***

ISSN 1018-5593
* *
* *
* *
***

European Commission

r f

L
f
rrc-ry ry; r> r; ~ .-y r*. rv r~.
rf
I

LA. 1/

_ L.
r-N / i . i ! ! 1 r

Properties and service performance

Serviceability deflections and displacements

in steel-framed structures

f(

Report
hi
EUR 15819 en STEEL RESEARCH
 European Commission

Properties and service performance

Serviceability deflections and displacements

in steel-framed structures
C. Bijl(\E Bijaard<2), R. Zandonini (3), D. Nethercot (4)
-'Centrum Staal
Groothandelsgebouw A-4
Stationsplein 45
Postbus 29076
3001 GB Rotterdam
The Netherlands
TNO Bouw
Lange Kleiweg 5
Rijswijk
Postbus 49
2600 AA Delft
The Netherlands
,3> Universit di Trento - Fac. Ingegneria
Via Mesiano 77
I-38050 Trento
Italy
w University of Nottingham
Nottingham NG2 7RD
United Kingdom

Contract No 7210-SA/612

1 July 1990 to 31 December 1992

Final report

Directorate-General
Science, Research and Development
1997 EUR 15819 en
LEGAL NOTICE

Neither the European Commission nor any person acting on behalf of the Commission
is responsible for the use which might be made of the following information

A great deal of additional information on the European Union is available on the Internet.
It can be accessed through the Europa server (http://europa.eu.int)

Cataloguing data can be found at the end of this publication

Luxembourg: Office for Official Publications of the European Communities, 1997

ISBN 92-828-0167-5

European Communities, 1997

Reproduction is authorized, provided the source is acknowledged

Printed in Luxembourg
 Contents

1 General aspects 8
1.1 Serviceability limits 8
1.2 Load levels and combinations 9
1.3 Deformation types 10
1.4 Reliability 10

2 Review of related studies and existing tests 11

2.1 General studies on serviceability 11
2.2 Loads to be considered 16
2.3 Economic aspects......... 16
2.4 Numerical analyses related to serviceability 17
2.5 Effects of cladding on deflection limits 20
2.6 Literature review on full-scale testing 22
2.7 Site measurements , 23
2.8 Summary 24
References 24

3 Existing code requirements and design practices 28

3.1 Serviceability limit states in current codes 28
3.2 Conclusions 34
References 35

4 Serviceabil itv problems in existing steel-fra med buildings 36

4.1 Review of previous studies r.. ....... 36
4.2 Lateral movements 38
4.3 Differential settlement 38
4.4 Ponding . 40
4.5 Serviceability survey 40
4.6 Conclusions 43
References 43

5 Non-structural components of steel framed buildings 44

5.1 Connections and details 45
5.2 Precast concrete cladding 46
5.3 Brickwork 46
5.4 Cinder blocks and clay tiles ^3
5.5 Stonework 54
5.6 Wood framed diaphragms $4
5.7 Glass curtain walls 62
5.8 Profiled steel sheeting 63
5.9 Sandwich panels 63
5.10 Conclusions 64
References 66

6 Structural modelling and calculation methods at serviceability load levels .... 68

6.1 Numerical study 69
6.2 Joint action and frame response 71
6.3 Joint action and frame performance in service *^4
6.4 Cladding action 91
6.5 Modelling of cladding action and cladding action on the considered frames 98
6.6 Conclusions 106
References 107

7 Testing and analysis of a full-scale steel framed building 109

7.1 Modelling of a two storey steel-framed office building ... ...~ 109
7.2 Static model predictions 118
7.3 Dynamic model predictions 120
7.4 Summary of model predictions 121
7.5 Testing of two-storey steel framed office building 121
7.6 Comparison of test results and model predictions 131
7.7 Theoretical principles behind building measurements 133
References 138

8 Evaluation of the actual structural behaviour at service load level

of full-scale structures by in situ dynamic tests 139
8.1 Structures investigated 139
8.2 Numerical models and results . 143
8.3 Physical tests 147
8.4 Comparisons 150
8.5 Conclusions . 152
References - 154

9 Summary of the major findings . ....... 155

9. 1 Design codes and serviceability problems in steel-framed buildings 155
9.2 Measured in-service behaviour of steel framed buildings 155
9.3 Design models at serviceability load levels 157
10 Conclusions and recommendations 160
10.1 Recommendations 160

11 Literature list for all related references 162

Introduction

The aim of the project is to match simplified design calculation methods to determine
the deflections and displacements of framed structures at serviceability load levels, to
the performance actually observed in service. The performance in service, automati
cally takes account of all available components providing rigidity to the frame, and
recognises real constraints or acceptable limits of deformation. This also means a criti
cal review of the traditional approaches to serviceability, the identification of the most
appropriate parameters to be checked and the definition of the most appropriate design
procedure.
The research project deals with the behaviour of steel framed structures for multi
storey and industrial buildings in particular the deflections and displacements under
serviceability loading conditions.
The expected results of the work is to provide answers to two fundamental questions:
What levels of in-service deformation are appropriate for various key components
in steel framed buildings, taking account of the effects of such deformations on the
performance of the complete structure?
What form of simplified design calculations are appropriate as a means of conduct-
ing a quantitative check that satisfactory in-service performance will result?

Three main tasks for the technical aspects of the work have been identified:
1. Identification of presently used design limits for serviceability deflections from
codes of practice, general custom and practice within the industry and from manu
facturers requirements, including supporting evidence for particular limits where
this is available.
2. Determination of frame deflections by different methods of calculation.
3. Evidence from on-site measurements and large scale tests of deflection levels ob
served in structural assemblies.

All the parties will be involved in the different aspects of the study. However, it was
agreed that one member would take lhe lead on each of these three topics of follows.
Task 1: University of Nottingham (contract 7210-SA/828)
Task 2: Universit di Trento (contract 7210-SA/418)
Task 3: TNO Building & Construction Research (contract 7210-SA/613)
Coordination and secretary: Centrum Staal (contract 7210-SA/612)
1 General aspects

With the adoption throughout Europe of the limit states basis for the design of steel
structures, attention has been focused more clearly on the different performance re
quirements that must be satisfied. Whilst much of the material of recently produced
design codes and standards e.g. Eurocode 3, concentrates on providing assistance
with the design for the ultimate limit state, these documents also set out the designers
obligation to ensure satisfactory performance at working load levels - the so-called
serviceability limit state. However, the actual guidance provided is much less detailed,
despite the fact that modem trends towards lighter and more open forms of construc
tion are likely to mean that serviceability considerations will increasingly control the
design.

1.1 Serviceability limits

Although several different serviceability checks should, in principle, be conducted for
all types of steel structure, this investigation concentrates on the particular require-
ments'.that static deformations under working load conditions should remam within ac
ceptable levels. Thus it does not address behaviour under dynamic loads and the par
ticular issues of vibration, natural frequency, excitation and resonance. Nor does it
deal with other serviceability requirements such as resistance to deterioration through
corrosion or fatigue damage.
For steel structures it has been customary to base the serviceability deflection check on
a comparison between deflection calculated on a linear elastic basis for a simplified
representation of the bare steel frame and sets of limits drawn from experience of the
satisfactory performance of previous structures designed on a similar basis. Little sci
entific evidence exists of either the levels of deflection that could safely be permitted in
different forms of construction without impairing their everyday performance or of the
actual deformations experienced by real structures for which complex interactions be
tween frames, joints, floors, walls, partitions and foundations mean that calculations
for bare frames have little real relevance. Thus the link between the calculations and the
codified deflection limits is an empirical one, serving merely as an indication (or not)
that the structure will not be too flexible, rather than as a mean of assessing actual in-
service behaviour.
Various aspects of serviceability limits are considered in chapters 2, 3, 4, 5 and 6 of
this report. These embrace not only the review of current codified limits presented in
chapter 2 but also evidence of actual in-service performance obtained both from the
work of other investigators and from measurements undertaken as part of the present
 study. The link between these two rather different types of limits is, of course, the
method used in the design process to estimate serviceability deflections.
If the calculations were to be performed in a way that accurately modelled all of the ef
fects present in the real structure that had some influence on its stiffness and thus on
the deflection obtained, and if the loading used for the calculations correctly repre
sented the way in which the actual in-service loading functioned, then the results could
be compared with actual performance limits i.e. the overall sway at which connections
between the main frame and the curtain walling ceased to function properly, the level
of floor beam deflection for which an unacceptable degree of cracking in the slab was
introduced etc. Such an approach to deflection calculations would, with the tech
niques, computing resources etc., likely to be readily available to designers for the
foreseeable future be impractical. It is also doubtful that all of the structural effects -
and, more importantly, the parameters necessary to describe them properly in every set
of circumstances, given the one-off nature of most structural designs - could ever be
faithfully represented by an analysis, no matter how complex.
Thus the reality of the situation is that designers must 'make do' with something sim
pler. The widespread use of linear elastic analysis - implemented increasingly often
nowadays by means of commercially produced software mounted on a personal com
puter - means that approaches to deformation checking at serviceability are likely to
continue to rely on this method for the calculations well into the future. What is
needed, therefore, is the setting of limits (for use with different types of unacceptable
in-service behaviour that ensure that the real structure will not exceed the performance
limits associated with real events that would render it unfit for use.

1.2 Load levels and combinations

Similar arguments to those advanced in the previous section concerning deflections
apply to the other side of the equation: the loading. Real loading on structures is, of
course, extremely complex. That this is the case may be accepted by considering the
difficulties associated with first assessing the correct magnitude of each component
type e.g. dead, imposed, wind etc. - especially when one considers that each of these
will itself have several components - and then deciding upon realistic combinations.
Although the subject has for many years been treated on a probabilistic basis - thereby
recognising its variable nature and the need to consider matters such as the design life
of the structure and an acceptable risk that an unwanted effect will occur - for the pur
poses of design calculations explicit values are required. Thus, once again, it is neces
sary to associate particular quantities with the design process, knowing that these have
been chosen so that the whole package of calculations will give a result that is satisfac
tory in the overall sense. The basis for selection must, however, be evidence that the
use of these levels has been correlated against assessments of the true behaviour of the
real structure under its actual loading conditions.
 It is also necessary to be careful, when selecting loading arrangements for checking the
serviceability conditions, mat both the load levels and the combinations used reflect in-
service rather than ultimate conditions. It is suspected that, just as for the response side
of the equation, far more attention has been directed to the collection, analysis and rep
resentation of data for me ultimate limit state.
Some rather limited, attention is given to the question of the loading cases that should
b considered when checking serviceability behaviour in chapter 2. This forms part of
the review of codified or similar material presently available on the subject The paral
lel aspect of actual in-service loading experienced by various forms of steel building is
not, however, considered.

1.3 Deformation types

The actual pattern of deformation that will be experienced by a steel building during its
lifetime will be complex. Traditionally serviceability checks in design consider only
particular forms under an associated idealised loading case e.g. maximum vertical de
flection of a floor beam under uniformly distributed vertical load. These are normally
selected so as to present particular unsatisfactory events e.g. local damage to non
structural partitions. Thus they represent spot checks on the structure's ability to re-
spond in an acceptable fashion. The link between these and the sort of deformation that
might be experienced in service due to the almost limitless combinations of a loading
that might occur during the lifetime of the building is, on the basis of the present
study, somewhat difficult to make.

1.4 Reliability
Clearly uncertainty plays a large role in all aspects of the prediction of structural be
haviour and its links to true behaviour. From the forgoing qualitative consideration of
the treatment of serviceability deflections this would appear to be especially significant
for that topic. Ideally the three main data items:
- loading;
- structure properties;
- limits
should all be based on probabilistic concepts as each should reflect the variability of
the subject e.g. use should be made of load combinations with an appropriate likeli
hood of their being experienced during the life of the structure. To do this properly
would, of course, require enormous volumes of data item to be available. Clearly this
is unrealistic and a suitably pragmatic approach is required.

10
 Review of related studies and existing
tests

This chapter reviews previous work on serviceability problems and the existing test
data (at service load levels). Also numerical analyses are reviewed together with the
techniques used to determine the characteristics influencing serviceability.

2.1 General studies on serviceability

What follows are the findings of a literature survey examining previous works under
taken in this field by Associations, research groups and individuals:
A Dutch document on serviceability requirements, translated into English (Canada
Institute for Scientific and Technical Information, CISTI, 1980) recommends the fol
lowing for the effects of static deformations and their allowable values:
- Water accumulation (on roofs): it can be prevented by judiciously determining the
point of water discharge.
- The subjective aspect: becomes more significant if the deformations become visi
ble. The needs arising from this subjective aspect can largely be satisfied by con
forming to the requirements that Z^/L < 250 (L is the span in question, ZM is the
actual sag in the final state), see figure 2. 1. The code requires that when calculating
Z,, any rotation occurring at the 'fixed-end' should be taken into account

deflection (ultimate value = 0 camber Zg

Sag (ultimate value = Z^ )

' Unloaded (including no self weight)
loaded (including creep)

2.1. Requirements for deflections (CISTI, 1980).

The use aspect: this is to ensure permanent serviceability of the floor structure.
Requirements depend on each individual situation and there is no general rule.
The construction aspect static deformations in the floor and roof of structures may
give rise to cracking or other damage in members which are supported by these
structures (a typical example is the cracking in partitions). As a recommendation for
beams or floors supported on two or more ends, the following limitations were
suggested:

11
 ^< 500 to 600, andaiso
OarJd < 10 to 20 mm
where,
L = span parallel to the partition wall
2jrj = additional deflection occurring after installation (hardening) of the wall

The maximum rotation (due to static loading and special influence) is suggested to
be:

wi<0max< 300
5O0

Other recommendations were also made concerning the dynamic effects on buildings
and structural components. An Ad Hoc committee produced a report establishing
structural serviceability research needs (Ad Hoc Committee on serviceability Research,
ASCE, 1986). The committee had the responsibility for:
- defining the serviceability problem and its scope on contemporary construction
- identifying areas where sufficient data appear to exist
- identifying the areas where additional research is necessary.
The authors of the report suggested mat good understanding of the behaviour of the
individual components in a structure is essential for an adequate assessment of the de
formations that occur in the structure as a whole.

In Australia a survey was conducted into the deflection limits of portal frames
(Woolcock and Kitipornchai, 1986). A questionnaire on specific items was sent to in
dividual designers and organisations. The result of the survey is summarised in table
2.1 for lateral deflection limits, and table 2.2 for rafter deflection limits. One feature
about the survey is the diversity of the engineer's opinions oh many aspects of ser
viceability, and would indicate that this issue is not clearly understood by many engi
neers.

Table 2.1. Recommended lateral deflection limit (Woolcock and Kitipornchai, 1986).

Type of building Limits Comments

Industrial a. steel sheeted walls, inceillings, no h . b relatve deflection between
buildings internal partitions against external 150 '200 adjacent fimes
walls or columns, no gantry cranes
b. as in (a), but with gantry h b - h may be taken at crane level
250 '250 - h/300 should be used for
heavy cranes
c. as in (a), but with external masonry h b
walls supported by steelwork 250 ' 200 ij

h . b
Farmsheds
100 '100

12
Table 2.2. Recommended rafter deflection limit (Woolcock and Kitipornchai, 1986).
Type of building and load Limits Comments
Industrial a. dead load L forroofpiteh>3
buildings 360
L for roof pitch < 3% but check for ponding or insufficient
500 roof slope
b. live load L check spread of columns if gantry crane present
240 if no ceiling
L
250
c. wind load L
250
Farmsheds a. dead load L check for ponding if roof pitch <3
240
b. live load L
180
c. wind load L
100

A study was undertaken in New Zealand by the Building Research Association

(Branz) on serviceability criteria for buildings (Cooney and King, 1988). Tables 2.1
and 2.2 give some examples for the limiting deflection values of horizontal and vertical
components respectively. Covered in the report are different aspects of deformations
such as deflections due to loads, thermal expansion, environmental and dynamic ef
fects. Tables 2.3 and 2.4 give the reason for limiting deformations as well as the load
combination. As stated by the authors, the deflection components being assessed
should be related to me assumptions made with respect to:
- the section modulus
- changes in section (e.g. composite sections)
- loading assumptions (intensity and distribution)
- duration of load
- flexibility of the support
- environment effects
- shear distortions.

For example in concrete structures the outcome of the analysis will be strongly influ
enced by the section modulus.
A state-of-the-art survey of the design for drift of steel framed buildings has been un
dertaken by the ASCE Committee on Design of Steel Building Structures (Galambos
and Ellingwood, 1986). This was accomplished by means of a survey of structural
engineering consultants (35 responses were received). Some of the results of this sur
vey, which pertain to this report, are listed below:
- More information is needed to define acceptable drift limits for human occupancy
and for the structural acceptability of curtain wall systems.
- Codes should indicate levels of quality. Perhaps an upper bound of lateral drift
should be specified.

13
 Inter-storey drift limits currently used by engineers vary between 1/600 and 1/200.
Drift should be codified because it is ignored by so many engineers.
Drift limits are meaningless if the limitations between drift and damage for common
cladding materials and partitions types are not defined.
Different building types should have different drift limits.
Many engineers believe that different drift ratios should be used for different
cladding types. This is rejected by the working group because the contribution of
the same cladding types is highly dependent upon connection type, location and
number.
It is meaningless to codify a set of drift values if it is not applied using consistent
methods for calculating drift. No component of deflection should be ignored if it
has a significant contribution to the total

Table 2.3. Examples of limiting deflection values for horizontal components (Cooney and King,
1988).

Reasons for limit Deflection limi Load combina

ing deflections tations tion Examples and comments
water accumulation S 1 D (allow for
(ponding) on roofs L<250 creep) plus
etc. for beams paral rainwater or
lel to line of snow melt
roof slope
beams that support 1 D+Lor D+S - reinforced concrete or steel beams support
surfaces which L<250 ing slabs
should drain water 1 D+Lor D+S - trafikable deck supported by timber beams
L<350 - non-traflicable deck supported by timber
1
D or D+S beams (always check that water flows as de
L<600 signed)
L = live load; D = dead load; S = snow load
differential settle L - beams supported masonry walls
ment 100 - beams supporting walls other than masonry

Table 2.4. Examples of limiting deflection values for vertical components (Cooney and King, 1988).

Reasons for limit Deflection limitations Load combina

ing deflections tion Examples and comments
sway of columns 1 D+W applies especially to
due to wind h<500 multi-storey buildings
and per storey: < 4 mm D -dead load
W = wind load
frame deflection due j horizontal deflection at eaves W W = wind load
to wind and earth
quake
frame spacing
L< 200 ' -;
and in end bay: < 40 mm
differential settle- J 8 1 - masonry
ment h 300
1

- other materials
1 h < 150

14
In the U.S codes the vertical deflections were seen to vary in the range L/360 to L/180
depending on the Standard used (in the USA each State has it own Standards) and on
the properties of the roof and floor of the structure.

A working programme was established 'to provide information relevant to the struc
tural serviceability of buildings with special attention to structural design' (Holicky and
Deak, 1989). For that purpose four sub-groups were formed with the following tasks:
- serviceability requirements
- design concepts
- deformation of floors and roofs
- floor vibration

A state-of-the-art document corresponding to the findings of the four sub-groups is

due to appear during 1993.
An IABSE seminar was held on the subject of serviceability limit states for steel
buildings at Zurich in 1989. The aim of the seminar was to address the following
questions:
- Have all relevant serviceability requirements for common buildings been identified?
- Which basic design principles should be applied to meet the serviceability require
ments for buildings?
- How shall the question of responsibility for adequate building performance be
handled?.

The participants in the seminar, coming from ten different countries, confronted the
following aspects of this problem:
- Deflections of continuous beams in steel frames with composite floors of steel and
concrete (Johnson, 1989).
- Limitation of lateral displacements in steel frame design (Golembiewski, 1989) and
(Tschemmernegg, 1989).
- Vibrations in buildings (Sato and Yoshida, 1989), (Yoshida and Sato, 1989),
(Uchida etal., 1989), (Kuhlmann,1989).

Although the outcome of the seminar was promising, the contributors admitted that it
may take a few years before agreement can be reached for the setting of serviceability
design rules.
In modern construction a number of problems associated with limit states related to ex
cessive static deformation (deflection, rotation, curvature) can be identified. The fol
lowing is a list of some of the most common problems (Galambos and Ellingwood,
1986):
- local damage to non-structural elements (eg. ceilings, partitions, walls, doors and
windows, etc.) due to deflections caused by load, temperature variation, shrinkage
or creep, and moisture changes

15
 - deterioration of the structure due to age and use (fatigue)
- extensive damage of non-structural elements due to excessive natural events (eg.
hurricane, tornado, etc.)
- noticeable deflections causing distress to occupants
- discomfort due to vibrations (produced by machines, traffic, blast or wind, etc.)

One of the main tasks of the construction industry is to ensure that such problems are
properly countered and minimised. The use of adequate materials in the construction,
properly connecting the various components of the structure (through efficient bolting
and welding), allowing for thermal expansions by providing sufficient separation be
tween deflecting primary structural elements and non-structural components, etc., are
all factors that will eventually help in reducing, to a considerable extent, problems as
sociated with deformations and deflections in the structure.

2.2 Loads to be considered

Depending on the circumstances, it may be necessary to consider (SCI, 1991):
- dead load
- imposed load
- all gravity loads (i.e dead and imposed)
- wind load
- wind load plus dead load
- 80% of (wind load plus imposed load)
- 80% of (wind plus imposed) plus 100% dead load.

