquakeEngineering Research Institute, Oakland, CA, 1994

6-13 Freeman, S.A., "Drift Limits: Are They Realistic."Structural Moments, Structural EngineersAssociation of Northern California, Berkeley, CA1980

6-14 Jeng, V, Kasai, and Maison, B.T.: A Spectral Methodto Estimate Building Separations to Avoid Pounding,Earthquake Spectra, Volume 8, Number 2,Earthquake Engineering Research Institute, Oakland,CA, 1992

6-15 Blume, J.A, Newmark, N.M, and Corning,L.H,Design of Multistory Concrete Buildings forEarthquake Motion ,Portland Cement Association,Skokie, IL, 1961

6-16 International Conference of Building Officials(ICBO), 1958 Uniform Building Code (UBC),Whittier , CA 1958

6-17 Fuller, L.P. "Going Global". World Architecture, 42 ,London, 1996

Chapter 06-Architectural Consideration

Documents

Transcript of Chapter 06-Architectural Consideration