The dead load need normally only be considered where its effects are not already com
pensated by the initial camber of the frame.
It must be stated that very often deflections are specified by engineers without a clear
definition of the load to be considered nor its duration (creep effect). Care must be
taken in considering load combinations for a particular deflection control criterion.
These combinations need to be applied to the most critical combination for the elements
which may influence the deflection. The choice of such combinations is often not ob
vious and should be worked out carefully depending on the type of element, the func
tion of the structure and the controlling effect Gong- or short-term).

2.3 Economic aspects

Limiting deflections to the right level of serviceability requirements in a structure is an
important issue as far as economy is concerned. A related article was published in a
seminar on 'serviceability limit states for steel buildings' held in Zurich (Golem-
biewski, 1989). He showed that, the value of h/150 in limiting the lateral deflection of
hall structures due to wind, and adopted by the Swiss Steel Construction Standard
(SIA 161, 1991), is a hard demand. A value of h/100 is sufficient. This was the result

16
 of many years experimental research undertaken in the old GDR which proved that
with this limit value of h/100 damage is not to be expected. As stated by Golembiew-
ski this difference is in fact significant, since sharpening h/100 to h/150 requires up to
15% more steel in the case of heavy roof claddings and up to 35% in the case of light
weight roof claddings.
The outcome of such a study strongly supports the idea that the emphasis of future re
searches should be on the exact definition of the limit state of serviceability.

2.4 Numerical analyses related to serviceability

In this section literature on numerical analyses related to serviceability problems is re
viewed.
An approximate method of analysis has been developed for multi-storey infilled frames
subjected to lateral loads (Stafford Smith and Carter, 1964). In this method an equiva
lent strut is described to replace the infill panels. The emphasis of this report is on cal
culating the ultimate strength of such a structure, but the basic analysis may also be of
use at service load levels.
A graphical method of predicting sideways deflection in the design of multi-storey
buildings has been proposed (Wood and Roberts, 1975). The method was based on
the Hardy-Cross moment distribution technique. The authors recommended a control
ling sway angle of 1/250 in every storey of the building for base frame design, and
1/400 for those storeys of clad frames where the proposed composite design method is
attempted. They suggested that future full-scale tests on realistically clad frames com
plete with windows, doorways etc., should be undertaken.
A revised method of analysis of multi-storey infilled frames subjected to lateral load
has been developed (Liauw and Lee, 1977). Tests and modelling were performed for
frames with full and partial connection between the in-fill panels and bare steel frame.
Estimates of in-fill panel stiffness are given. As with the previous model, emphasis is
placed upon ultimate failure load predictions, but useful information at service load
levels was also observed.

The results of finite element analyses on shear walls for moderately tall frames with
cut-outs are described (Chang-Koon, 1987). A new finite element is described which
gives good results and can be more easily used than conventional elements. The major
conclusions of this study are:
- The new element is well adapted for the study of structures with both structural and
non-structural components.
- The accuracy of this element decreases as the cut-out ratio (ratio of the width of the
opening to the width of the entire element) increases. It is recommended that cut
out ratios of 0-30% should be used.

A model used to predict the lateral load behaviour of an entire wood frame structure
has been developed (Schmidt and Moody, 1989). This model assumes that there are a

17
series of shear frames arranged in rectangular fashion between a rigid floor and roof.
Major conclusions from this study are as follows:
- Lateral and rotational behaviour of a structure consisting of non-linear shear walls
between rigid floors and roofs can he estimated.
- Reasonable agreement is observed between mis model and full scale tests.
The strength limit states design of structures is nowadays treated in a large number of
software packages (including buckling analysis, plastic analysis, etc.). On the other
hand, though the introduction of the serviceability limit states design is still very mod
est, it is increasingly recognised by the design packages. Among these are the SAP-90
(Wilson and Habibullah, 1989), and Dispar (Charney, 1990). The Dispar (acronym
for Displacement PARticipation factor) program is suitable for planar frame and
framed tube structures. It is particularly interesting since it is based on a numerical
factor, representing the member's contribution to the displacement occurring at a speci
fied point and a specified direction (Charney, 1990). Of further interest is the fact that
the factors may be easily broken down into components of flexural, axial, shear and
joint deformation (in % value) .
A sizing technique which can be used to minimise the material required for the lateral
bracing of multi-storey buildings has been published (Baker, 1990). This technique,
based upon energy principles, has been described for steel frames with a number of
different lateral bracing systems. This method is most useful when lateral drift arid not
strength is the controlling factor. Such a method could be of use when describing
multi-storey structures acting compositely with non-structural components.
The effect of non linearity and joint flexibility on the lateral drift of steel structures has
been investigated (Ho and Chan, 1991). The program GMNAF developed by the au
thors recognises geometric and material nonlinearities and includes the - effect in the
analysis. They found that ignoring the P-5 effect leads to 9% error while a misuse Of
joint stiffness may result in an error of up to 45%.

Efforts to model an eight storey reinforced concrete building that was destroyed during
an earthquake are given (Wood, eLal., 1991). The model that was used includes inte
rior partitions and the exterior cladding. Lateral drift and stiffness were calculated.
The results of a comparison between analysis and modelling for a 13 storey steel-
framed office building have been published (Maison and Ventura, 1991). Measured
lateral motions were recorded during two earthquakes. During these earthquakes the
building responded in a predominantly linear elastic manner. Six models were made
each with increasing levels of sophistication. The models, in order of increasing so
phistication, accountedTor the effects of the following pararneters:
- Planar model. Frames are modelled with beam-column and beam elements.
- Three dimensional model. Frames are modelled as before. The mass of each floor
is assumed to act at the centroid of the buildings

18
- Same as the previous model except that the beam to column joints are modelled.
This changes the effective lengths of each beam and column. Beams and columns
are assumed to be rigidly connected.
- Same as the previous model except that - effects are included.
- Same as the previous model except that the mass on each floor is more realistically
distributed.
- Same as the previous model except that full composite action between the floor
slabs and steel- frame is assumed.

The results of comparisons between these models and the measured response of the
building suggested the following conclusions:
- A linear elastic model can be used to adequately predict the true dynamic behaviour
of a steel moment-resisting frame.
- The parameters that have the most effect upon the building studied are the mod
elling of the frames themselves and the modelling of rigid joints between the beams
and columns.
- Parameters of relatively small importance are the effects of beam slab interaction,
- effects and the distribution of mass on each floor.
- Design-type analytical models can reasonably model the true seismic response of a
steel-momnt-resisting frame provided that the applied loads are small (the struc
tural response remains linear and elastic).

A simple method of approximating the lateral stiffness of elastic moment-resisting

frames has been developed (Schultz, 1992). It is stated that a single value can be used
to represent the stiffness of each story in an elastic rectangular frame which is sub
jected to regular distributions of lateral load. It is shown that three existing expressions
for lateral stiffness are only valid when beams are flexurally suffer than the columns.
This proposed method includes the effects of the following parameters, which are
stated to have a significant influence upon lateral stiffness:
- the effect of unequal heights for adjacent stories,
- the influence of top and bottom boundaries (for each floor),
- the stiffening effect of the basein low-rise frames.

Several examples are given for various frame geometries and boundary conditions.
The major conclusions are as follows:
- The proposed method may only be used for moment resisting frames which are
fixed at the base and only flexural deformations are considered.
- Three correction factors in the proposed method, h, Cs and xs, provide reasonably
good estimates of story lateral stiffness even for frames with columns that are as
much as ten times suffer than the beams.
- Three correction factors in the proposed method, h, Cs and xs, provide reasonably
good estimtes of story lateral stiffness even when storey heights and member
stiffness (beams and columns) differ by up to 50%.

19
2.5 Effects of cladding on deflection limits
A review and evaluation of various structural systems employed in current building
practice has been published (ACI Committee 442, 1971). The principal objective of
this document is to explain the different lateral load carrying mechanisms in buildings.
Part of mis document, however, deals with serviceability criteria. It is mentioned that
buildings built before 1950 had less rigorous lateral load limitations than those im
posed in the 1970's (L/300 as opposed to L/500). The reason for this change was ex
plained to be due to changes in non-structural building components. Before 1950
heavy masonry partitions and exterior cladding were common. Since 1950 the trend
has been towards lighter and lighter partitions and cladding. Lateral drift limits are
stated to be rather arbitrary in nature.
Steel framed buildings where the wall and the roof are sheeted have been extensively
studied (Bryan and Davies, 1972). The presence of the sheeting was considered as a
stressed skin which imposes restraining forces on the frame. This has led to a signifi
cant decrease of the frame lateral deflection calculated on the usual assumption that all
loads are carried by a set of internal frames. The authors described the stressed skin
design concept and the difficulty arising from coupling the behaviour of the framed
system with that of the sheeting.
The interaction Of structural elements with cladding in tall buildings has been studied
(Dubas, 1972). The deflection behaviour of a multi-storey frame under lateral loads
was studied taking into account the cladding effect He showed that cladding can con
tribute by as much as 30% to a decrease in the frame deflection, this decrease is de
pendent on various parameters, of which the number of storeys of the frame is particu
larly significant (for tall buildings for instance the effect of cladding becomes more im
portant).

The effect of using wall cladding on the behaviour of high rise steel buildings has been
investigated (Scalzi and Arndt, 1972). It was found that the lateral deflection may be
reduced by 20 to 50% depending upon height and length to width ratio of the floor
plan, without increasing the weight of the structural frame.
Design methods have been proposed for including structural frames and non-structural
elements (in-fill panels), when subjected to lateral loading (Wood, 1978). Test results
of frames with in-fill panels are used to verify theoretical and design predictions. Only
ultimate load levels were examined, not service load levels. Major conclusions from
this study include the following:
- The shear strength to be gained by including in-fill panels is considerable. Due to
the magnitude of this effect simple, conservative, rules can be proposed.
- Further work must be done to include the effects of openings such as doors and
windows.

In-fill panels in steel framed structures have been examined with respect to provide
extra passive damping (Gasparini, etal., 1981). Providing extra damping for light and

20
flexible structures can help to meet serviceability requirements for lateral movements,
accelerations and vibrations. Design examples are given and the following conclusions
made:
- Significant increases in the damping of a steel frame may be realised by incorporat-
ing in-fill panels into the structural analysis.
- The resulting structural system (structural elements and in-fill panels) may be anal
ysed using a simple method which utilises an equivalent plane stress rectangle finite
element to represent the actual panels.
- Preliminary design of such systems can be done by simple hand calculations.
- In-fill panels can be designed to have considerable stiffness and acceptable energy
absorbing characteristics.

A survey into the serviceability of buildings has been conducted (Huggins and Barber,
1982). It consisted of a questionnaire sent to Canadian consulting engineers and
building officials. Cases of deflection, distortion, and vibration distress were reported,
m summary it was concluded that where deflections may lead to wall cracking as well
as window and door opening distortion, it may well be that some upper limit should be
applied to permissible deflection in addition to the usual L/360, etc., requirements
(alternatively joints between the wall separations should be provided to follow the de
flection without cracking).
The case history of a high-rise steel framed office building that was designed as a
composite structure with the exterior cladding to reduce lateral drift has been published
(Tomasetti, 1986). The type of cladding used were flat steel plates about 8 mm thick,
with cut-outs for window openings. It is stated that these panels were not included in
ultimate limit state calculations, but included in serviceability limit state calculations.
Details of the connections between cladding and the structural system are shown. It
was concluded that:
- Including the cladding in service limit state calculations reduced service sway by a
factor of two.
- The cost efficiency of the underlying steel frame was significantly improved.
- Additional usable floor space was obtained due to a reduction in the size of the
structural frame.

The effect of the envelope on high rise building frames has received attention
(Bergmann, 1988). Particular attention was given to the interaction of the envelope and
me supporting structure.
The influence of non-structural panels and cladding on serviceability criteria for lateral
drift has been published (Bat, eLal, 1991). It is stated that lateral movements in Rus
sian design codes are given as a function of the integrity of non-structural elements.
Non-structural elements examined were 'blank and glazed walls and partitions', and
the 'pliability of joints between cladding elements and skeleton'. The displacement of a
building as a whole is limited to L/500. The maximum displacement of any one storey

21
 is limited to L/300 and L/700, depending upon 'the mode of connection of the walls
and partitions to the building skeleton' and the 'material from which the panels and
cladding are made'.
Numerical and experimental studies were performed to estimate the actual stiffening ef
fect of concrete cladding connected to a bare steel portal frame (Gaiotti and Smith,
1992). Both modelling and testing were performed on a one-storey one-bay frame.
The objective of this study was primarily to determine the forces that are developed in
connections between the concrete cladding and steel frame. As such connections were
classified into three groups and individually studied. The contribution of individual
components to the total lateral stiffness was estimated. Results indicated that the frame
with concrete cladding and frame transmit substantial shear forces. It is not clear,
however, if the objective of this study is to examine cladding-frame interaction at ser
vice or ultimate load levels. Lastly, an offset diagonal brace was proposed which could
represent the bracing action of the cladding.

2.6 Literature review on full-scale testing

Test programs in the following review include only those from which static lateral
stiffness may be estimated at service load levels. This implies that the lateral stiffness
is determined using the response spectrum of the building obtained from a dynamic
excitation. Due to the lack of references on steel-framed structures, reports for other
framing systems such as reinforced concrete have been included in this review.
A complete panel brick multi-storey building in an abandoned quarry has been tested
(Sinha and Hendry 1976). The buildings behaviour was recorded under simulated lat
eral wind loads (serviceability load level), and at the lateral collapse load. Lateral loads
were provided using hydraulic Jacks.
At serviceability load levels test results indicate that the building can satisfactorily be
analysed by replacing the actual structure with an equivalent frame in which the
columns have the same sectional properties as the walls.
Static and dynamic tests on a portal-framed storage building were undertaken (Strand
and Pirner, 1978). The structure consists of six two-hinge portal frames with spans of
12 m. Portal frame hinges are located at the eaves. The total length of the building is
30 m. Exterior cladding consists of steel sheets 1mm thick with rib heights of 50 mm,
connected to 160 mm deep channel purlins. The purpose of these tests was to examine
the interaction of the steel frame and thin- walled steel profiled cladding, when the
structure is subjected to wind loads. Particular interest is expressed about the dynamic,
not static, contribution of wind load and its effect upon the fasteners between decking
and steel frame.
Static and dynamic tests were performed on both the bare steel frame and the com
pleted structure. Static measurements showed a significant improvement in lateral
stiffness and load carrying capacity due to the exterior cladding. This effect can be rea
sonably predicted using existing stressed skin design recommendations. The influence

22
 of the dynamic components of wind loads were shown to have negligible effects upon
the immediate and long term lateral stiffness of the completed structure.
The results of an experimental study investigating the lateral stiffness of an 80 m high
office building have been published (Jeary, etal., 1979). The structural system of this
building is a concrete shear core with cast-in-place floor slabs and external reinforced
concrete columns. Non-structural blockwork (internal partitions) are used. Exterior
cladding consists of both glass and lightweight panelling.
A shaker (eccentric mass vibrator) was used to excite the structure. The shaker was
placed near the top of the building. The conclusions of this testing program are the
following:
- The exterior cladding and interior partitions contributes significantly to the overall
lateral response of the structure. If the cladding and partitions did not contribute,
the building would have unacceptable lateral movements and vibrations. Evidence
of the participation of the interior partitions is easily confirmed as cracking regu
larly occurs during and after the passage of high winds.
- The cladding and partitions have effectively changed the location of the shear centre
of the building. As a result lateral loads produce a significant torsional response.
- Typical design assumptions ignoring non-structural components can lead to signifi
cant differences between assumed and actual in-service behaviours.

Testing of a 46-storey, 190 m tall, reinforced concrete office building has been re
ported (Jeary, etal., 1979). The structural system consists of a massive heavily rein
forced concrete core, with three 'leaves' attached and supported by large concrete cor
bels at their bases. The three 'leaves' are of different heights, and contain office space.
Each leaf consists of steel columns, onto which are attached cladding. Concrete floors
in each 'leaf are cantilevered from the concrete shear core and attached to the steel
columns at the periphery.
The buildings response is measured using both wind loads and a shaker placed on the
43rd floor of the building. Conclusions drawn from this study are primarily concerned
with comparisons between design values and measured wind loading responses.
Design values are stated to be grossly in error, the buildings response being smaller
than predicted. Only passing mention is made concerning modelling considerations.

2.7 Site measurements

A summary of the results of a survey of existing data on ninety-eight buildings is
available (Skempton and MacDonald, 1956). The work consisted of actual measure
ments of building settlements. Based on differential settlements (from field or labora
tory tests) and angular distortion of the building, the authors proposed a series of val
ues of the ratio 67L for different types of buildings and soils. It was found, for exam
ple, that the settlement characteristic principally causing cracking was probably the an
gular distortion. The latter is conveniently expressed by the ratio of the differential set-

23
 dement and the distance L between two points. From all the field data on cracking in
buildings a general limiting value of /L = 1/300 was obtained.

2.8 Summary
The literature survey reported in this chapter revealed the existence of a wide range of
documents related to different aspects of serviceability. These studies are not suffi
ciently complete, however, to permit researchers to answer the following questions:
a. load levels and intensity to be used in determining deflection limits,
b. basis (background) for the values used in the deflection limits,
c. contribution of the different components (non-structural elements) in the structure
overall deflection.

Concerning point (c), the following can be said. Previous tests of full-scale buildings,
regardless of the framing system, indicate mat exterior cladding and interior partitions
participate structurally at service load levels. Their participation may even be dominant
at service load levels. The effects of such a participation can be broadly divided into
two groups as follows:
- Beneficial. This includes increased stiffness, thus smaller lateral movements, and
decreased lateral vibrational problems.
- Detrimental. The building does not behave as expected by the designer. In some
cases this means that cracking of non-structural elements may occur. The buildings
response may be so different from that predicted by the designer that other unfore
seen problems occur, such as introducing a torsional response into what was
thought to be a symmetrical structure.

Only in the case of portal frames with thin-walled profiled steel sheeting are the effects
of cladding presently accounted for.

References
ACI Committee 442, 'Response of buildings to lateral forces', Journal of the American Concrete
Institute, p. 81-106, 1971.

ASCE Committee on Design of Steel Building Structures, 'Wind drift design of steel-framed build
ings. State-of-the-art report', Journal of the Structural Division, ASCE, Vol. 1 14, No.9, Sep.1988, p.
2085-2108.

Ad Hoc Committee on Serviceability Research, Committee on Research of the Structural Division,

'Structural Serviceability. A critical appraisal and research needs', Journal, Journal of the Structural
Division, ASCE, Vol.112, No.l 12, p. 2646-2664, December 1986.

WJ7. Baker, Sizing techniques for lateral systems in multi-storey buildings, Proceedings Tall
Buildings: 2000 and beyond, p. 454-554, November 1990.

. Bat, V.A. Otstavnov and LI. Lemysh, On deflections and displacements, CTB W85 Reference
Document R91: IB, 1991 (Translated from Russian).

24
R. Bergmann, Structural serviceability aspects of building envelopes in tall buildings, Proceedings of
the symposium/workshop on serviceability of buildings, National Research Council, Canada, p. 293-
322, May 1988.

ER. Bryan and JAI. Davies, Stiffening effect of light cladding, Proceedings of the International Conf.
on Planning and Design of Tall Buildings. Lehigh University, Bethlehem, Pennsylvania, August 21-
26, 1972, p. 643-651.

Canada Institute for Scientific and Technical Information (CISTI), Deformation requirementsfor build
ings, joint publication of the Steelbuilding Association, Rotterdam and the Concrete Association,
Zoetermeer. Report NRa CNR TT-1969. Canada, Ottawa, 1980.

F. Charney, DISPAR for SAP. A post processor for the SAP90 Finite Element Analysis Program,
Advanced Structural Concepts Division, J JL Harris and Company, Denver, Colorado, 1990.
-, Sources of elastic deformation in laterally loaded steel frame and tube structures. Design methods
based on stiffness, J.R. Harris and Company, Denver, Colorado, 1990.

C.-K. Choi and M.S. Bang, 'Plate Element with Cut-out for Perforated Shear Wall', Journal of the
Structural Division, ASCE, VoL 113, No. 2, February 1987.

R.C. Cooney and AJB. King, Serviceability criteria for buildings, Building Research Association of
New Zealand. BRANZ study report. Report SR14, 1988.

P. Dubas, Interaction ofstructural elements with cladding, Proceedings of the International Conference
on Planning and Design of Tall Buildings. Lehigh University, Bethlehem, Pennsylvania, p. 675-683,
August 21-26, 1972.

R. Gaiotti and B.S. Smith, 'Stiffening of moment-resisting frame by precast concrete cladding',
Precast/Prestressed Concrete Institute (PCI) Journal, September/October 1992, Vol.37, No.5, p. 80-
92.

D.A. Gasparini, L.W. Curry and A. DebChaudbury, 'Damping of Frames with Visco-elastic In-fill
Panels', Journal of the Structural Division, ASCE, May 1981, Vol. 107, No. ST5, p. 889-905.

D. Golembiewski, On the limitation of lateral displacement in designing steel frames,

Seminar/Workshop on serviceability limit states for steel buildings. IABSE Working commission ,
Zurich, April 1, 1989, p. 8-15.

WJM.G. Ho and SL. Chan, On the effect of non linearity and joint flexibility in lateral drift determi
nation of steel buildings, Hong Kong Polytechnic, Department of Civil and Structural Engineeringj
Research Report CE/010491, April 1991.

M. Holicky and G. Deak, Basic principles of the control of serviceability, Draft for CUB Programm
W85, February 1989.

M.W. Huggins and J J5. Barber, 'Building deflection, distortions and vibrations. A survey', Canadian
Journal of Civil Engineering, March 1982, VoL 9, No.l, p. 133-137.

A.P. Jeary, B.E. Lee and P.R. Sparks, The determination of Modal Wind Loads from Full-Scale
Building Response Measurement, The International Conference on environmental forces on engineer-
ing buildings, held at Imperial College, London, July 1979.

A.P. Jeary and B.RlIis, A Study of the Measured and Predicted Behaviour of a 46-Storey Building,
The International Conference on environmental forces on engineering buildings, held at Imperial
College, London, July 1979.

25
Ri*. Johnson, Deflection of continuous beams in steel frames with composite floors of steel and con
crete, Seminar/Workshop on serviceability limit states for steel buildings. IABSE Woiking commis
sion , Zurich, p. 3, April 1, 1989.

U. Kuhlmann, Slender footbridges, Seminar/Workshop on serviceability limit states for steel build
ings, IABSE Working commission , Zurich, April 1, 1989, p. 55.

T.C. Liauw and S.W. Lee, On the behaviour and analysis of multi-storey infilled frames subject to
lateral loading, Proceedings of the Institution of Civil Engineers, Part 2, Vol. 63, September 1977, p.
641-656.

J. Maison and CJE. Ventura, 'Dynamic Analysis of Thirteen-Story Building', Journal of the
Structural Division, ASCE, Vol. 117, No. 12, December 1991, p. 3783-3803.

K. Sato and M. Yoshida, Predicted and observed vibration of high-rise buildings and related human
sensitivity during typhoon, Seminar/Workshop on serviceability limit states for steel buildings,
IABSE Working commission , Zurich, April 1, 1989, p. 18-25.

JJB. Scalzi and A. Arndt, Plate wall cladding, Proceedings of the International Conference on Planning
and Design of Tall Buildings, August 21-26, 1972, Lehigh University, Bethlehem, Pennsylvania, p.
653-665.

RJ. Schmidt and R.C. Moody, 'Modelling Laterally Loaded Light-Frame Buildings', Journal of the
Structural Division, ASCE, Vol. 115, No. 1, January 1989, p. 201-217.

Ali. Schultz, 'Approximating Lateral Stiffness of Stories in Elastic Frames' ,

Journal of the Structural Division, ASCE, Vol. 118, No. 1, January 1992, p. 243-263.

B.P. Sinha and A.W. Hendry, 'Structural testing of brickwork in a disused quarry', Proceedings,
Institute of Civil Engineers, Part I, Vol. 60, February 1976, p. 153-162.

A.W. Skempton and D.H. MacDonald, 'The allowable settlements of buildings', Journal of the
Structural Division, ASCE, May 1956, p. 727-784.

B.S. Smith and C. Carter, 'A method of analysis for infilled frames', Proceedings of the Institution of
Civil Engineers, Vol. 43, September 1969, p. 31-48.

M. Strand and M. Pirner, 'Static and dynamic full-scale tests on a portal structure', The Structural
Engineer, September 1978, Vol. 56b, No. 3, p. 45-52.

RL. Tomasetti, A. Gutman, LPiew and LM. Joseph, Development of thin wall cladding to reduce
drift in high-rise buildings, IABSE colloquium on thin-walled metal structures. Stokholm, 1986, p.
239-246.

F. Tschemmernegg, On the limitation of deflection of sway-frames, Seminar/Workshop on service

ability limit states for steel buildings, IABSE Working commission , Zurich, April 1, 1989, p. 16-
17.

N. Uchida, M. Kawamura, T. Aoyagi and H. Kirihara, Vibration offloor beams due to walking occu
pants. Study on measured data, Seminar/Workshop on serviceability limit states for steel buildings,
IABSE Woiking commission , Zurich, April 1, 1989, p. 31-52.

. Wilson and A. Habibullah, SAP90 Finite Element Analysis Progra. Computers and Structures,
In., Berkeley, California, 1989.

26
R.H. Wood and E.H. Roberts, 'A graphical method of predicting sideways in the design of multistorey
buildings', Proceedings of the Institution of Civil Engineers, June 1975, Vol. 59, Part 2, p. 353-372.

R.H. Wood, 'Plasticity. Composite action and collapse design of unreinforced shear wall panels in
fimes', Proceedings of the Institute of Civil Engineers, June 1978, Part 2 No. 65, p. 381-411.
-, 'Effective lengths of columns in multi-storey buildings. Part 3', The Structural Engineer, No.9,
Vol. 52, September 1974, p. 341-346.

S.L. Wood, R. Stark and S.A. Green, 'Collapse of Eight-Storey RC Building During 1985 Chile
Earthquake', Journal of the Structural Division, ASCE, February 1991, Vol. 117, No. 2, p. 600-619.

S.T. Woolcock and S. Kitipornchai, 'Deflection limits for portal fimes. Steel Construction', Journal
of the Australian Institute of Steel Construction, Vol. 20, November 1986, No.3, p. 1-12.

M. Yoshida and K. Sato, Predicted and observed vibration of high-rise buildings and related human
sensitivity during typhoon, Seminar/Workshop on serviceability limit states for steel buildings,
IABSE Working commission , Zurich, April 1, 1989, p.26-27.

27
 Existing code requirements and design
practices

The increased understanding of building materials and structural behaviour achieved

during the last few decades, has led to the use of lighter structures having greater
flexibility. This means that structures are more likely to experience larger deformations
in service. The traditional design concept based on allowable stresses is not therefore
sufficient to ensure acceptable performance in service as well as an adequate margin
against failure. For the structure to perform acceptably (drift, settlement, vibration,
etc.) at working loads, specific serviceability requirements, which are dependent on a
number of parameters (type of structure, type of foundations and soil, environmental
conditions, etc.) are intended to prevent any excessive deformations from taking place
under load combinations likely to be experienced in service. The diversity of control
ling parameters means that easy-access, simplified serviceability design rules are not
readily available. However, in most recently produced Codes and Standards an attempt
has been made to specify design requirements for serviceability, though sometimes in
a very modest way.
A review has been conducted that has shown that numerous serviceability design cri
teria exist but that these are spread diversely through codes, papers, journal articles,
technical reports, standards etc., or are simply the customary practice of individual
engineers. In this chapter are reviewed the serviceability requirements as reported in
some design codes/standards.

3.1 Serviceability limit states in current codes

Deformation limits for roofs and walls of profiled sheeting have been reviewed
(Douhan, 1980). Swedish regulation on this matter were also reviewed. The origin of
such limits was also specified based on some test results. The series of tests was con
ducted on trapezoidal sheeting of aluminium with uniformly distributed load. For ex
ample, f or a sheeting of 40 mm depth and 0.6 mm thickness it was found that folding
began when the least radius of curvature of the profile was 25 mm, and since the re
lationship to distributed load is:
1 M qL2
L: span (m); EI: flexure (Nm2)
~ FJ " 8EI

and deflection y: max. deflection (mm)

y 385EI
y 40L 40
therefore,
384p 384p

28
 y L
for = 25 m for a simply supported profile
L_240
-JL. for a two-span profile
L"330
These limits may be used as a basis for a design criteria for profiles of low bearing ca
pacity, m table 3.1 are given some examples of limiting values.

Table 3.1. Swedish sheeting manufacturers regulations for limiting deformations (due to dead and
snow and wind loading) (Douhan, 1980).

Conditions Deflection limitation

roofs with insulation and felt L
for L< 4.5 m
150
for4.5m<L<6m 30mm
L
for L> 6.0 m
200
roofs with insulation and strip metal L
150 'JL_
roofs with insulation and double sheeting
90
uninsulated roofs L
90 .L.
walls not subject to special requirements
L 90
walls subject to stringent requirements as regards appearance
: no

Design for serviceability is increasingly becoming an integral part in the Standards of

many countries. This interest is, as described earlier, mainly dictated by safety as well
as economical considerations.
Two categories of limit states have defined in the American steel design code (Fisher
and West, 1990):
- the strength limit states controlling the safety of the structure which must be met,
- the serviceability limit states defining the functional performance of the structure
that should be met

The LRFD specification lists five topics which relate to serviceability concerns. They
are:
a. camber
b. expansion and contraction
c. deflections, vibrations and drift
d. connection slip
e. corrosion.

Special interest was given to topic (c). Tables 3.2 and 3.3 give a summary of some of
the recommendations adopted by the AISC for roofing and cladding deformations.
The British steel design code makes a provision for serviceability limit states design
(BS5950, 1990). The code, however, specifically excludes portal frames. Two types

29
of limit states are considered: deflection and durability. For the latter, the code sug
gests that the following factors should be considered at the design stage:
- the environment
- the degree of exposure
- the shape of the members and the structural detailing
- the protective measure, if any
- whether maintenance is possible.

Table 3.2. Serviceability considerations - Roofing (Fisher and West, 1990).

Roofing type Structural Deformation Recom. Loading
element TPaxiTpym
membrane roof roofing expansion hor. movt 150 to 200 thermal
joints feet
metal deck (two span) vert defl. J-L 3001b cone, load at
200 mid-span
metal deck vert defl. J-L live load
240L
metal deck verLdefl. J-L
2401j
2001b cone, load at
mid-span
steel joists vert. defl.
J-L
240L
live load

joist girders vert defl. J-L live load

240
roofs vert defl. J_L dead load + live load
240L
roofs slope 1. 1 u drainage
to j111***1
metal roofs through
fastener type
purlin vert defl. J-L
250L
snow load

metal roofs standing purlin

seam
vert defl. J-L
150 L
snow load

Table 3.3. Serviceability considerations - Cladding (Fisher and West, 1990).

cladding cladding type and support cMormation recom. maximum loading

support type element
metal panels/bare frame drift perpendicular to J_ J- h 10 year wind
wall 60 100
metal panels/girts or wind
columns
nor. deflection
J-L
1201j
10 year wind

foundation precast walls/bare fiame drift perpendicular to J-h

100
10 year wind
wall
reinforced masonry drift perpendicular to J-h
200
10 year wind
walls/bare frame wall
masonry walls/girt or nor. deflection 10 year wind
wind column 24"L<1.5in
vert deflection dead load +
masonry walls/lintel ^L<0.3in live load
columns pre-assembled units/ bare racking J-h 10 year wind
frame 500h

30
In Table 3.4 are shown some deflection limits as suggested by BS 5950.
The new Australian steel design code states that responsibility for selecting deflection
limits rests with the designer, but gives some recommendations (AS4100, 1990). The
code gave some suggestions concerning vertical as well as horizontal deflections limits
for certain types of structure (see tables 3.5 and 3.6).

Table 3.4. Deflection limits for certain structural members (BS 5950, 1990).
Structure Member Limit
a. Deflection on beams due - cantilever length
to unfactored imposed 180
loads - beams carrying plasteror other brittle fin
ish * (L = SpaD)
L
- allotherbeams
200
b. Horizontale deflection of - top columns in single-storey buildings
columns other than poral 3 O"1"*
names due to unfactored - in each storey of a building with more than
imposed and win loads one storey height of storey
300
c. Crane gantry girders - vertical deflection due to static wheel load span
- horizontal deflection (calculated on the top 600
flange properties alone) due to crane surge span
500

Table 3.5. Suggested vertical deflection limits for beams (AS 4100, 1990).

Deflection to be consid Deflection limit for Deflection limit for

Type of beam ered spanLt) cantilever L)
beam support deflection which occurs 5-<J- 5-<J-
ing masonry after the addition or at L "SOO L "250
partitions tachment of partitions where provision is made where provision is made
to minimise the effect of to minimise the effect of
movement otherwise movement otherwise
.1
< J
L -1000
i<J-
L "SOO
all beams total deflection < 1
1
L ^250 L ^ 125
t Suggested deflection limits in this table may not safequard against ponding.
$. For cantilevers, the values of 6/L given in this table apply, provided that the effect of the rotation at
the support is included in the calculation of .

Table 3.6. Suggested horizontal deflection limits for beams (AS 4100, 1990).

Building clad in steel or aluminium sheeting wihthout gantry cranes and withhout
internal partitions against external walls 150
Building with masonry walls supported by steelwork
240

The new European steel design code defines the following serviceability limit states for
steelwork (Eurocode 3, 1991):
- deformations or deflections which affect the appearance or effective use of the
structure (including the malfunction of machines or services)

31
- vibration, oscillation or sway which causes discomfort to the occupants of a build
ing or damage to its contents
damage to finishes or non-structural elements due to deformations, deflections, vi
bration, oscillation or sway.

Concerning deflections, Eurocode 3 defines the deflection of a beam m (sagging in

the final state relative to the initial straight line, see figure 3.1) as being the algebraic
sum of three types of deflections , , -*.

Omax = + 2 -
where
max = sagging in the final state relative to the straight line joining the supports.
= pre-camber (hogging) of the beam in the unloaded state, state (0)
= variation of the deflection of the beam due to the permanent loads immedi
ately after loading, state (1).
$2 = variation of the deflection of the beam due to the variable loading plus any
time dependent deformations due to the permanent load, state (2).

(oU - -

- J ',
Trow ^ /wWK
""-
^ (D
* 'max
(2)

3.1. Vertical deflections to be considered (Eurocode 3, 1991).

The vertical deflection limitations for buildings are given in table 3.7 in which L is the
span of the beam (for cantilever beams this length is doubled). In table 3.8 are given
the horizontal deflections at the tops of the columns Oi is the height of the column or of
the storey, ho is the overall height of the structure).

Questions concerning the design of portal frames, omitted by the BS5950, has been
discussed in the Journal of the Steel Construction Institute (Steel Construction Today,
SCI, 1991). For a portal frame such as that shown in figure 3.2, the deflection limits
for pitched roofs are given in table 3.9a/b (table 3.9a. for horizontal and 3.9b. for ver
tical deflections). A wide range of side and roof cladding materials is covered in the
table.

32
Table 3.7. Recommended limiting values for vertical deflections (Eurocode 3, 1991).
Conditions Anax 2
- roofs generally L L
200 250
- roof frequently carrying personnel other than for maintenance L L
250 300
- floors generally L L
250 300
- floors and roofs supporting plaster or other brittle finish or non-flexible L L

V
cic
partitions
- floors supporting columns (unless the deflection has been included in the
global analysis for the ultimate limit state)
where Omax can impair the appearance of the building

Table 3.8. Recommended limiting values for horizontal deflections -- sway (Etnocode 3, 1991).

- portal frames without gantry cranes

- other single storey buildings

- in a multistory building: in each storey

on me structure as a whole
250
L
400
L
250

Table 3.9a. Deflection limits for pitched roof steel portal frames (SCI, 1991). Horizontal deflection at
eaves level, due to wind or imposed roof load or 80% (wind and imposed).

Typeofroof
h
150
h
300
b
300
ho
500

Absolute Differential deflection relative to adjacent

350
L
500

deflection fiame (see figure 32)

1

side cladding: profiled metal sheeting

4> -

fibre reinforced sheeting <Jl

"150 -'..

brickwork _Vh2 + b2
"3 - 660
hollow concrete blockwork ^/h2 + b2
- 500

precast concrete units .Vh2 + b2

- 330
roof cladding: profiled metal sheeting
-200
fibre reinforced sheeting .-.
<-*- ' ' 'I
"250
.

felted metal decking -

<Jl-
- -- - ' ..,, ..-.-..^ . 1
-400

33
 Table 3.9b. Deflection limits for pitched roof steel portal frames (SCI, 1991). Vertical deflection at
ridgev(for slopes > 3*), due to wind or imposed roof load or 80% (wind and imposed).
Type of roof Differential deflection relative to adjacent frame
(see figure 32)

profiled metal sheeting .Vb2 + s2

*I "* 125

fibre reinforced sheeting .-Vb2 + s2

I * ** 165

felted metal decking, supported on purlins ,Vb2 + s2

j ad - 250

felted metal decking, supported on rafter ,Vb2 + s2

z^r and
- 250
- - - - - ._.....
200

S =r + 0./2)

3.2. Portal frame definitions (SCI, 1991).

3.2 Conclusions
The investigation carried out on the serviceability requirements has shown the impor
tance of the issue. This is illustrated by the increase in popularity of the subject, being
recognised by many Codes and Standards. The diversity of influencing parameters
made easy-access to simplified serviceability design rules very difficult and patchy.
Due to these diversity, no attempt was made to compare serviceability limits from dif
ferent codes. This difficulty was further increased by the lack of a clear definition of
the load combination or the structural component to which the deflection limit applies.
For example, for cantilever beams the Australian Code specifies a limit of L/125
whereas the British Code gives a limit of L/180, L being the span of the beam. As can
be seen the difference is quite remarkable and as stated in Section 3.4, the economic
aspect will be very decisive for design. In order to be able to produce a unified code,
experimental data on real structures may have the last word to decide on what deflec
tion limits should be adopted in design for serviceability.

34
References
Ad Hoc Committee on Serviceability Research, Committee on Research of the Structural Division.
'Structural Serviceability. A critical appraisal and research needs', Journal of Structural Engineering,
ASCE, Vol.112, No. 112, p. 2646-2664, December 1986.

Australia Standards AS4100.1-1990. Steel Structures Code, 1990.

BS 5950, The use of structural steel in buildings (1990). British Standard Institution.

Commission of the European Communities, Eurocode 3. Design of steel structures. Part 1. General
Rules and Rules for Buildings, 1991.

L. Douhan, Deformation limits for roofs and walls for profiled sheeting, Swedish Council for
Building Research, Document D32: 1980, Stockholm, Sweden, 1980.

J.M. Fisher and M.A. West, Serviceability design considerations for low-rise buildings, American
Institute of Steel Construction (AISC), 1990.

T.V. Galambos and B. Ellingwood, 'Serviceability limit states. Deflections', Journal of Structural
Engineer, ASCE, No. 112, Jan.-Apr., 1986, p. 67-84.

T.V. Galambos, P.L. Gould, M.K. Ravindra, H. Suryoutomo and R.A. Crist, Structural deflections.
A literature and state-of-the-art survey, U.S. Department of Commerce, National Bureau of Standards.
Building Science Series 47, October 1973.

D. Golembiewski, On the limitation of lateral displacement in designing steel frames,

Seminar/Workshop on serviceability limit states for steel buildings, IABSE Working commission ,
Zurich, p. 8-15, April 1, 1989.

SIA 161, Stahlbauten, Edition 1979.

Steel Construction Today, The Steel Construction Institute (SCI), Advisory Desk. VoL 5, No. 4, p.
203-206, July 1991.

35
 Serviceability problems in existing steel-
framed buildings

This chapter contains a compilation of reports of unsatisfactory structural behaviour in

existing steel-framed buildings. This review is restricted to the following types of
problems which may be observed during normal in-service use:
- lateral movements;
- differential settlement;
- ponding.
In some instances exceptional loading conditions have been included The following
subjects, however, are not included in this compilation:
Static floor displacement. Static floor displacements are excluded because such de
flections are primarily due to the floor system itself. The contribution of the rest of the
building (including the steel frame) to vertical floor deflections are of secondary impor
tance. Much guidance can be found in the Uterature and in design codes concerning the
calculation of static floor deflections.
Dynamics and vibrational displacements. Dynamic and vibrational problems are not
included in this literature review because they are the object of an on-going CIB study.
This study is being undertaken by the CIB working group W85 'Structural
Serviceability'.

The period surveyed includes the years between 1970 and 1991. Key words used in
the uterature search were steel construction, serviceability, performance, test, testing,
deflection(s), deformation(s), drift(s), side-sway, shear stiffness(es), ponding, wind,
stability, failure, pseudo-static and dynamic. Several thousand references were re
viewed using the Brix, Iconda, Ntis, Compendex and Pascal data systems. Additional
references, available at TNO, Nottingham and Trento, were also reviewed.
Articles containing previous reviews of serviceability problems in steel-framed build
ings are first presented. Articles of particular interest are presented under one of the
following headings: lateral movements, differential settlement and ponding.
Preliminary results are included of a survey on serviceability sent to designers in the
United Kingdom, France, Italy, Switzerland and the Netherlands.

4.1 Review of previous studies

A review of the most prevalent types Of structural failures has been attempted [Janney,
1972]. The results of this review are summarized as follows:

36
 Ponding. Ponding can be considered as a serviceability and/or an ultimate load
problem. It is due to a combination of the load-defection characteristics of a relatively
flat roofs structural system, improper or inadequate drainage and sufficient source of
water such as a torrential downpour. Typical errors are the placement of drains near
columns, drains that are too small and inadequate roof cambering.
Bearing. Bearing failures are normally considered as ultimate limit state problems.
This type of failure is most commonly associated with precast beams or slabs.
Shear. Shear failures are normally considered as ultimate limit state problems. This
type of failure is related to bearing failure, and is most often due to punch through in
slabs near columns.
Stability. Stability failures are normally considered as ultimate limit state problems.
It is stated that stability failure often occurs as the final result of other structural prob
lems.

A survey of serviceability problems experienced by 385 designers throughout Canada

was conducted [Huggins and Barber, 1981]. This survey was intended to gather in
formation in the following general categories: vibration, deflection of supporting
members, bulging of cladding and excessive side sway. The major results of this sur
vey are as follows:
Vibrations. Most problems were connected with open web steel joists.
Deflection of supporting members. Several different problems were reported. First,
cracking of masonry or partition walls because of excessive deflections. Rainwater
penetration due to exterior curtain wall movements. Concrete cracking due to thermal
shock, inadequate shoring and heavy construction loads at an early age.
Bulging of cladding. These problems were reported to be both frequent and costly.
Most problems were due to poor design details and/or poor workmanship and super
vision.
Excessive side-sway. Few (two) cases were reported, but all proved to be costly to
repair. One case was due to an incompatibility of deflections between adjoining slabs,
the other due to support rotation.

A review of problems related to the proper assessment of serviceability limit states is

reported in the literature [Leicester and Pham, 1987]. They state that there are two in
teracting serviceability problems:
- Problems of designing for serviceability.
- Problems of assessing actual serviceability in the case of a claimed serviceability
failure.

Two basic types of serviceability data are needed, in-service performance and com
plaint thresholds. The in-service performance of building structures was examined in
this review, for government buildings in Australia (circa 1981-1984). The following
serviceability related problems were reported, as a percentage of all complaints:

37
 - wall cracks 65%
- jammed doors and windows 15%
- vibration 10%
- floor deflection 7%
- window distortion 3%

Conclusions from this article include the following:

. - The first step towards the rationalisation of serviceability design limits is the speci
fication of 'legal' performance limits.
- With available data the in-service performance of most structural components can
be estimated.
- Most available data on serviceability are of limited use. Surveys can be used as rep
resentative but biased sample of compiami thresholds.

A review of the causes behind premature deterioration of exterior cladding in high rise
structures has been reported [Cheung and Khan, 1990]. This study illustrates the diffi
culty behind determining the exact cause of an observed failure. For the case of
cladding, three failure mechanisms can be identified: Moisture deposition, building
distortions and chemical or ultra-violet attack. Cracking can result from any one of
these mechanisms. This indicates that not all serviceability problems may be directly
attributed to excessive building movements.

4.2 Lateral movements

A steel-framed high-rise building subjected to extreme winds has been examined
[Minor, etal., 1972]. The structure (The Great Plains Life Insurance Building, Lub
bock, Texas) was subjected to tornado force winds. Cracking of the exterior cladding
(brick-work) and of interior walls was documented. Permanent lateral drift was mea
sured as a function of building height

4.3 Differential settlement

A 35 year old publication is included in this review because it closely addresses the
subject studied [Skempton and MacDonald, 1956]. A summary of the results of data
on 89 buildings with differential settlement problems were presented. Tentative values
for damage limits in terms of angular distortion were established. Damage (or the lack
thereof) at different settlements was reported for load-bearing brick wall structures and
for steel or reinforced concrete frame buildings with panel walls similar constructions.
Damage (and lack of damage), is reported for both structural and non-structural ele
ments as a function of angular distortion. The results of this study may be summarised
as follows:
- Damage to non-structural specimens, built and tested in the laboratory, is reported
at angular distortions smaller than L/300.

38
- Non-structural damage in buildings occur at angular distortions as small as L/300.
This is primarily for panels made of brick, tiles and clinker blocks.
- Structural damage in buildings is not reported until an angular distortion greater
than L/200 is observed.

Table 4.1. Summary of structural damage reports due to ponding.

Case Normal Emercency Slope Parapet Remaries
nr. discharge outlet
1 clogged no 0.18% yes
2 clogged no 1% yes' '

3 small no 0.3% yes failure of bolted connection

4 NLP 0.28% no slip in rigid connection
5 high no 0.4% yes full discharge capacity at 70 mm
6 NLP no 0.4% yes
7 small no 0.8% yes .

8 NLP no unknown yes torsional buckling of beam

9 small no 0.1% 200 mm
10 clogged no 0.61 % yes
11 NLP no 1 % 200 mm torsional buckling of beam
12 NLP no 0.06% yes very sensitive for ponding
13 NLP yes 0.3% yes emerc. outlets located wrong
14 small no 0.65% 850 mm
'.
15 clogged 1% 200 mm
. ' '
no
16 small no 1% 150 mm
17 small no 1% yes vertical discharge opening
18 small no 0.5% yes
19 NLP 100 mm 025% 250 mm slope locally too low
20 small small 0.9% 140 mm structural system sensitive to non-uniform
loading

Normal discharge
NLP: rain pipe is not located at lowest point
dogged: partly clogged with leaves branches and other rubbish
small: in some cases the full discharge capacity is only reached when there are some em's water on
the roof; in other cases the slope causes a non uniform flow
high: the opening of the rain pipe is placed too high
Emercency outlet
no: there is no emercency outlet created
not relevant because there is no parapet
...mm: emercency outlet placed at a height of ... mm above the roof
Slope
%: actual slope near damaged section
Parapet
yes: the roof has a parapet
...mm: the height of the parapet above the roof

39
4.4 Ponding
Reports on structural damage due to ponding, available at TNO Building and Con
struction Research, have been collected and analysed. Individual reports are not in
cluded in the reference list For steel structures most cases relate to partial collapse
caused by ponding of rain water.
The results of this survey are presented in table 4. 1. It can be concluded that similar
characteristics in case of failure are:
- Slope of damaged part was less or equal to 1%.
- Water dis charge capacity was too small either by dimensions, location (placed to
high) or accumulation of rubbish.
- Strength and stiffness of the roof were not based on the height of the parapet or lo
cation of emergency outlets.

4.5 Serviceability survey

As part of this study a questionnaire on serviceability sent to designers in the United
Kingdom, France, Italy, Switzerland and the Netherlands. An English language copy
of this questionnaire is shown in figure 4.1. Note that it is divided into the parts:
- The impact of design code provisions for serviceability.
- The effectiveness of design code provisions for serviceability.
The objective of this questionnaire was to evaluate, from a practicioners point of view,
the following:
- The types and uses of steel framed buildings for which existing serviceability crite
ria require significant expenditure on the part of designers, fabricators or construc
tors. Such areas are of importance for further study.
- The types and uses of steel frames buildings for which problems most often occur.

Preliminary results of this survey are given in table 4.2. Final results, including
Germany and England will be available at a later date.

40
 Table 42. Preliminary results of the questionnaire on serviceability. Part A: The impact of design
code provisions for serviceability (Note: Moie than one answer can be given to each question).
Nr. Question CH F GB I NL
1 type of structure: bare steel 4 5 I
7 14
composite 1 1 ; 4 1
other _ _ _ __-

2 type of construction: portal 3 1 4 7

multi-storey 4 3 9 5
truss 3 2 10
other (a) _ _ .
1
3 principle building use: commercial 2 1
_

6 81
residential 2 .
6 2j
industrial 2 5 4 12
other (b) _ .
2
4 design modification: increase member size 41 3 4 14
further investigate limits 1 -r 5 1
use alternative design 1 5 10
other (0 -i _____
1 - . .- -

a. Responses included the following: Long spanning structures

b. Responses included the following: Public utility buildings
c. Responses included the following: Changing user requirements

Table 4.3. Preliminary results of the questionnaire on serviceability. Part B: The effectiveness of de
sign code provisions for serviceability (Note: More than one answer can be given to each question).
Nr. Question CH F GB I NL
1 type of structure: bare steel 4 4 7 14
composite 1 1 - 3 2
other . _ _ . _'
2 type of portal 3 1 3 8
construction: multi-storey 3 4 7 4
truss .
1 12
j other

3 principle commercial 2 1 _
6 9
building use: residential 2 2 . _
4 ' 2!
industrial 2 3 ._
3 13
other (a) -2l
4 type of problem: jammed doors/windowa 2 1
crane rail deflection 2 1 ' . _
3
beam/floor deflection 4 _ .
4
beam/floor soffit deflections
pending 2
_

3
._ _
2
3
3
10
foundation movements _
1 1 10
cracking of non-stmctural comp. 3 2 3 3
stniOure/non-stnjcture _______ . _
2 1
erection (tolerances, etc.) 2 4 3
construction (stability, etc.) 1 2 5
cairpies/cantitivers/balconies 2 _ _
3
other :

a. Responses included the following: Public utility buildings, light framed buildings and new roof in
stallation.

41
Survey on serviceability criteria for steel framed buildings
Preambe. The purpose of this questionnaire is to ascertain the impact and effectiveness of serviceabil
ity design code provisions for steel framed buildings. Serviceability limits that are of interest for this
survey are static onlv: ResDonse to dynamic loads, vibrations, resistance to corrosion and fatiaue are
not included.

Please answer each question by cireling the appropriate response:

A. The impact of design code provisions for serviceability

Identification of steel framed buildings where serviceability provisions require the most design
modifications.

1. Type of structure Bare steel

Composite
Other
2. Type of construction Portal
Mutistorey
Truss
Other
3. Principle buildings use Commercial
Residential
Industrial
Other
4. Design modification Increase member size
Further investigate limits
Use alternative design
Other

The effectiveness of design code provisions for serviceability

Identification of steel framed buildings where serviceability problems have most often oc
curred.
1. Type of structure Bare steel
Composite
Other
2. Type of construction Portal
Mutistorey
Truss
Other
3. Principle buildings use Commercial
Residential
Industrial
Other
4. Type of problem Jammed doors/ windows
Crane rail deflections
Beam/floor deflections
Beam/floor soffit deflections
Ponding
Foundation movements
Cracking of non-structural components
Structure/non-structural attachments
Erection (tolerances, etc.)
Construction (stability, etc.)
Canopies/cantilevers/balconies
Other
4.1. English language version of the serviceability questionnaire

42
4.6 Conclusions
Several references were found which review serviceability related damage to existing
buildings. Most of this information, however, is from occupant surveys. Reports on
damage at service load levels are not accompanied by drawings or details of the par
ticular structure for which damage was reported. Comparisons between serviceability
limits and observed damage for structural and non-stractural elements were reported.
This is the only study of the effects of a specific deformation type on building compo
nent performance.
The difficult task ofresearchers and code writers in establishing serviceability limits is
well documented. Legal serviceability limits imply estabshing acceptable damage
risks for each building component (stractural and non-structural). It remains to be ex
plained, however, how data such as occupancy surveys can be used to 'adequately'
estimate individual component behaviours. Such observations, however, must be used
when data is not available, field measurements or numerical models including non-
stractural elements.

It is important to note that damage due to lateral movements at serviceability limits are
often confined to non-stractural elements. This implies that at service load levels non
structural elements receive a large proportion of deformation imposed forces. This is
verified by Skempton and MacDonald's observations of real building behaviour at ro
tations near L/300. Thus, for normal steel construction practices, all non-stractural el
ements should be checked at service load levels. Exceptions may occur when stractural
elements control a design at service load levels. Such cases may be due to large lateral
drifts in slender structures, unacceptable dynamic response or due to excessive floor
deformations and vibrations.

References
M.S. Cheung and J. Khan, Premature Deterioration of Building Enclosures in High-Rise Structures,
Council on Tall buildings and Urban Habitat, Tall Buildings: 2000 and Beyond. Fourth World
Conference, Hong Kong, November 1990, p. 201-214.

M.W. Huggins and JD. Barber, 'Building deflections, distortions and vibrations. A survey', Canadian
Journal of Civil Engineering, Vol 9., No. 1, 1982, p. 133-137.

J.R. Janey, ACS: Structural failures: modes, causes, responsibilities, ACSE National meeting on
structural engineering, Qeveland, Ohio, April 1972, p. 11-20.

R.H. Leicester and L. Pham, Serviceability limits, The Institution of Engineers, Australia, 1987
(CIB-W85).

JE. Minor, K.C. Mehta and J.R. McDonald, ACSE: Structural failures: modes, causes, responsibili
ties, ACSE National meeting on structural engineering, Qeveland, Ohio, April 1972, p. 49-73.

A.W. Skempton and D.H. MacDonald, 'The allowable settlements of buildings', ASCE Journal of the
Structural Division, May 1956, p. 727-784.

43
Non-structural components of steel-
framed buildings

This chapter contains a review of typical non-stractural components and attachments

between structural and non-structural members used in steel-framed buildings. Non-
stractural components are classified at follows:
- Cladding. Non-stractural components which form the exterior walls.
- Partitions. Non-stractural components which form the interior walls.

The term 'non-structural component' refers to cladding and partitions that are not in
cluded in ultimate-limit-state design calculations. This does not imply, however, that
these components have no stractural integrity.
The in-plane strength and stiffness of non-structural components can have a substantial
effect upon a buildings lateral response at service and ultimate load levels. Reasons for
examining non-stractural components and attachments can be stated as follows:
1 . At present, designers limit lateral deformations (thus the applied force) in each non-

stractural component of a building. This is normally done by limiting individual

storey drifts. The model used by the designer, however, only takes into account the
strength and stiffness of the bare steel frame. In reality, the structure is much suffer
that the bare steel structure..
2. Non-structural components often contribute to the stractural response of steel
framed buildings at service load levels, whether or not the designer intended for
them to participate or not This participation that can lead to serviceability problems
such as cracking, gaps forming between cladding panels, windows or doors that
do not open or close properly, etc.
3. An estimate of the lateral in-plane stiffness of non-structural components is needed
if the designer wishes to include such components in building design calculations.
For service load levels initial stiffness may be assumed. This implies that cracking,
buckling, concrete crushing, plastification attachment deformations, etc. do not oc
cur. If significant degradation does occurs at service load levels, the structure is
clearly not serviceable. This greatly simplifies the models that may be used to in
clude non-strachiral components.

In order to address these three points, existing information on the following topics is
examined in this chapter:
- Previous investigations in this field.
- The initial in-plane stiffness of non-structural components typically used in steel
framed buildings.

44
 - The applied load or deformation at which service problems occur in non-stractural
components.
- The maximum load carrying capacity of non-structural components.
- The strength and deformation capacity of attachments.
- Typical constructive details for attachments.

Experimental and theoretical investigations of the compressive and out-of-plane bend

ing strength and stiffness of non-stractural components are not included in this review.
Such studies, while interesting, have little effect upon the lateral behaviour of steel-
framed buildings at service load levels. The stractural behaviour of the steel frame it
self and other components normally included in ultimate-limit-state design calculations
(reinforced or prestressed concrete, etc.) are not treated in this chapter. Components
that are included in ultimate-limit-state design calculation of some steel framed struc
tures, such as sandwich panels and profiled sheeting, are included in this review.

A literature review was undertaken for years between 1960 and 1992. Key words used
in searching the search were brick, masonry walls, joints, shear walls, mortar, cinder
blocks, clay tiles, timber, wood, plywood, adhesive and stone. Several thousand ref
erences were reviewed using the Brix, Iconda, Ntis, Compendex and Pascal data sys
tems. Only references investigating the lateral strength and stiffness of non-stractural
components commonly used in steel-framed buildings were retained. Additional refer-
ences, available at TNO, Nottingham and Trento, were also reviewed.
The results of this literature search are presented according to their primary subject
matter. Articles reporting upon related past studies are first presented. Articles of par
ticular interest for one component or connection type are presented under one of the
following headings: connections and details, brickwork, cinder block and clay tiles,
stonework, wooden framing, glass curtain walls and sandwich panels.

5.1 Connections and details

A review of all common types of metal ties and anchors used to connect brickwork to
structural elements has been completed [Grimm, 1976]. Comprehensive lists with
common sizes and dimensions as well as material properties are given for each type of
tie and anchor. Illustrations showing details of the connection between brickwork and
support and included. Several design recommendations for ties and anchors are pre
sented.
The use of stainless steel components in buildings as masonry support systems has
been investigated [Fallon, 1992]. Basic properties of stainless steels are reviewed.
Three current types of stainless steel masonry support systems are reviewed: continu
ous angles, individual brackets and composite bracket/ angle. Good construction de
tails are shown, advantages and disadvantages for each type are discussed.

45
 Connection dtails have been published for me case history of a high-rise steel framed
office building that was designed compositely with the exterior cladding to reduce lat
eral drift [Tornasela, etal., 1986]. A review of different anchor connectors for stone
slabs has been published [Amrhein, eLal, 1990]. Test results for different anchors and
stone qualities are given.
Details of steel frame connections used for the replacement of defective precast con
crete wall panels with brick masonry veneer have been classified [Cowie, and Ameny
1990].

5.2 Precast concrete cladding

Precast concrete units are typically used as exterior cladding on steel framed buildings.
Such units may be reinforced or prestressed, and constructed using normal or light
weight concrete.
A review of previous studies which examine the effects of such non-stractural ele
ments on the service behaviour of steel framed structures has been published [Smith,
Gaiotti, in 1990]. A reference list is given containing 12 studies between the years
1973 and 1989. It is stated that these studies clearly indicate, theoretically and experi
mentally, that non-stractural precast concrete units significantly contribute to the in
service behaviour of steel framed buildings. Further, it is stated that
- The positive effect of such cladding of lateral building stay is typically ignored
(conservative).
- The effect of forces on the cladding and their connections are typically ignored
(non-conservative).

Design criteria for precast cladding and typical connections between the steel frame and
cladding are reviewed. A calculation procedure given in this paper, including both
structural and non-stractural elements. The conclusions of this paper are as follows:
- Usually specified connections between the steel frame and precast concrete units in
North America do not isolate the cladding. This implies that the cladding and
structure act together under lateral load conditions.
- A significant potential exists to use such cladding to brace steel framed buildings.
- A design procedure is proposed to include such panels in a structural analysis.
- Use of this procedure has shown that building lateral deflections may be reduced
by between 36% and 68%.
- Cladding failure using existing construction practices occur not in the precast units
but in the connections between the units and the steel frame.

5.3 Brickwork
Review of selected references
Brickwork is commonly used as exterir cladding in steel-framed buildings. Some use
of brickwork as interior partitions, however, may be found. Two distinct types of

46
brickwork are commonly used in steel-framed buildings. These are referred to as fol
lows:
Cast in place. This type of brickwork is normally placed by hand, brick by brick.
Brickwork is supported by the concrete floor slabs or on a steel section which is at
tached directly to the steel frame. The joints between bricks normally consists of stan
dard mortar. Such mortar contains cement, lime and sand or just cement and sand.
Mortar is often referred to by the ratios of each material in the resulting mix.
Precast. Precast brickwork is generally supported by stractural steel angles or
channels. These angles are directly attached to the steel frame. The joints between
bricks normally do not consist of standard mortar. This is because normal mortar will
not withstand the forces and vibrations associated with moving the finished panels to
the job site. Special mortars consisting of polymer glues are often used.

Brickwork is often placed outside the last row of columns. Tie backs (thin metal
strips) to the steel frame are then used to inhibit the brickwork from moving out-of-
plumb. This provides an exterior wall which is architectarally pleasing, as it is free
from interruptions. In such cases, however, the only direct shear connection between
the brickwork and stractural frame is on the bottom (support) edge. When brickwork
is placed in the same plane as the row of steel columns direct shear connections be
tween the vertical edges of the brickwork and the steel frame may be present The col
umn to brickwork details must be carefully examined to ensure that no gap has been
left between column and cladding.
The influence of 16 variables on the behaviour of brick walls has been published
[Grimm, 1975]. Influences (given in the form of equations) are based upon the cumu
lative results of several past studies. Of particular interest to this review, lateral shear
force and stiffness was included in this review. The following general remarks and
observations about the shear strength and stiffness of brickwork were made:
- Shear strength is a function of me bond strength of the mortar and the frictional re
sistance at brick-mortar interface (itself a function of applied axial load and the ori
entation of the mortal joints with respect to the applied load).
- Shear failure, however, may be due to tensile failure of the mortar.

Methods of predicting the shear strength of a brickwork are proposed. Two formulas
for general use are proposed as follows (one without axial load, one including axial
load). Without axial load, the strength of brick masonry may be predicted based upon
diagonal tension in split disks, as follows:

Fdt = kVfm

in which:
Frit is the shear strength of brickwork in diagonal tension as measured using 15 in di
ameter split disks with bed joints at 45 to the applied load, in lb/in2

47
k is an experimental constant (values are normally between 2.5 and 4.5, average =
3.5)
fm is the compressive strength of brick masonry prisms, in lb/in2

With axial load, the shear strength of brick masonry may be predicted using the fol
lowing expression:

fs=1.4fb + <i>fp
in which:
fs is the shear strength of brickwork, in lb/in2
(fp must be greater than Q
fb is the bond strength of the mortar to brick, in lb/in2
is the coefficient of internal friction (values are between 0.6 and 1.33, average =
0.68)
fp is die axial compressive stress, in lb/in2

It is stated that lateral shear tests of full size brick walls are difficult to perform.
Several methods are currently in use. Several types of small-scale tests are also widely
used.
The modulus of rigidity of brickwork increases with compressive strength. Typical
values range from 1030 N/mm2 to 8960 N/mm2. A formula for general use is given
for computing me shear stiffness of brickwork, as follows:
E
Ev =
2(1 + )

in which:
Ev is the shear modulus of rigidity of brickwork
E is the tangent static modulus in compression of brickwork (values given for
brickwork range from 3300 N/mm2 to 18300 N/mm2).
is Poisson's ratio (values for brickwork range from 0. 1 1 to 0.2)

A good literature review is available at the end of this article, containing 79 references,
most of which are dated between the years 1955 and 1975.
The load capacity of brickwork has been investigated, when subjected to a diagonal
compressive load combined with a compressive edge load acting in the plane of the
wall and normal to the direction of the mortar joints [Yokel and Fattal, 1976]. 32 wall
specimens were tested, and three general failure modes were observed. These are the
following:
- separation along mortar joints
'- splitting about the axis of principle stress
- splitting in the approximate direction of the principle stress.

48
Failure modes due to crushing of the bricks themselves or mortar joints spUtting in the
direction normal to the plane of the wall was observed. Theoretical failure hypotheses
for this type of loading and failure modes were developed. Three failure hypothesis
were examined, as following:
- failure by critical normal stress.
- failure by a critical biaxial combination of normal principle stresses.
- failure at a critical in-plane tensile strain.

The major conclusions of this study are as follows:

- Failure under combinations of compressive diagonal and edge loads can occur by
joint separation or splitting. Failure modes change from joint separation to joint
sptitting as the compressive edge load is increased.
- An equation is developed which predicts failure load when joint splitting is ob
served. This is the following:

0 = +

in which:
ic is the nominal shear stress at failure
is the shear strength at = 0.
is the nommai vertical stress
is a coefficient thought to be related to friction
- When failure is caused by splitting, the failure load may be predicted assuming that
the failure originates in the centre of the panel and is caused by a critical combina
tion of principle stresses. The load combination must, however, be within the fol
lowing range:

-l<^>-5
Xd

where:
td is equal to (0.707)Pd/bt
Pd is the diagonal load capacity when Pv = 0 (measured from tests without verti
cal edge load),
b is the panel width
t is the panel thickness
Pv is vertical edge load.

A state-of-the-art report covering three types of masonry walls: brickwork, clay

tiles and cinder blocks has been published [Fattel and Cattaneo, 1977]. This article
is organized as follows:
- A general introduction is given concerning common types of masonry and their
uses in wall systems.

49
- A review of previous tests and test results is presented. This includes tables of test
results for compressive strength, shear strength, tensile strength in flexure, mod
ules of elasticity and the shear modules of rigidity.
- Sampling and test methods are discussed.
- Limit states for masonry walls are described under combined loading conditions.
- Several appendices as included at the end of this reference which do not have a di
rect bearing upon this study.

Testing types and conditions for the lateral shear strength and stiffness of brickwork
are given. Tables of test values for the failure load are also included. The influence of
openings and their inclusion in design formulae are discussed.
It is stated that barring premature compressive crushing of the masonry at the diago
nally loaded corners, shear specimens will fail along the loaded diagonal by shear
cracking in the mortar joints, or by tensile cracking of the bricks, or by a combination
of both. A formula is proposed which included these modes of failure, as follows:

in which:
fv is the masonry shear strength
Pu is the diagonal compressive load at failure (measured from axially loaded speci
mens)
A is the approximate cross-sectional area parallel to lhe mortar joints
u is a coefficient representing the influence of the compressive shear strength
(between 0.3 and 0.5)

A substantial bibliography, with 81 entries, is included at the end of this article. Most
articles in this reference list are between the years from 1950 to 1977.
The results of an analytical technique to predict the failure load of a brick shear walls
subjected to compression and in-plane shear forces has been published [Samarasinghe,
etal, 1981]. The failure criteria used was the following Mohr-Coulomb relationship:

= 0 +

in which:
is the average shear stress at failure
to is the initial shear bond strength
is the coefficient of friction
oh is the average normal stress at the joint

The underlying hypothesis for the use of this failure criteria is that the failure load is
not sensitive to the stress distribution within the panel. Such distributions can, how
ever, be influenced by wall geometry, boundary conditions and load application. As a

50
result, model predictions must be used with caution when changing any of the previ
ous values. The conclusions draw from comparisons with full-scale in-plane shear
tests are as follows:
- Failure initiation can be reasonable predicted using this failure criteria.
- The shear strength of brick panels does not seem to increase linearly with applied
axial load. Non-linearities as of the order of 10%.

An experimental program has been conducted consisting of six masonry infilled steel
frames [Dawe and Mcbride, 1985]. The effects of joint reinforcement, wall-t-column
tie systems and the use of bond beams were evaluated. Results of analytical methods
suggested by other researchers are suggested for predicting pre-crack stiffness, and
ultimate load carrying capacity are presented. An increase in ductility and ultimate load
of the brickwork due to me steel frame was cor_firmed.
The previous study was further investigated [Dawe and Young, 1985]. Here the ef
fects of openings (such as doors and windows), bond and friction between panel and
frame, airspace between roof beam and panel, and column-to-panel ties were investi
gated. In addition to full-scale testing, finite element analysis were made. In all cases,
initial stiffness (stiffness prior to cracking) was predicted using hand calculations. It
was found that the initial cracking load was reduced due to the presence of openings.
Data sheets have been gathered based upon tests performed under biaxial stress
[Dhana, 1985]. An example is given in figure 5.1. The experimental evaluations are
based upon representative stress-strain relations and a failure surface for solid brick
masonry. The finite element formulation accounts for non-linear deformation, cracking
and sliding of the joints between bricks.

*^~mw PMMtflt Emmm Mhmm lwmc rtin

_ fu_l_i-i-l. Fnat ml
Fnw
_-r !--__. Fram #i

_f

o
-j

Deflection, bb
5.1. Data sheet for brick wall tests.

51
A study of joint failure under lateral shear load was made [Atkinson, etal, 1989].
Shear load-displacement behaviour during static and cyclical loading was investigated.
The conclusions of this investigation are the following:
- Peak and residual lateral shear strengths of brickwork are well represented by the
Mohr-Coulomb failure criteria.
- Residual shear strength does not seem to be affected by the number of shear cycles.
- A joint coefficient of friction of 0.7 can be taken as the lower bound estimate for a
wide range of brickwork types.
- Cohesive joint strength show a wide scatter, reflecting wide variations between
mortar qualities.
- The results of this study can be used to include joints in brickwork in finite element
studies.

Shear failure criteria for masonry joints subjected to axial and in-plane shear force have
been published [Riddington and Ghazili, 1990]. Tests confinning the theoretical model
were also presented. A review of past references is first presented. The Mohr-
Couloumb type failure criteria previously proposed by other researchers is criticised as
it requires the coefficient of friction, , to change as a function of the applied average
normal stress, On. No physical reason for such a change has been found.
Three distinct failure modes are defined: joint slip, mortar tensile failure and brickwork
crashing. It is stated that joint slip is the dominant failure mode with small axial load,
joint compressive stresses less than about 2 N/mm2. The published failure criteria is
applicable for joints under higher axial stress (above 2 N/mm2), and due to mortar
tensile failure. Mortar tensile behaviour is defined as a function of the ultimate local
shear stress and local normal pre-compression stresss. Calculations must be made us
ing the aid of a computer.

Summary
m Brickwork is used structurally in other building types, thus a significant amount of
research has been devoted to examining its structural characteristics. Most of this ex
perimental and theoretical research has been performed to determine the ultimate
strength of brickwork (walls constructed using bricks and mortar). Some work has
been preformed to estimate shear stiffnesses of brickwork subjected to in-plane shear
forces. Brickwork has been examined for a wide range of loading and boundary con
ditions. The effects of openings, such as windows and doors has also been studied.
Little or no research has been directed at defining initial in-plane shear stillnesses,
limiting load levels for serviceability criteria or providing guidance as to maximum
shear deformations that may be imposed prior to unacceptable service behaviour.
Unfortunately, little work has been performed examining real boundary conditions
between brickwork and steel-framed structures. Some work has been done to classify
the stractural behaviour of common components used to connect brickwork to load
bearing frames.

52
5.4 Cinder blocks and clay tiles
Review of selected references
Cinder blocks (also referred to as concrete masonry) are made from cement and stan
dard or lightweight concrete aggregate. Cinder blocks are classified according to their
dimensions but are normally between 30 and 40 cm long, 15 to 20 cm height and 5 to
15 cm thick. Cinder blocks normally have one or more interior openings. If the ratio of
interior openings to the total net area exceeds 25% concrete block are referred to as
open and are usually non-load bearing. Clay tiles are similar to cinder blocks but nor
mally have smaller wall thicknesses. This results in substantial weight savings and in
creases their thermal insulation properties. The joints between blocks normally con
siste of standard mortar consisting of cement, lime and sand or just cement and sand.
Non-load bearing cinder blocks and clay tiles are typically used in steel-framed build
ings as interior partitions. When used as such they are often left exposed and provide
excellent thermal and fire resistance. Being less aesthetically pleasing than brickwork,
such materials are not frequently used as exterior cladding. When used on exterior
walls, however, they are often covered by other claddings.
Both cinder blocks and clay tiles are normally layed by hand directly on a concrete
floor slab. Often partitions are located in the same plane as rows of columns. If the
vertical edges of the partitions fit tightly against the columns direct shear connection
with the steel frame may be assumed. As with brickwork, column to partition details
must be carefully examined to ensure that no gap exists. Stracturally, cinder blocks
behave similarly to brickwork, and like brickwork, are commonly used as stractural
components in other building types. As such their stractural behaviour has received
some interest in the literature. As is the case for brickwork, when estimating the shear
behaviour of cinder block and clay tile partitions, openings (such as doors and win
dows) and joints (such as expansion joints or construction joints) must be taken into
account

A state-of-the-art report has been published which covers three types of masonry
walls: brickwork, clay tiles and cinder blocks [Fattoi and Cattaneo, 1977]. The con
clusions of a study on brick were found to be applicable to burnt clay and masonry
irtfill walls [Dawe and Mcbride, 1985].
Shear tests have been conducted on concrete masonry walls [Hegemier and
Krishnamoorty, 1977]. This research is part of a larger effort to develop a basis for
predicting the earthquake response and damage to concrete masonry walled stractures.
A number of different tests are reported, starting with simple tests and culminating in
biaxial loaded panels subjected to quasi-static and dynamic loads. Test results include
information on both strength and stiffness. No design formulae are proposed. No in
formation on the initiation of cracking is provided.

53
 Summary
u The literature search indicates that a significant amount of experimental and theoret
ical research has been performed which may be used to determine the ultimate strength
of cinder block walls or clay tile walls. This is largely due to the similarity of such
walls with brick walls. Little work has been performed specifically for cinder blocks
or clay tile walls.
Little or no research has been directed at defining initial shear stiffnesses, limiting
load levels for serviceability criteria or providing guidance as to maximum shear de
formations that may be imposed prior to unacceptable service behaviour. Little work
has been performed examining real boundary conditions between brickwork and steel-
framed stractures.
Existing research suggests that cinder block walls and clay tile walls may be easily
included in finite element calculations for steel-framed stractures at working load lev
els. At service load levels cracking should be greatly limited. Due to the lack of work
examining real boundary conditions between brickwork and steel frames, care must be
taken when connecting steel frame elements and brickwork elements. Analysing the
results of the finite element calculations to determine whether or not service problems
will occur (cracking in the brickwork or excessive lateral movements of the stracture)
may pose a problem when using existing knowledge.

5.5 Stonework
Stone and mortar walls are normally used for their aesthetic value as exterior cladding.
They are far more expensive than brickwork, cinder blocks or clay tiles, thus their use
is limited.
A study of safety factors for thin stone curtain walls for buildings has been published
[Tawressey, 1990]. It is stated that even though the use of such cladding is wide
spread in the United States, no national standards exist In this review, no direct men
tion is made to in-plane shear forces.

5.6 Wood framed diaphragms

Review of selected references
Wood framed diaphragms consist of a large variety of materials and construction
methods. In general, most consists of vertical studs, with occasional cross-bracing,
covered by sheathing (plywood or gypsum sheets). Connections between sheathing
and studs are normally made in one of two ways: by nailing or gluing. For most
wooden diaphragms horizontal wooden studs are provided at the top and bottom of the
frame (sometimes referred to as headers or sole plates). This increases the shear resis
tance of the frame and facilitates connecting the frame to floor and/or roof.
In steel-framed buildings wooden diaphragms are often used as interior partitions.
These partitions are normally placed in the same plane as columns. Such partitions

54
rarely extend from floor to roof but stop short of the roof, above the level of hung
ceilings. The details of connections between column and partition must be studied
carefully to ensure that no gap exists. If gaps exist, the only direct connection between
the wooden diaphragm and the steel frame is at the floor slab.

A state-of-the-art review on wood framed diaphragms for the ASCE select committee
on wood of the stractural division was published [Carney, 1975]. This reference is
highly recommended when reviewing all prior wood diaphragm research. The purpose
of this review was to provide engineers with lists of design related documents for
wood diaphragms (in the form of a bibliography) and to give key research documents.
No design formulae are given in the review itself. The references in this state-of-the-art
review list is considerable, with more than 120 entries, which are generally more re
cent than the mid 1960' s. A incomplete list of subject information contained in this
bibliography is listed as follows:
- The effect of connections between plywood and frame (rigid or non-rigid) in rect
angular diaphragms.
- Strength and stiffness of roof diaphragms with different percentages of decking
edge glue.
- The effect of differential deflections in shear diaphragms (for roof panels)
- The design of diaphragms consisting of transverse boards, diagonal boards, double
diagonal boards or plywood.
- Non-rectangular flexible diaphragms.
- Diaphragm action of plywood panels in combination with fiberboard sheathing, or
from one plywood panel in combination with fiberboard sheathing, or from two or
three separate plywood panels per wall, regardless of other sheathing.
- A review of the internal action of timber diaphragms, together with a study of the
stresses caused by curving of sloping portions of the diaphragms to conform to
common to common truss shapes.
- Deflection formula and useful design curves for determining maximum allowable
deflections of walls of various types of construction.
- Test method for complete diaphragm assemblies, including framing.
- Test method for complete shear-waU assembHes, melding framing.
- Methods for evaluation of sheathing materials on a standard wood frame.
- A full-scale house racking test results.
- Design methods for timber diaphragms.
- Full-scale test of housing modules.
- Shear tests of plywood sub-floors and gypsum ceilings for residential housing.
- Lateral tests of wooden roofing with decking.
- An analysis of a shear wall and wood frame building.
- Diaphragm action of diagonally sheathed wood panels
- The effect of openings in roof diaphragms.

55
- Test results from diaphragms made of many types of wood, new fastener types and
spacings.
- The adequacy of diaphragms with long-span framing.
- Static and cyclic tests of horizontal roof diaphragms made of plywood, chip[board
and flaxboard.
- Comparisons of the racking behaviour of many different types of wooden wall
construction.

Data sheets have been gathered from tests performed on wood shear panels bonded
with flexible adhesives [Richards, etal., 1975]. These are shown in figure 5.2. The
goal of these tests was to develop and partially substantiate a computer program for the
stress analysis of the panel.

3/4" Plywood Sheathing

2- by 4- inch
Framing Mm.r

Experimental Panel (1 ft = 0.305 m; 1 in. = 25.4 mm)

0 ^--WIO-X 4

-nett STOP'
MACH

LOCK
ICH
\
TIE MOD-
COM)
CELL -0*0 BLOCK

__

- 10 ION HrOM-UC JACK W XI90

5.2. Data sheet and schematic for wood panel tests.

56
The results of an investigation on the shear strength of rectangular light-frame nailed
walls has been published [Tuomi and McCutcheon, 1978]. Comparisons are given
between a method used to predict shear strength and test results. The proposed for
mula is as follows:

R = r-[(Kn + Km)p + (a2Kna + b2Knb + a2^ + b2Kmb)f]

in which:
R is the theoretical racking load
r is the lateral nail strength (one nail)
a is the ratio of heights of interior to rjerimeternail patterns
b is the ratio of widths of interior to perimeter nail patterns
f refers to interior nails
refers to perimeter nails
Kn is the racking coefficient of for horizontal nail spacings (given in table form)
Km is the racking coefficient of for vertical nail spacings (given in table form)
I-na is the racking coefficient for horizontal nail spacings, adjusted for the height of
the interior nail pattern (given in table form)
Knb is the racking coefficient for horizontal nail spacings, adjusted for the width of
the interior nail pattern (given in table form)
Kma is the racking coefficient for vertical nail spacings, adjusted for the height of the
interir nail pattern (given in table form)
Kmb is the racking coefficient for vertical nail spacings, adjusted for the width of the
interior nail pattern (given in table form)

The major conclusions of this study are as follows:

- Good correspondence is obtained between model and test strengths. This model
can not be used to predict shear stiffness, however.
- The shear strength of a wooden diaphragm may be predicted based upon panel ge
ometry, the number and spacing of nails, the lateral resistance of one nail and a
small scale test Using only small scale tests is much more cost efficient than per
forming full-scale tests.

Simplified formulas have been proposed, based upon previous work on profiled steel
sheeting, for the design of wood frame shear walls with sheathing attached by nails or
other types of discrete fasteners [Easley, 1982]. Tests were performed to provide a
comparison with formula predictions.
The formula predicting initial shear stiffness is as follows:
r- 1
G =^b
Kw + Gt

57
in which:
G ' is the linear shear wall stiffness
b is me panel length between end fasteners rntre lines
K is the constant in the linear load-slip relation for fasteners (obtained for load-slip
curves of
nail joint tests)
w is the panel width between side fastener centre lines
G is me shear modulus of elasticity of the sheathing material
t is the thickness of the sheathing panels
is me panel attachment modulus, equal to the following:

=%+_____2
w

in which:
ns is the number of side fasteners at each side
nSj is me number of fasteners in each interior stud
le is the second moment of inertia of end fasteners
Is is the second moment of inertia of interior stud fasteners

The maximum shear force may be obtained by equating the maximum experimental
values obtained from nail test results to the lower value predicted by the following two
equations:

-_ Nb
Fs=T
or;

"e w

in which:
Fs is me maximum side fastener force
is the shear force per unit length on the shear wall
Fei are the resultant end fastener forces
Xei is the distance from the panel centre line to the end fastener
% is the number of end fasteners at each end

The major conclusions from this study, appropriate for this review, are as follows:
- The stiffness formula provides a good approximation of the initial slope of the load
deflection relation for a typical shear wall.
- The allowable design loads for shear walls can be based on allowable values for the
maximum force in the wall sheathing fasteners.

58
- The afore mentioned formulae should be applicable to shear walls of any size
sheathed with fibreboard, metal, or other types of sheathing, and those using
screws, staples or other types of discrete sheathing fasteners.
- The formulae presented are valid only if no separation occurs in the framing mem
ber joints between the studs and the header of sill when the wall is loaded.

A numerical model has been developed to predict the shear strength of wood-stud
shear walls [McCutcheon, 1985]. The method is based upon energy methods and in
cludes nail load-slip relationships and linear deformations due to shear distortions of
the sheathing material. This method can not be used for predicting lateral deforma
tions, however. The major conclusions of this study ar as follows:
- Lateral nail tests or small-scale racking tests can be used as the basic data needed to
predict the shear performance of this type of wall system.
- A simple power curve, f(x)=A-B, representing nail non-linearities give good re
sults in this model when compared with test results.

A numerical model has been developed for predicting the lateral shear behaviour of
wood framed houses [Gupta and Kuo 1987; two references]. This model includes up
lift in studs, a parameter not included in previous models. The most important conclu
sions from this work are as follows:
- If uplift does not occur the nail force-slip relationship has the strongest influence
over lateral behaviour.
- Uplift can greatly decrease the lateral stiffness of wooden shear walls. The effect of
dead loads can partially or fully compensate for this loss.
- For very long walls the effect of uplift on shear stiffness is greatly reduced.
A mathematical model has been presented for analysing partially composite wooden
diaphragms [Gutkowski and Castillo, 1988]. These walls consisted of a stud frame
with sheathing (plywood, wood or gypsum board) placed on one side or both sides. A
non-linear partial connection is assumed between studs and sheathing. Data sheets
from this reference are shown in figures 5.3 to 5.5. Major conclusion from this article
are summarised as follows:
- The model performed well even in non-linear regimes.
- The model is dependent upon determining the behaviour of the connection between
stads and sheathing, usually due to nails.
- The model may be extended to include buckling of studs and buckling of the
sheathing by identifying limiting stress levels in each component

The static and dynamic characteristics of wood framed shear walls have been examined
[Rliatrault 1990]. Simple stractural analyses are presented with predict the following:
- The stiffness of such walls subjected to lateral static load
- The maximum static lateral load carrying capacity
- The stractural response due to earthquake excitation.

59
 .s.f
^
^
.;<?'' :

,V ft
% 'i'!

tud tr mmm '.H'

tr* _:"-""*
\ >

j eoo

# (Ini fttue cun) T ttttTm

_ (l-_) < I' furmij W*E_S

O. A O. 1 .

-fleetien D. In.

5.3. Data sheet and schematic for wood-gypsum panel tests.

These analyses are contained in a computer program developed for micro-computers.

It is further stated that wood framed shear walls have been developed in the past ten
years to the point where they are commonly used to provide lateral stiffness and
strength in wood framed stractures. It is thought that the computer analyses could be
easily incorporated into a larger stractural analysis.

Summary
Wooden diaphragms are widely used for the construction of residential housing.
As such, it has received a significant amount of attention in the literature. Most of this
experimental and theoretical effort has been directed towards estimating ultimate
strengths. Little work, until recently, has been concentrated on estimating shear stiff
nesses or deformations due to in-plane shear forces.

60


O
J

200

5.4. Data sheet and schematic for wood-gypsum panel tests.

Wooden diaphragms, in general, have been examined for a wide range of loading
and boundary conditions. The effects of openings, such as windows and doors has
also been studied. Major factors influencing the shear strength and stiffness of wooden
diaphragms have been found to include the following:
- The shear stiffness of the sheathing.
- The force versus slip relationship of the connection between sheathing and studs
(nails or glue).
- The presence (on absence) of uplift at the bottom of each frame. If the frame is
long, uplift becomes much less important
- The presence of openings in the frame, such as windows or doors.
Little or no research, however, is directed at defining limiting load levels for ser
viceability criteria, such as cracking, or providing guidance as to maximum shear de
formations mat may be imposed prior to unacceptable service behaviour.

61
 _*""
' - :\y:
\ :

:. "
tud fr-mmm ::* : __*"-*

Af
___-__-_ "'
'ff II
I
TJ
//
O
J
200 i Letead

~- 10 cxpriMnt_l curve

Protra- WA-ELS

O
O O . 2 O. A O. O.
D-41tlan D, in.

5.5. Data sheet and schematic for wood-gypsum panel tests.

Unfortunately, little work has been performed examining real boundary conditions
between wooden frames and steel-framed structures. Some work has been done to
classify the stractural behaviour of common components used to connect brickwork to
load bearing frames.

5.7 Glass curtain walls

Glass curtain walls are used as exterior cladding for medium and high-rise steel-
framed buildings. No data was found in the literature concerning their structural be
haviour under lateral loading. One general reference, including other stractural aspects
of glass curtain walls, was found.
A series of articles conprning the use and design of glass curtain walls has been pub
lished [Glass in Building Design and Construction, 1967]. The topics covered in these

62
 articles are design, testing, practice, in-service experience. Tolerances for movement
of the frame relative to the glass panels are discussed. Most stractural emphasis is
placed upon wind loads and proper support of the glass in the frames. Shear testing
and shear stiffnesses are not covered.
An overview of recent developments in glass curtain wall design for high rise build
ings have been published [Gartner, 1990]. Typical edge details are shown but no in
formation is given about lateral movements or maximum in-plane shear strength.

5.8 Profiled steel sheeting

Profiled steel sheeting normally are formed using thin-walled material, 0.75 mm to L5
mm in thickness. The profiling process normally is undertaking in continuous cold-
role forming machin. At the end of this process panels consisting of a few ribs
(between 2 to 6), are cut to length and bundled for shipment to the construction site.
Profiled steel sheeting is commonly used as exterior cladding on walls and roofs of
steel framed buildings. Profiled steel sheeting is not commonly used as interior parti
tions.
A substantial body of knowledge concerning the structural behaviour of profiled steel
sheeting is available in the Hterature. It is not the intent of this report to review this lit
erature in a systematic manner, as the results of this knowledge have been included in
existing design recommendations and codes. These documents are freely available to
designers, and have been in use for a number of years.
A review of deformation limits for profiled sheeting has been published [Douhan,
1980]. In this review me functional stability of attachments subjected to lateral forces
were included. Swedish and other national practices are reviewed.
A series to tests have been performed on profiled steel sheet wall subjects to shear
forces, loaded monotonially and cyclically until collapse have been prepared
[Italsider] . Data sheets from this test program are included in figures 5.6 and 5.7.

5.9 Sandwich panels

This is a relatively new category of stractural and non-stractural element consisting of
two parts: exterior thin walled steel sheets and a lightweight core. Sandwich panels are
gaining wide acceptance as an economical solution for both exterior and interior usages
with steel-framed stractures.
Recommendations for the structural use of sandwich panels in steel-framed stractures
are currently being drafted. Due to a lack oftest data these recommendations will most
likely suggest that only the exterior steel sheet directly connected to the sandwich panel
be taken into account Design recommendations for this steel sheet will be referred to
the appropriate sections of the ECCS recommendations for stressed skin design.
In plane shear tests on sandwich panels have been conducted, however, published re
ports are not yet publicly available. No study of the use of sandwich panels as non-
stractural elements in a steel framed structure are known.

63
 1

i
20
SO: i\_
i
/\ A _- _- J_
_J7_

I 0 400 1
t

/
1

% ve
bf sagoma
Ht 300
f * \:!
Pl
ir.
.
y
/ -J S
4 L__ sezione S-S
-ti
i
I

1 i
autottenantt
PN 120
CI cucitura

ts.s
I schema di assembiagge

dl tesaggi.

5.6. Data sheet and schematic for simple steel sheeting tests.

5.10 Conclusions
Existing codes acknowledge mat a buildings lateral response is in part governed by the
type of non-structural components used. Limits on lateral movements are defined ac
cording to standard uses of non-stractural components. Very general lateral drift limits
have been adopted that are applied to all buildings regardless of construction type.
For each of the following subjects the present state-of-the-art is summarised and sug
gestions for further work are presented.

Attachments between the steel frame and non-structural element. Few references
are available on this subject Standard details should be classified according to their
structural behaviour (strength and stiffness). Recommendations should be made
defining minimum acceptable structural behaviour. This could be done for two cases:
- Attachments isolating non-structural components from the structural system. Such
attachments should allow a pre-defined minimum rotation to occur before signifi
cant lateral force is transferred to non-stractural components. Strength requirements
may be based only upon the weight of non-stractural components.

64
 .
/ /_.-.
/ Ux42

e- -

ii

;? _ />
<jT

5.7. Data sheet and schematic for simple steel sheeting tests.

- Attachments for non-structural components that participate with the structure at

service load levels. The initial stiffness of the attachment should be significant The
strength of the attachment should be sufficient to carry the weight of the non-
stractural components and fail only after significant serviceability problems have
occurred in non-stractural components.
In-plane shear stiffness of non-structural elements.Jistimates of initial in-plane
shear stiffness are needed for all types of non-stractural components. Initial in-plane
shear stiffness equations need only characterise the initial stiffness of cladding and
partitions. Cracking, plastification, out-of-plane deformations, and other problems all
occur at load levels that exceed serviceability limits states. Numerous existing test data
can be used to estimate such values. In most on these tests, however, researchers ex
amined only the maximum load carrying capacity and failure modes at ultimate limit

65
states. Further, guide-lines could be put fourth for estimating the effects of small and
large openings.
Definition of load limits or imposed shear deformation limits. The maximum ap
plied load or deformations that each type of non-structural element may withstand be
fore serviceability problems may be expected remains to be explored. No direct infor
mation was published on this subject Many tests have been performed but have yet
been analysed with respect to serviceability. It must be noted that not all non-stractural
elements need be included in calculations at service load levels. Adequate deformation
characteristics, however, are necessary in all cases to ensure adequate serviceability.

References
LE. Amrhein, R.H. Hatch and M.W. Merrigan, Anchor connections of stone slabs, Fifth North
American Masonry Conference, June 1990, p. 1417-1424.

R.H. Atkinson, B.P. Amadi, S. Saeb and S. Sture, 'Response of Masonry Bed Joints in Direct
Shear', Journal of the Structural Division, ASCE. VoL 115, No. 9, February 1989, p. 2276-2296.

JJM. Carney, 'Bibliography on Wood and Plywood Diaphragms', Journal of the Structural Division,
ASCE. Vol. 101, No. STI 1, February 1975, p. 2423-2436.

J.W. Cowie and P. Ameny, Replacement of defective precast concrete wall panels with brick masonry
veneer without relocating building tenants, Fifth North American Masonry Conference, June 1990, p.
1371-1384.

JJL. Dawe and R.T. McBride, Experimental investigation of the shear resistance of'masonry panels in
steel frames, Proceedings of the 7th international brick masonry conference, Melbourne Australia.,
February 1985, p. 791-801.
JJL Dawe and T.C. Yong, An investigation offactors influencing the behaviour of masonry infillin
steelframes subjected to in-plane shear, Proceedings of the 7th international brick masonry conference,
Melbourne Australia, February 1985, p. 803-814.

Dhana, Page and Kleeman, The behaviour of brick masonry under biaxial stress with particular refer
ence to infilled frame', Proceedings of the 7th international brick masonry conference, Melbourne
Australia, February 1985.

L. Douhan, Deformation limits for roofs and walls of profiled sheeting, Document D32:1980,
Swedish Council for Building Research, Stockholm, Sweden, 1980.

J.T. Easley, M. Foomani and R.H. Dodds, 'Formulas for Wood Shear Walls', Journal of the
Structural Division, ASCE. Vol. 108, No. ST11, November 1982, p. 2460-2478.

J.B. Fallon, 'Stainless steel in building', Steel Construction Today, Vol. 6, No.l, January 1992, p.
7-11.

S.G. Fattal and L. Cattaneo, Evaluation of Structural Properties of Masonry in Existing Buildings,
National Bureau of Standards, Building Science Series 62, 1977, p. 1-117.

A. Filiatrault, 'Static and Dynamic Analysis of Timber Shear Walls', Canadian Journal of Civil.
Engineering, VoL 17, No. 4, August 1990, p. 43-651.

F. Gartner, Design Aspects of Curtain Walls of High-Rise Buildings, Tall Buildings: 2000 and
Beyond, Fourth World Congress, Hong Kong, November 1990, p. 307-316.

66
CT. Grimm, 'Strength and Related Properties of Brick Masonry', Journal of the Structural Division,
ASCE. Vol. 101, No. STI, January 1975, p. 217-233.
-, 'Metal Ties and Anchors for Brick Walls', Journal of the Structural Division, ASCE. Vol. 102,
No. ST4, April 1976, p. 839-858.

RJM. Gutkowski and AJL. Castlo, 'Single- and Double Sheathed Wood Shear Wall Study', Journal
of the Structural Division, ASCE. VoL 114, No. 6, February 1988, p. 1268-1284.

Glass in Building. Design and Cbnsiracirn, Building Research, May-June 1967.

A.K. Gupta and GJ?. Kuo, 'Modelling of a Wood-Framed House', Journal of the Structural Division,
ASCE. Vol. 113, No. 2, February 1987, p. 241-259.
-, 'Wood-Framed Shear Walls with Uplift', Journal of the Structural Division, ASCE. Vol. 113, No.
2, February 1987, p. 260-278.

G.A. Hegemier and G. Krishnamoorty, An Experimental Study of Concrete Masonry Under Seismic
Type Loading, National Bureau of Standards. Building Science Series 106, September 1977, p. 114-
154.

Italsider, Indagine sperimentale sulla duttilit di pannelliparete in lamiera grecta. Il comportamento

delle strutture portentidi acciaio alle azoini -/smicfe, Monography no. 7.

WJ. McCulcheon, 'Racking Deformations in Wood Shear Walls', Journal of the Structural Division,
ASCE. Vol. Ill, No. 2, February 1985, p. 257-269.

JA. Richards, R.P. Kerfoot and Drueger, 'Wood shear panels bonded with flexible adhesives', ASCE,
Journal of the Structural Division, Vol.101, No, Stl, January 1975, p. 131-149.

JJL Riddington and M_Z. Ghazili, Hypothesis for shear Failure in Masonry Joints, Proceedings,
Institute of Civil Engineers. Vol. 89, Part 2, March 1990, p. 89-102.
W. Samarasinge, A.W. Page, and A.W. Hendry, 'Behaviour of Brick Masonry Shear Walls', The
Structural Engineer, Vol. 59B, No. 3, September 1981, p. 42-48.

B.S. Stratford Smith and R. Gaietti, Effect of Precast Concrete Cladding of the Behaviour of Tall
Buildings, Tall Buildings: 2000 and Beyond. Fourth World Congress, Hong Kong, November 1990,
p. 317-332.

J.G. Tawressey, Factors of safety for thin stone in building curtain walls,
Fifth North American Masonry Conference, June 1990, p. 1417-1424.

RJL. Tomasetti, A. Gutman, LP. Lew and LJvL Joseph, Development of Thin Wall Cladding to
Reduce Drift in Hi-rise Buildings, IABSE Colloquium on Thin-Walled Metal Structures, Stockholm
1986, p. 239-246.

RJL. Tuomi and WJ. McCutcheon, 'Racking Strength of Light-Frame Nailed Walls', ASCE, Journal
of the Structural Division, VbL 104, No. ST7
July 1978, p. 1131-1140.

F.Y. Yokel and S.G. Fattal, 'Failure Hypothesis for Masonry Shear Walls', Journal of the Structural
Division, ASCE. Vol. 102, No. St3, March 1976, p. 515-532.

67
Structural modelling and calculation
methods at serviceability load levels

Steel frame buildings are complex spatial systems composed of linear elements (beams
and columns) and plane elements (floors, diaphragms, partitions and cladding). Figure
6.1 provides a schematic representation of all these components with reference toa
typical building with a steel braced skeleton. The stractural response is the final result
of the interaction between the bare frame and the 'non stractural elements' .

Column
protection

concrete

Bracing system cladding

Profiled composite decking - , *. _ _-
r ^^ Sprayed fire protection

6.1. Schematic of a typical building with a steel braced skeleton.

The design approach is generally based on hypotheses which reduce the scheme of the
building to systems composed of bi-dimensional skeleton frames. Beam-to-column
joints are traditionally considered hinges or rigid joints and the design is thus based on
the simplified model of simple and rigid frames respectively. However, the sophistica
tion of the computing tools nowadays available to the practising engineer has led to an
increasing refinement of design analyses, which is more and more exploited, due to
the strong competition among different structural materials. The relationship existing
between the degree of refinement of the model adopted and the performance required
of the model is clearly recognized for ultimate limit states, yet guidelines for a practical
appraisal of this relationship with reference to serviceability are lacking. This condition
is clearly reflected by recent stractural Codes [AISC-LRDF, 1986; Eurocode 3, 1992;
Eurocode 4, 1992], still based on the traditional design philosophy.

68
 This situation is a heavy burden for the design, with reference to steel and steel-con
crete composite frames, for which the current trend towards lighter systems makes
serviceability increasingly important It underlines the significant imbalance in the de
sign quality for the ultimate conditions and for the serviceability conditions.
The research work, carried out at the University of Trento (in the framework of the
ECSC research programme focussing on the static deflection of steel framed build
ings), is aimed to evaluate the influence that the model used in the stractural analyses
has on the frame response at the serviceability load levels. A numerical analysis has
been developed with the aim of focalizing the influence of the beam-to-column joints
and the cladding action on the frame response. An experimental study has also been
performed with the aim of evaluating the interaction between shear and bending mo
ment on the joints in beams with partially restrained ends. The main details related to
die design of the specimens and to their behaviour are then briefly presented.

6.1 Numerical study

The numerical study was performed using a finite element program [Poggi, 1988].
The approach allows a second order plastic zone analysis of partially restrained frames
and the beam-to-column joints can be modelled as inelastic springs (figure 6.2). The
particularly efficient formulation adopted permits to model a beam and two attached
joints by means of a single joint-beam element; the finite dimensions of the joint are
token into account via me rigid end element

m t
W

%
.rigid rigid
X
S

t t
s
- l
"o A^A1 B == _
B _)
A A1 3

6.2. Modelling of the beam-to column joints as inelastic springs.

The program enables analyses to be carried out which incorporate both geometric and
material non-linearities. All loads are increased proportionally through a common load
multiplier ot up to collapse.

69
The numerical study of the behaviour of semi-continuous steel frames was focussed
on the parameters: (a) the type of the steel frame, (b) the load conditions, (c) the joint
action and (d) the cladding action.
a. As reported in figure 6.3, two different types of bi-dimensional sway frames,
which can be thought as cut off from spatial stractures, were considered: frame A
(one span and three storeys) and frame (three spans and two storeys). Cross-
sections for the elements are HE 180 and IPE 330 for columns and beams re
spectively. The choice of these profiles is related to a 'strong beam-weak column'
design approach. Initial imperfections in the frames, due to residual stresses and
mechanical tolerances, are considered in accordance to Eurocode 3. Some prelimi
nary analyses were performed on the frame 1, Le., frame with a system to pre
vent horizontal displacements.

o -o TRUSS SniOTATDJG
F .immun. C"*"*"6 ACTION aF mffm, IIIHHIIB

y y
^_tq
g, " gq
cFlU_l_____| aPliiniiiiinlifmiiiiuliiiiiiiiui
y
y
aF |H____3 1

-b
_.
/
/
_=_. ,
FRAME
1p= frame imperfection BEAMS : IPE 330
IFRAME A| according to EC3 COLUMNS : BE 180

SWAY PRAMF NO-SWAY FRMRI

I FRAME Al I FRAME Bl I FRAME Bll

6.3. Two different types of bi-dimensional sway frames.

b. Each span of the frame is loaded with an uniform vertical load q (equal to 40.0
kN/m); meanwhile each storey of the structure is stressed by a horizontal force
simulating wind action. Different load conditions, functions of the parameter
(ratio between the horizontal force and tibe total vertical load for each storey), were
considerd. The value of ranges from 0.0125 (1/80) to 0.1 (1/10) in order to
cover a wide range of practical applications.

70
 c. The behaviour of the beam-to-column semi-rigid joints was taken into account via a
non linear rotational spring in order to take into account the stiffness and strength.
The column base joints were considered fixed joints.
d. The claddings action is simulated through a truss element in tension [Liauw, 1988];
the equivalence has been established only in elastic field.

Comparison of the results of the analyses performed is made referring to the service
load levels and assuming the serviceability limits given by Eurocode 3, i.e.,:
- maximum horizontal displacements < H/500 (H is the total height of me frame);
- maximum horizontal storey displacement < h/300 (h is the height of the storey)
H/500 and h/300).

6.2 Joint action and frame response

General
The traditional design of steel frames assumes that connections behave according to the
ideal models of hinge and rigid joints. Nevertheless, due to experimental and numeri
cal studies into the behaviour of the beam-to-column joints which have been performed
since 30's, it has been pointed out that this type of connection has a great influence on
the response of steel frames. Joint action, simulated via a non-linear rotational spring,
was investigated considering partial strength joints characterized by different degrees
of stiffness and bending capacity. The ideal cases of hinge and rigid joint were consid
ered in order to make comparisons between the response of the ideal frame and the re-
sponse of the frame with semi-rigid joints respectively.
Joint response and joint action in steel frames were investigated in the last decade
[Anderson et al. 1987; Nethercot and __andonini, 1988; Cosenza et al. 1988: linear
analysis; Cosenza et al. 1988: non-linear analysis]. Current knowledge enables us to
develop procedures for the design of semi-continuous frames as indicated by recent
European and American Codes, which realistically permit an explicit frame design
based on the effective characteristic of the joints, i.e., using the more realistic model of
semi-rigid joint [Eurocode 3, 1992; Eurocode 4, 1992; AISC-LRFD 1986].

Full Fun
strength m strength
Mpl,Bd
1 0
/ Rigid

Partial
strength
0.8-

U.ti-
r Partial
strength
0.4-
Semi rigid s<mi-rigid

0
Flexible Pinned

0.2 0.4 0.6 0.8 1.0 1.2 1.4

Braced Frames

0.2

C
te -Flexible Pinned

0.2 0,4 0.6 0.8 1.0 1.2 1.4 1.6

Unbraced Frame 5
=
JIbf_
LbMpLRd

6.4. Classification of the joint in braced and unbraced frames.

71
The approach used in the Eurocode 3 for the classification of the joint is based on the
comparison between its characteristics (stiffness and strength) and the properties of the
beam. In figure 6.4 the domains for the classification are reported for braced and un
braced frames respectively; the moment of the joint is divided by the moment resis
tance of the beam and the the stiffness of the joint is divided by the stiffness of the
beam.
The non dimensional - curve of the joint in the Eurocode domain for the classifica
tion can give three different cases:
- The rotational non-dimensional response of the joint falls in the rigid part In this
case the rigid frame model cam be sufficiently accurate for stractural analysis and
design.
- The rotational response falls in the semi-rigid region: it is necessary to use the
model of the semi-continuous frame.
- The rotational response falls in the pinned region. The simple frame model can be
used.

Effect of the shear on the rotational behaviour of the joints

Usually the joint response is determined via the sole relation governing the rotational
behaviour in the plane of the joint Recently, experimental studies on shear force-
bending moment interaction in flexible connections were carried out at the University
of California-Berkeley [Astaneh and Nader, 1989]. The tests were performed on can
tilever beam models properly loaded at the free end, in order to simulate different load
histories. The effect of the presence of the shear on - curves of flexible joints
seems important Definitive conclusions require tests on full beam-joint subassem
blages. Due to the importance of joint action in order to achieve a realistic appraisal of
the beam deflection in braced frames, the present study includes a limited series of
beam tests with different end restraints.

_.' L = 6m ,

C 5m
BEAM IPE 300 i = < 0
( 2m

6.S. Specimen for testing flexible connections.

The aim of this experimental analysis is to define the shear-moment interaction for sev
eral forms of flexible connections. Twelve specimens composed of beams with par
tially restrained ends were appositely prepared (figure 6.5). The cross section of the

72
beam is an IPE 300 profile having a span of 6 m, a value of length currently used in
practice applications; the beam was loaded by two vertical forces. Two series of tests
were performed: the first one was characterized by a distance between the load points
equal to 2 m (6 specimens); in one this distance was 5 m (6 specimens). The end re
straints of the beams have been connected to vertical elements reproducing the case of
very stiff columns.

-L 150 150-15*18 -L 150xl50-xi5*18

57

L 100*100*6*10 L 100*100*6fl0

a) b) c)
TOP & SEAT DOUBLE WEB TOP & SEAT ANGLE
ANGLE ANGLE wrra
DOUBLE WEB ANGLE

6.6. Details of the three different types of steel joints tested.

In particular for each series of tests, (see figure 6.6) three different types of steel joints
were considered:
- joints with double web angles;
- joints with top and seat angles;
- joints with top and seat angles with double web angles.

Two different angle thicknesses were adopted for each type of joint appraising the in
fluence of this parameter on the joint stiffness. In order to sinrulate different
shear/moment ratios at the joint two tests were conducted for each beam-joint configu
ration, differing in the value of die distance i between the two applied forces F. The
main mechanical and geometrical characteristics of the speciments are reported in table
6.1. The testing apparatus (see figure 6.5) is composed of triangular frames to which
are fixed very stiff columns. Each of these frames is connected to a concrete block;
- A very stiff beam used in order to apply the load at two load points corresponding
to interaxes of 2 and 5 m.
- Circular hollow elements connecting the vertical elements of the triangular frames.
- A couple of hydraulic jacks. Each of these can give a maximum load of 600 kN.
- A rigid frame which tranforms the load of the hydraulic jacks in a concentrated load
on the very stiff beam.

73
Table 6.1. Main mechanical and geometrical characteristics of the specimens. Bolts M20 are nade of
steel grade 10.9, with failure stress of 1 130 N/mm2.

Thickness Yield stress Bolts M20 10.9

Load dis web top & seat web an top & seat
tance i angle angles gles angles beam column beam
Test (m) (mm) (mm) (N/mm2) (N/mm2) (N/mm2) side side
PB2W9 2 8.4 307 280 8 4
PB2W10 2 10.4 317 280 8 4
PB2TS16 2 16.4 308 280 8 8
PB2TS18 2 18.6 324 280 8 8
PB2TSW9 2 8.4 16.4 307 308 280 16 12
PB2TSW10 2 10.4 18.6 317 324 280 16 12
PB5W9 5 8.4 _ 307 280 8 4
PB5W10 5 10.4 317 280 8 4
PB5TS16 5 16.4 308 280 8 8
PB5TS18 5 .
18.6 324 280 8 8
PB5TSW9 5 8.4 16.4 307 308 280 16 12
PB5TSW10 5 10.4 18.6 317 324 280 16 12

L = 6m
i

F X

j~_- |__-

6.7. Position of inductive transducers.

74
The measuring system has been developed in order to appraise the main parameters af
fecting the behaviour of the specimens. In three cross-sections of each beams four
electrical strain gages were positioned (a couple of strain guages on each flange of the
beam). Inductive transducers were used, according to the scheme reported in figure
6.7. These transducers permit us to appraise the beam deflections and the main pa-
rameters affecting the joint responses.
Twelve tests were performed and the phase of evaluation of the experimental data is in
progress. According to the criteria reported in Eurocode 3 and to the rotational predic
tion methods for flexible connections [Chen and Kishi,1990] the response of the
specimens was estimated. Also the beam-line design method [Zandonini and Zanon,
1991] has been used, modified in order to enable an analysis of beams loaded at two
points. For each joint the straight lines associated with the ultimate limit state of bear
ing for web angle and for top and seat angle and bolt shear were considered with the
serviceability deflection limit states of L/250 (figure 6.8).

, -00.00

. 175.87

' 150.00

50.00
38.71

I I W I I I
0.01 0.0178380.02 0.03 0.04
ROWZIONE ( rod )

: 200.00

. 1757

' 150.00

s_.es
50.00

u I ' 1 1 II I II |[| | || I I I ll| I II I II| I I I I I I I I I jIi In 1 1 1 1 1 1 I I II I I I I

0.00 0.01 0.0170410.0- 0.03 0.04 0.05 0.08
ROTAZIONE ( rod )

6.8. Test results. Above: test PB2W9, with t = 8.4 mm and fy = 307 N/mm2. Under test
PB2W10, with t = 9.9 mm and fy = 317 N/mm2.

75
In each test some cycles Goading the specimen to a prefixed level of load and then un
loading until the zero value of the load was achieved) were performed before the col
lapse of the specimen. A short summary about these tests is briefly presented here.

The first series of tests was characterized by a distance between the load points equal
to two metres. In these tests the collapse of the specimens was due a beam mechanism;
the joints were characterized by a rotational ductility sufficient to activate a plastic
hinge in the midspan of the beam.

PB2W9: tests on beam with end connections composed of web angles. The
specimen was loaded with three cycles performed at the load levels of
50, 100 and 190 kN. This last cycle is very close to the collapse of the
specimen (figure 6.9).
PB2W10: tests on beam with end connections composed of web angles. The
specimen was loaded with three cycles performed at th load levels of
50, 100, 150 and 190 kN. This last cycle is very close to the collapse
of the specimen (figure 6. 10).
PB2TS16: tests on beam with end connections composed of top and seat angles.
The specimen was loaded with three cycles performed at the load levels
of 50, 100, 150, 200 and 240 kN (figure 6.11).
PB2TS18: tests on beam with end connections composed of top and seat angles.
The specimen was loaded with three cycles performed at the load levels
of 50, 100, 150, 200 and 240 kN (figure 6.12).
PB2TSW9: with end connections composed of web and top and seat
tests on beam
angles. The specimen was loaded with three cycles performed at the
load levels of 50, 100, 150, 200 and 250 kN (figure 6. 13).
PB2TSW10: tests on beam with end connections composed of web and top and seat
angles. The specimen was loaded with three cycles performed at the
load levels of 50, 100, 150, 200 and 250 kN (figure 6.14).

The second series of tests was characterized by a distance between the load points
equal to five metres (i.e. the load points close to the connections). In these tests the
collapse of the specimens was due a instability of the web panel of the beam in the
zone between the connection and the load point

PB5W9: tests on beam with end connections composed of web angles. The
specimen was loaded with three cycles performed at the load levels of
50, 100, 150, 200, 250, 300, 350, 400 and 450 kN (figure 6.15).
PB5W10: tests on beam with end connections composed of web angles. The
specimen was loaded with three cycles performed at the load levels of
50, 100, 150, 200, 250, 300, 350, 400, 450 and 500 kN. This last
cycle is very close to the collapse of the specimen (figure 6. 16).

76
 fixed

experimental

simpty support

K ss _ 4.6 kN/mm

K f - 21.1 kN/mm

K exp. - 6.5 kN/mm

30 40
MRECnON (mm)

6.9. Load/deflection cycle ci test PB2W9.

/fixed 'experimental

simply supported

K ss - 4.6 kN/mm

Kf = 21.1 kN/mm
exp. -7.3 kN/mm

40 50
-SUCTION (mm)

6.10. Load/deflection cycle oftest PB2W10.

experimental

250

_ 200
CM
O simply suppo
<
O
", 'so

ss = 4.6 kN/mm
Kf = 21.1 kN/mm
exp. = 12.6 kN/mm

0 .

o
DRECnON (inri)

6.1 1. Load/deflectiei cycle of test PB2TS16.

77
 fixed experimental

simply supported

K ss _ 4.6 kN/mm

K f -.21.1 kN/mm
Kexp. = 12.4kN/mm

40 SO
DERICTION (mm)

6.12. Load/deflection cycle ci test PB2TS18.

300 + fixed /
/ experimental

simply supported

K ss - 4.6 -N/mm

K f -21.1 kN/mm
K exp. - 15.4 kN/mm

40
DEFLECTION (mm)

6.13. Load/deflectic_ cycle of test PB2TSW9.

fixed experimental

simply supported

K ss ' - 4.6 kN/mm

K f = 21.1 kN/mm
K exp. - 16.3 kN/mm

30
DEFLECTION (mm)

6. 14. Load/deflection cycle of test PB2TS W10.

78
PB5TS 16: tests on beam with end connections composed of top and seat angles.
The specimen was loaded with three cycles performed at the load levels
of 50, 100, 150, 200, 250, 300, 350, 400, 500 and 600 kN (figure
6.17).
PB5TS18: tests on beam with end connections composed of top and seat angles.
The specimen was loaded with three cycles performed at the load levels
of 50, 100, 150, 200, 250, 300, 350, 400, 500 and 600 kN (figure
6.18).
PB5TSW9: tests on beam with end connections composed of web and top and seat
angles. The specimen was loaded with three cycles performed at the
load levels of 50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550
and 650 kN (figure 6. 19).
PB5TSW10: tests on beam with end connections composed of web and top and seat
angles. The specimen was loaded with three cycles performed at the
load levels of 50, 100, 150, 200, 250, 300, 350, 400, 450, 500, 550
and 650 kN (figure 6.20).

For each test the stiffness values related to the relationship load- mid span deflection
in elastic and plastic range. A comparison between these values and the values related
to the ideal cases of simple and fixed beams shows that the experimental elastic stiff
ness is very close to the ideal case of a fixed beam for all the specimens except the case
of specimens with web connections.
On the basis of these values, the stiffness of the connections has been evaluated and
the values are reported in table 6.2.

Table 62. Stiffness of the tested connections.

Stiffness
Kf Kss Kexp.
Test Load distance i (m) (kN/mm) (kN/mm) (kN/mm) (kNmm/rad)
PB2W9 2 21.1 4.6 6.50 3.55-106
PB2W10 2 21.1 4.6 7.30 5.32-106
PB2TS16 2 21.1 4.6 12.60 25.50-106
PB2TS18 2 21.1 4.6 12.40 24.30-106
PB2TSW9 2 21.1 4.6 15.40 51.50-106
PB2TSW10 2 21.1 4.6 1630 66.30-106
PB5W9 5 210.6 15.7 28.98 5.70-106
PB5W10 5 210.6 15.7 30.13 6.24106
PB5TS16 5 210.6 15.7 62.40 24.60-106
PB5TS18 5 210.6 15.7 63.60 25,50106
PB5TSW9 5 210.6 15.7 64.51 26.10106
PB5TSW10 5 210.6 15.7 74.9 34.10106
Kf = stiffness of the fixed beam
Kss = stiffness of the simply supported beam
I-exp. = experimental stiffness of the beam
K, = stiffness of the joint

79
 fixed experimental

700 -

600

supported
500

40p -f

300

K ss = 15.7 kN/mm

K f = 210.6 kN/mm

K exp. - 28.98 kN/mm

20 25
OBJECTION ImnO

6.15. Load/deflecti(Hi cycle of test PB5W9.

fixed experimental

simply supported

Kss = 15.7 kN/mm

K f = 210.6 kN/mm

K exp. = 30.13 kN/mm

15 20 25
DEFLECTION (mm)

6.16. Load/deflection cycle of test PB5W10.

fixed experimental

simply supported

Kss = 15.7 kN/mm

K f - 210.6 kN/mm

K exp: - 62.4 kN/mm

20 25
DEFLECTION (mm)

6.17. Load/deflection cycle of test PB5TS16.

80
 experimenta)

simply supported
500
u.

O
O 400
Kss - 15.7 kN/mm

g 300 Kf = 210.6 kN/mm

K exp. - 63.6 kN/mm
200

20 25
DEFLECTION (mm)

6.18. Load/deflection cycle oftest PB5TS18.

800

700

300 -
K wc. - 15.7 kN/mm
200 -
K inc. - 210.6 kN/mm

100 ICjptrim.- 64.51 kN/mra

20 25
raCClAtn-d

6.19. Load/deflection cycle of test PB5TSW9.

800 --

/fixed experimental
700

600
simply supported
_ 500
CM

O
<
2 *oo
<
Kss = 15.7 kN/mm
O 300
K1 = 210.6 kN/mm
200 K exp. = 74.9 kN/mm :

100

15 20 25
DEFLECTION (mm)

6.20. Load/deflection cycle of test PB5TS W10.

81
The values shown in table 6.2 have been put in non dimensional form according to the
EC3 criteria for the classification of the joints. Figures 6.21 and 6.23 are related to the
domain for a no-sway frame; figures 6.22 and 6.24 are related to the domain for a
sway frame. It is possible to observe that the joint are classified as semi-rigid or rigid.
These first data show that the influence of the joint characteristics is non neglegible
also in cases of joints traditionally considered as hinge (web angles joints). The
numerical analyses performed on the frame 1 (see 6.3) on the other hand confirm this
influence on the serviceability level of the frame. The influence of the shear on the
rotational characteristics of the joints in elastic phase is hon neglegible, as it appears
from table 6.2.

m
1.2 -
P82TSWlo/ /P82TSW9 PB2TS16 PB2TS18

RIGID

SEMI - RIGID

o
0 0.05 0.1 O.JS

6.21. M-F curves for specimens PB2... in a no-sway hame.

PB2TSW10/ / PB2TSW9 P62TS16 P82TS18

RIGID

SEMI - RIGID

o :

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.18 0.18 0.2

6.22. M-F curves for specimens PB2... in a sway frame.

82
 m

PB5TSW10/ P85TSWS / PB5TS16

RIGID P8SWI

BS WS

SEMI - RIGID

0.15

6.23. M-F curves for specimens PB5... in a no-sway frame.

m
1.2 PB5TS18

P8STSW10 PBSTSWS P8STS16

RIGID

reswto
SEMI - RIGID
PBSW9

0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18

6.24. M-F curves for specimens PB5... in a sway frame.

Joints considered in the numerical study

The joints considered in this numerical research were chosen in order to cover a wide
range of stiffness and strength values. The - curves, deduced from experimental
studies, are presented in figure 6.25 in non-dimensional form, according to the EC3
criteria for joint classification. It is possible to observe that:
joint S 1 : is classified as a hinge for both rotational stiffness and bending capacity
[Davison, et al., 1990];
joint JT12: is characterized by an initial stiffness in the semi-rigid range and a
bending capacity in the hinge domain [Davison, et al., 1987];
joint JT13: is defined as semi-rigid joint for both stiffness and strength [Davison,
et al., 1987];
joint EPB1-1: the initial stiffness is in the rigid domain but the inelastic response is in
the domain of the semi-rigid joints [Bernuzzi, et al., 1991].

83
 m=M/MP)b Lb=5m

LOWER BOUND

'SI
0.4 0.8 1.2 1.6 2.0 0 = 0EIb/(LbMp>b)
625. - curves deduced from experimental studies.

Joints characterized by the - law boundary between rigid and semi-rigid domain
and semi-rigid and hinge domain, i.e, upper bound (EC3 U.B.) and lower bound of
the EC3 (EC3 L.B.) domain for semi-rigid joints were also considered in the numeri
cal study.
The cases of rigid joints and hinges are respectively represented by the vertical and
horizontal axis in this non-dimensional domain.
The results of the analysyse performed have been organized in order to appraise and
evaluate:
- The influence of the order analysis (Le., 1st and 2nd order analysis) on the consid
ered frames in the case of simple and rigid frames.
- The joint action on the semi-continuous frame.
- The influence of the model adopted in the design of the stractural response consid
ering a cladding system composed of metal sheeting.
- The minimum values of the shear stiffness required to the cladding panels in order
to satisfy the serviceability limits of Eurocode 3.

The service load multiplier otg was considered by dividing th ultimate load multiplier
Ou by the factor 1.43 according to plastic design criteria.

6.3 Joint action and frame performance in service

Prior to the numerical research into the response of the serviceability loads of sway
frames, some analyses were performed on no-sway frame 1. The aim of this study
was to evaluate the joint action in the case of braced frames. Elasto-plastic analyses
were performed considering the simple frame and semicontinuous frames with semi
rigid joints: SI, EC3 L.B., JT13 and EC3 U.B.
The serviceability load multiplier, defined by considering the case of the simple frame:
otsji = ctu/1.43, shows that all the correspondenting vertical displacements are lower
than Lb/250, the EC3 limit. Considering the joint action (ccsj = ocuj/1.43), it is possi
ble to observe a substantial increase in the serviceability load multiplier (figure 6.26):

84
 Frames with joint EC3 L.B. have a serviceability load 25% greater than the value
related to the simple frame.
Frames with joint EC3 U.B. have a serviceability load that is 78% greater than the
value related to me simple frame.

EC3 .B.

10 20

626. Relation between the service load multiplier and the deflection in ame 1.

The numerical study into the response of sway frames was at first devoted to apprais
ing the influence of the order of the elasto-plastic analysis on the response of the con
sidered frames.
First and second order analyses were performed considering only the ideal cases of
simple and rigid frames, in order to evaluate the bounds of variability for the semi
rigid joints. The incremental analyses were performed till the collapse of the frame was
achieved and the results have been evaluated considering the load multiplier versus
the maximum horizontal displacement V.
Simple frame, in figure 6.27 the -V response of frames is presented for the two
different types of analysis and for equal to 0.0125. It is possible to observe a tangi
ble difference, decreasing with the increase of , between the 1st and the 2nd order re
sponse. In fact, with the smaller value of the 1st order curve shows a strictly linear
trend, different from the 2nd order one. This underlies the important rule played by the
geometric noh-linearities; in the first case the collapse is achieved wim the formation of
the plastic hinge in the midspan of the beams (beam mechanism) while in the other
case it is due to plastic hinges on the column ends due to geometrical effects (panel
mechanism). For higher values of the collapse mechanism is due in all the consid
ered cases to the panel mechanism and the differences in the -V curves are due only
to the second order effects. It is possible to observe the strong influence of 2nd order
effects; the differences, in fact, are more than 50% for = 0.0125 progressively de
creasing with the increase of me parameter.

85
 1. -
a
a 1 ORDER

ORDER

1.0 -

L_______ V

0.5 -
Or
HINGE /S-0.0125

V [mm]
100 200 30

6.27. -V response of frames B for two different types of analysis ( = 0.0125).

Rigid frame. In figures 6.28 and 6.29 the -V curves of frames A and are pre
sented for the two different types of analysis and for equal to 0.10. Here the trend of
the curves is non dependent by the order of the analysis and the rigid frame seems less
sensitive to the type of analysis. The analysis of the collapse mechanisms shows that
in the case of first order analysis, collapse is due to the beam mechanism for equal to
0.0125 and to the mixed mechanism for higher values of . In the second order anal
ysis the collapse is due to the interaction between instability and plasticity. The differ
ences between the displacement at the same load levels of the two types of analysis are
slighter with reference to the simple frame analysis with a range of 10% to 14% for
frame A and of 6% to 10% for frame B.

1.5:
a I 0EDER

ORDER

1.0 FRAME Al
V

RIGID JOINTS
0.5

j?=0.10

V[mm]
0
0 100 200 300

6.28. -V response of frames A for two different types of analysis ( = 0.10).

86
 I ORDER
1.5
a

ORDER

IfRAM Bl
1.0

"fftR vSTC *5

0.5
[3=0.10

RIGID JOINTS

0.0
V [mm]
0 100 200 300

6.29. -V response of frames B for two different types of analysis ( = 0.10).

These preliminary analyses have shown that the considered frames are very sensitive
to the second order effects. Due to the slenderness, it seems necessary also to evaluate
the geometric non-linearity and the results here presented are mainly related to second
order elasto-plastic analyses.
The joints considered in the part of the numerical study aimed at appraising the joint
action are usually classified as hinge (joints SI, JT12, JT13, EC3 L.B.) or rigid joints
(EPB1-1, EC3 .B.) on the basis of the type of the joint. The - non-dimensional
curves in figure 6.25 shown, on the other hand, that in all these cases joints are
characterized by non neglegible stiffness and strength and, at least for joint JT13 and
EPB1-1, the semi-rigid model could be adopted.
As was previously mentioned, the behaviour of the joint was taken into account by a
rotational spring characterized by a multilinear - law. This permits to fit the exper
imental curve with a satisfactory degree of accuracy.
The analyses were performed untili collapse was achieved and the results are here or-
ganized on the basis of the following criteria:
- Frame type (frame A or frame B, in figure 6.3).
- Load condition (=F/qLb).
- Type of ideal joint assumed in the traditional design (i.e., hinge or rigid joint) in
order to evaluate the differences between the traditional design (based on simplified
models) and the design of the semi-continuous frame.

The service load multipliers considered in this part of the study were referred to the
ideal cases of simple frame ^ (= /1.43) and rigid frame (^ (= aUjI/1.43). The
term Osj is related to the semi-continuous frame, i.e. as,j is defined as a j divided by
1.43.

87
In figures 6.30 and 6.31 the responses of the frame are reported for the lower value of
; in figures 6.32 and 6.33 they are related to the load condition with the higher value
of .
The influence of die joint action on the frame response is non-neglegible; in the case of
joints SI and JT12, characterized by the initial stiffness in the semi-rigid zone of the
C3 domain and by a low bending capacity (no greater titan 15% of the plastic mo
ment of the beam) the frame modelled as semi-continuous results more rigid than the
simple frame. It suggests that in these cases, i.e. joints classified, according to EC3
criteria, as hinge for strength or stiffness, the accurancy of the analysis can be im
proved by taking into account the actual behaviour of the joint

a
3.0
-o- HINGE
-f- SI
-0- EC3 LB.
--012
2.5 -ft- JTIO
-*- EC3 U.B.
-*- RIGID
2.0
IFRAME Al
S-r V
1.5
|-g=0.0125
1.0

0.5
SA

0.0
/50p V [mm]
0.0 100 200 300 400

6.30. Response of unsheeted frame A for the lower value of (= 0.0125).

3.0
IFRAM- Bl
EC3 U.B. V
RIGID
2.5

EPB l-l
V. % tl c - * .

2.0 JT13
=0.0125
O-S-r 3~_

1.5
JT12
SI

1.0 HINGE

tt:SA
0.5

h/soo V [mm]
0.0
50 100 150 200 250

6.31. Response of unsheeted frame for the lower value of (= 0.0125).

88
 1.5
a I FRAME Al

EC3 U.B. V
RIGID
EPB 1-1

1.0

* JT13
EC3 LB.
=0M
JT12
0.5
Sl

as HINGE
H/600

o.o
V [mm]
100 200 300 400 500 600

6.32. Response of unsheeted frame A for the higher value of (= 0.10).

1.5 a

IFRAME Bl
RIGID EPB 1-1

EC3 .B.
JT13
1.0
EC3 LB.
Sx
JT12 /g=0.10

si
0.5

HINGE
a

0.0
/500 V[mml
0 100 200 300

6.33. Response f unsheeted frame for the higher value of (= 0.10).

The responses of the rigid frame and of the frame with joints characterized by the EC3
U.B law are practically coincident This confirms the fact that when the - law of
me joint is in the rigid zone of the domain for the classification, the simplified model of
rigid frame can be assumed in the structural analysis, neglecting the actual behaviour
ofthejoint
The service multiplier, previously defined, and the sway indexes H/V (where is the
total height of the frame and V is the maximum horizontal displacement at the servece-
ability level) enable the characterization of the tha frame responses. These parameters
were determined by incorporating the 'actual' joint behaviour, and the sway indexes
were evaluated for the service loads corresponding to the ideal case.

89
The significant influence of joint action is also apparent from table 6.3 which shows,
for frame A and frame B, the sway index for the ideal case, the ratio between the sway
index of ideal frame over the sway index of semi-continuous frames, the increment of
the service load with respect to the ideal case and the sway index of the semi-continu
ous frame.

Table 6.3. Lateral drift without cladding (SX = Sway Index = .

Rame A | Frame
S.I. at S.I.h Ch S.I. at D S.I. at S.I.h Oh SX at
Joint Cts.h S.I.J j as.j | Cts.h S.I, aj cts.j
hinge 60 1.00 1.00 60 1 81
1.00 1.00 81
SI 1
203 0.30 1.43 118 195 0.42 1.35 117
JT12 655 0.09 2.33 163 368 0.22 1.43 195
80
EC3 LB. 583 0.10 2.58 188 511 0.16 1.68 277
JT13 1784 0.03 2.69 265 327 0.25 3.23 328
hinge 39 1.00 1.00 39 51 1.00 1.00 51
SI 1
56 0.70 1.36 41 109 0.47 1.31 69
JT12 347 0.11 2.09 136 234 0.22 1.94 126
20
EC3 LB. 266 0.15 2.72 89 208 0.25 2.23 110
JT13 645 0.06 3.09 141 450 0.11 2.44 161
hinge 34 1.00 1.00 34 48 1.00 1.00 48
SI 1
71 0.48 1.33 44 83 0.58 1.31 56
JT12 387 0.09 1.93 106 328 0.15 1.89 106
10
EC3 L.B. 214 0.16 2.59 77 267 0.20 2.28 91
JT13 599 0.06 2.94 135 549 0.09 2.56 128
rigid 677 1.00 1.00 677 1 566 1.00 1.00 566
1
EC3 U.B. 662 1.03 1.01 656 508 1.11 1.02 483
80
EPBl-1 308 2.19 0.99 295 271 2.09 0.93 323
rigid 259 1.00 1.00 259 257 1.00 1.00 257
1
EC3 U.B. 251 1.03 1.02 217 250 1.03 1.02 243
20
EPBl-1 197 1.32 0.98 204 190 1.35 0.98 196
rigid 1
192 1.00 1.00 192 I 197 1.00 1.00 197
EC3 U.B. 185 1.04 1.00 183 1 192 1.03 1.01 190
10
EPBl-1 154 1.25 0.94 175 1 172 1.15 0.99 175

The degree of continuity provided even by the most flexible connections substantially
increases the stiffness of the frame (up to 30 times for frame A and to 42 times for
frame with equal to 0.0125). The ultimate frame strength also improves remark
ably, hence higher service loads might be admissible (Ogj). Frame stiffness, however,
is not sufficient to allow this potential increase of resistance to be fully exploited (see
die values of the sway index determined at Osj).
Frame response is far less sensitive to a variation in joint flexibility witii respect to the
rigid model. This applies to both ultimate strength and stiffness in service. However
the recognition of the semi-rigid behaviour of joint EPBl-1 implies a remarkable in
crease in the flexibility of the frame model.
An appraisal of frame performance in service based on this model would lead to the
frame being regarded as inadeguate for equal to 0.0125. since extended end plate

90
 joints are traditionally considered 'rigid' and frames accepted in the past on the basis of
rigid frame analysis would be rejected now if joint action is incorporated in die design
analysis.

6.4 Cladding action

General
Cladding panels are considered non-structural elements and, as previously reported,
they are generally neglected in the design of the skeleton frame. On the other hand,
their contributions to the stiffness and strength of steel frames has been recognised and
intensively investigated in the last decades from many authors.
The earlier studies into the behaviour of infilled frames were made in the UK, in the
US, and in the URSS [Thomas, 1953; Benjamin and William, 1957; Polyakov,
I960]. Experimental and numerical research was conducted in the past into different
forms of cladding panels. Tests were performed in order to evaluate the shear stiffness
and the strength [Wood, 1978] of the panels as well as the mechanical characteristics
of the connections between the cladding and the frame skeleton. Numerical studies, on
the other hand, evaluated the influence both of the height/width ratio and also of the
openings (due to windows and doors) on the response and mechanical properties of
the panel.
On the basis of these studies some methods for the prediction of the behaviour and of
the strength of the cladding have been developed. They result very usefull for the de
sign phase; predicting the response of the cladding panels on the basis of the geometri
cal andmechanical characteristics, it is possible to take into account the cladding action
via very simple elements (trusses or systems made up of trusses) which are able to re
produce the response of these 'non-stractural' elements. This improves the degree of
reliability of the analysis and of the design. One of these methods was developed for
panels of steel sheeting and it is briefly presented in the next section of this paragraph
[Bryan, 1972]. An application was performed in order to assess the influence of the
tiiickness of the coils on the response of the panel.
The results of experimental studies can be directly used in design analysis only if the
specimen is coincident to the cladding panels of the frame (same sizes or same
height/width ratio and same type of connections to the beams and/Or to the columns).
At me end of mis chapter some tests related to different types of cladding are presented
organized as a data-bank in data sheets where the main characteristics of the specimens
are illustrated. It appears that no rules related about the sizes of the specimens, about
the loading histories and to the measuring arrangement have been established. This
prevents from a direct comparison between the responses of cladding different only for
the sizes, tested in different researches. A simple procedure to extrapolate these exper
imentais results to systems characterized by different length/width ratio is presented in
the last section of this paragraph.

91
The Bryan method for panels of steel sheeting
The method developed by Bryan to predict the elastic response of panels of metal
sheeting (figure 6.34), i.e., the shear flexibility of the panel, is based on the following
hypotiiesis:
- end gables of the frame cross-braced or sheeted;
- sheeting directly fastened to the purlins;
- sheeting rails and side laps of adjacent sheets firmly stitched;
- presence of shear connectors.
If the end-gables or the stiffened internal partition exceed 4 times the width for rectan
gular roof buildings and 2.5 for pitched ones, this method can not be applied (as the
suggested frames are single storey or no more than than two or three storeys only if do
not arise instability problems). The method makes three contributions to the total
cladding flexibility c: the flexibility due to the sheet deformation (ci), the flexibility due
to the sheet fasteners (C2) and the flexibility (03) due to the purlin/rafter connection.
Each sheeting is characterized by a constant K evaluated on the basis of a computer
procedure developed by Mohsin and Bryan: its value can be found by approximating
the equivalent rectangular corrugation as shown in figure 6.35. The value of K is then
determined by considering the 4q/d and h/d ratios relevant to the fastening frequency
of sheetings.
A correction factor f depends on the number of intermediate purlins or beams and it is
correlated with the strain energy. The slip s (between me panel and the frame skeleton)
can be determined only by considering the experimental curve shear load-displacement
for the fastener.
The equations to calculate the shear flexibility c of cladding are shown in figure 6.36,
while the symbology of these equations is presented in table 6.4. It should be noticed,
about the flexiblity due to sheet fasteners C2, that there are two possible cases to con
sider:
case 1: sheeting fixed to purlins and shear connectors, would normally apply to the
design of new structures;
case 2: sheeting fixed to purlins only, would apply to the analysis of existing struc
tures.

An application of this method has been developed in order to appraise the influence of
the thickness of the steel panel on the value of shear flexibility. Referring to the frames
A and B, it has been considered a cladding panel with a width of 4 m and a lenght of 5
m (figure 6.35), composed of 6 rows (nsn = 6) of three steel sheetings of 0.9 m of
width with six seam fasteners per side laps (ns = 6) which are about 80 mm. The cross
section of the cladding is a syimmetric one whose height is equal to 55 mm and the
connection is composed of 8 self-tapping screws for the panel-to-column joint n^ = 8,
(see figure 6.34). Cases with thickness equal to 0.6 mm and to 1.5 mm were consid
ered applying this method.

92
 Seam or sidelamp fasteners

Sheet-
connector

Sheet purlin fastners

connection
\a Intermediate
purlin

^\^^
connector
\ Rafter
* Edge purlin

6.34. Arrangement of panel sheeting.

ns = 6 a=5Q00 mm

nsc = 8 b =4000 mm
np = 4

_^~L = d-6 = 900=>(pnsh)-a

side lap = 5
= 400

joint for side lap = ^- = 80 mm

6.35. Equivalent rectangular comigaon for metal sheeting, used by Monsin and Bryan.

93
Table 6.4. Symbology of the equations, used in figure 6.36.
a ; length of shear panel ( to direction of sheet) mm
A cross sectional area of edge member ( to direction of span of sheet) mm2
b depth of shear panel (// to direction of span of sheet) mm
c overall shear flexibility of panel mm/kN
d pitch of corrugation mm
E modulus of elasticity (= 207 kN/mm2 for steel) kN/mm2
f correction factor to allow for intermediate purlins
h height of corrugation mm
K sheeting constant
I width of crest of corrugation mm
np number of purlins
ns number of seam fasteners per side lap
nsc number of sheet/connector fastenen; per rafter
nsh number of sheet widths per panel
pitch of sheet/purlin fasteners mm
s slip per sheet/purlin fastener per unit load mm/kN
SS slip per seam fastener per unit load mm/kN
ssc slip per sheet/connector festener per unit load mm/kN
spr top movement of purlin relative to rafter per unit load mm/kN
t sheet thickness mm
V Poisson's ratio (= 025 for steel)

c-ci*C2*C3

Flexibility cf - Sheet deformation j

C1=C1_1+C1_2 + C1.3

FLEX. DUE TO SHEET DISTORSION Cu- (0.144adf,K) / (Et3 b3) [mm/kN]

FLEX. OUE TO SHEAR STRAIN IN THE SHEET ci ? = (2-a-f,) (1+2-h/d) (1+n) / (Etb) [mm/kN]
FLEX. OUE TO AXIAL STRAIN IN PURLINS c-t -, = (23 f,) / O-b^A-E) fmm/kN]

Flexibility C2 - Sheet fasteners

C2 = C2.1 + C2.2 + C2.3

SHEET/PURUN FASTENER Case(l) c,, = (2-a-s-p/^ [mm/kN]

Case (2) c,i = (2-s-pW6np--+a;iMr2/a:i [mm/kN]
SEAM FASTENER Cased) c,,= (ru*-D-s/i-u [mm/kN]
Case (2) c,-iws/n, [mm/kN]
SHEET/CONNECTOR FASTENERS Case (1) c, , = 2& / iw [mm/kN]

Flexibility 03 - Purlin/rafter connection

C3 =2-s/ n [mm/KN]

6.36. Shear flexibility of a panel.

94
The data related to die metal sheeting are reported in die table 6.5. The main steps re
lated to the application of the method are illustrated in the tables 6.6 and 6.7 for the
lower and higher value of the thickness, respectively.

Table 6.5. Data related to the metal sheeting, calculated in table 6.6 and 6.7.
d 150 mm d3 56.7538545 mm
I 61 mm teta 1.32154419 _d
e 30,5 mm d4 7 mm
h 55 mm eq 75 mm
dl 89 mm
V1 0.5 -
_ 14 mm h/d 0.36666667 -

Table 6.6. Main steps related to the application of the Bryan method for metal sheeting with a low
thickness of 0.6 mm.

cladding width a 5000 mm

cross sectional area A 5400 mm2
cladding height b 4000 mm
pitch of corrugation d 150 mm
modulus of elasticity E 207 kN/mm2
correction factor to allow for intermediate purlins f 1 _
height of corrugation h 75 mm
sheeting constant (1/1) K 33 -
equivalent width of crest corrugation feq 75 mm
number of purlins np 4 -
number of seam fasteners per side lap ns 6 -
number of sheet/connector fasteners per rafter nsc 8 -
number of sheet widths per cladding nsh 6 -
pitch of sheet/purlin fstenas 900 mm
slip per sheet/purlin fastener per unit load s 0.35 mm/kN
slip per seam fastener per unit load SS 0.35 mm/kN
slip per sheet/connector fastener per unit load ssc 0.35 mm/kN
sheet thickness t 0.6 mm
Poisson's ratio (= 0.25 for steel) ni 0.25 - j

sheet distortion c.1.1 4.2035 mm/kN

shear strain in the sheet c.1.2 0.0503 mm/kN
axial strain in purlins c.1.3 0.0047 mm/kN
SHEET DEFORMATION el 4.2584 mm/kN
sheet/purlin fastener c.2.1 0.1969 mm/kN
seam fasteners c.2.2 0.2917 mm/kN
sheet/connector fasteners c.2.3 0.0875 mm/kN
SHEET FASTENERS c.2 0.5760 mm/kN

SHEAR ELEXffiDJTY c = cl + c2 4.8345 mm/kN 1

SHEAR STIFFNESS k=l/c 02068 mm/kN |

95
Table 6.7. Main steps related to the application of the Bryan method for metal sheeting with a high
thickness of 1.5 mm.

cladding width a 5000 mm

cross sectional area A 13500 mm2
cladding height b 4000 mm
pitch of corrugation d 150 mm
modulus of elasticity E 207 kN/mm2
correction factor to allow for intermediate purlins f 1 -
height of corrugation h 55 mm
sheeting constant (1/1) K 33 -
equivalent width of crest corrugation 75 mm
number of purlins np 4 -
number of seam fstenas per side lap ns 6 -
number of sheet/connector fasteners per rafter nsc 8 -
number of sheet widths per cladding nsh 6 -
pitch of sheet/purlin fasteners 900 mm
slip per sheet/purlin fastener per unit load s 0.35 mm/kN
slip per seam fastener per unit load SS 0.35 mm/kN
slip per sheet/connector festener per unit load ssc 0.35 mm/kN
sheet thickness t 1.5 mm
Poisson' s ratio (- 0.25 for steel) ni 0.25 -

sheet distention c.1.1 0.2690 mm/kN

shear strain in the sheet c.1.2 0.0174 mm/kN
axial strain in purlins c.1.3 0.0019 mm/kN
SHEET DEFORMATION c.l 0.2883 mm/kN
sheet/purlin fastener c.2.1 0.1969 mm/kN
seam festeners c.2.2 0.2917 mm/kN
sheet/connector festeners c.2.3 0.0875 mm/kN
SHEET FASTENERS C.2 0.5760 mm/kN

SHEAR FI__<ffiI_rTY c = c.l + c2 0.8644 mm/kN

SHEAR STIFFNESS k=l/c 1.1569 mm/kN

First the equivalent width of die crest of corrugation (as it is shown in table 6.5), and
then the constant of steel sheeting K were calculated. On the basis of these values, the
shear flexiblity was evaluated using the equations contained in figure 6.36; me values
of elastic shear stiffness related to the two considerd thickness, kt=o.6 mm = 0.2
kN/mm and k^LSmm = 1.1 kN/mm show the non neglegible influence of me thick
ness on Ae shear response in me elastic field.

The FEM method

As was previously pointed out, die experimental results related to die shear stiffness of
cladding panels can be used in the design phases only if die H/B ratio and the connec
tion system are equal. These hypotieses are rarely satisfied. An approach aimed at al
lowing the use in elastic field of die experimental data in the case of different H/B ratio
has been developed. The procedure can be summarized in the following steps:

96
 1. On the basis of the results of the tests (i.e., F*, load applied to the panel and s*,
displacement in elastic range) evaluation of me shear elastic stiffness (s*/F*).
2. Evaluation of the fictitious elastic (E) and tangent (G) modulus using approximate
formulae.
3. Modelling die specimen using finite bi-dimensional element (plane stress element).
An iterative procedure based on finite element analyses will give me values of E
and G for me equivalence between the numerical and experimental stiffnesses.
4. Finite element analyses of panels characterized by different H/B ratios in order to
determine the elastic shear stiffness Kei versus me geometrical ratio H/B.

Some details related to step (3) are shown: from the experimental study the elastic
shear stiffness is defined; using, for example, the Airy theory with appropriate bound
ary conditions it is possible to estimate the elastic modulus E*:

*_________ F*B2
S ~3E*J + 8GJ
where:
F* = applied force
H =panelhigh
=panel length
J = inertia modulus

With me hypothes that G =|-E*, it follows:

8F*H3+9F*B2
E*=-
24s*J

As it appears in figure 6.37, this value is then used in a first finite element analysis on
the panels in elastic range. In the case of lack of agreement between this numerical
stiffness and the experimental ones, a change in the values of E and G and a hew anal
ysis are required until die two values are practically coincident

starting value of E and G

numerical analysis

Check of the shear stiffness

in elastic range

Satisfied

Stop

6.37. Iteration to find the appropriate value of E*.

97
 Some iterations (tiiree or four) are generally required in order to evaluate the fictitious
values of E* and G* that make me panel equivalent (for elastic shear stiffness) to the
specimen tested. These values are men used in finite element analyses on the same type
of panel characterized by different H/B ratio (step 4% in order to produce the curve
shear stiffness- sizes ratio, which is very useful for me design of the frame taking into
account the cladding action.

6.5 Modelling of cladding action and cladding action on the considered

frames
Modelling of the cladding panels
The interaction between the frame skeleton and me cladding elements is fairly com
plex, as it depends on me responses of these two systems as well as on the type of
cladding-frame connection. The degree of this interaction has been recognized as a
primary factor affecting the frame response. Three different types of design ap
proaches are possible [Mazzolarli and Piluso, 1990].
a. The main structure is designed to resist vertical and horizontal loads, while the
panel contribution is taken into account in the serviceability limit state only, when
checking maximum sway and storey drifts.
b. The whole structure is designed so diat panels and frames together have to resist
everykind of load.
c. The frame resists only to vertical loads while horizontal forces caused by earth
quakes or wind are supported by cladding panels.

The choice Of beam-to-column connections should be made according to me above

design criteria: case (a) requires rigid connection and case (b) can accept semi-rigid
joints connections. As die bracing effect can be guaranteed by claddings, pin-ended
connections also can be used in case (c), which provides maximum economy in reduc
ing botii structural weight and manufacturing and erection cost. In spite of these
advantages, examples of structures like type (c) are not used in seismic zones,
probably due to a lack of specific knowledge and experience.
Different models can be used in order to evaluate die cladding action on die frame de
sign. Using the refined analysis programs, it would be possible to use bi-dimensional
elements in non-linear ranges and to model the connection with non-linear springs
which can take into account anelastic slips. More simple approaches can be used in or
der to evaluate die cladding action in die elastic range [Anderson, 1985], Here the
cladding action is simulated with a system of elastic boundaries (axial springs in each
storey). Anomer simplified approach, followed in mis research, consists of modelling
cladding in the elastic field by an 'equivalent truss' [Mazzolani and Sylos Labini,
1984]. The cross sectional area of each truss is evaluated as appears from Figure 6.38.
The truss action must be equivalent to die cladding action and to the relationship be
tween die area of die cross section of the truss and me shear flexibility of die cladding.

98
 Cladding - Truss equivalence

L=-
cos S
F
N=
cos S
NL
l

ul--: - -
cosS cosa
Ul = s-cosS

. F 1
A=-
s E cosJS

Flexibility. c = Stiffness: .=---

el c s

. 1 1
A=
c E cos3

6.38. Truss equivalent of cladding action.

On die basis of E and L it is possible to evaluate the cross sectional area A of die
equivalent truss. The truss simulating cladding action is in tension and frame A has di
agonal elements in every span while frame take mem only in die central one. As a
joint between the truss and die frame it a perfect hinge is assumed.

Cladding action on the considered frames

The cladding action on me frame response was evaluated using the model of the
'equivalent truss'. The axial stiffness of this element (or its horizontal stiffness equiva
lent to the shear stiffness of the panel) was evaluated on die basis of an experimental
study performed on a simple steel sheeting (see figure 6.39). The main results of diese
tests are reported in the sheets contained at die end of titis chapter; it was refered to die
cladding type 1A. The dimension of the tested panel were equal to 3200x2640 mm,
which brought out a dimension ratio H/B = 0.73.

99
 SCREWS

o
3
OJ

800 i B00 I 800 I 800

3200 mm

50|~ . . . . A . . . . /\-4
shear force [kN] J

639. Experimental test on simple steel sheeting.

The load-displacement diagram underlines a non-linear behaviour un to a point of dis

continuity. During dus first range of load, die panel don't change its geometrical char
acteristics, but it is only possible to appreciate a local bearing in relation to die number
screws. At the maximum load of 88 kN, when die displacement is about 25 mm, the
contemporaneous collapse of me rivets in the central row which cause a sensible re
duction of me load capacity of the panel. Considering die secant stiffness at this mo
ment, an experimental value of kei = 5.4 kN/mm was assumed. Masonry and concrete
infills present significany higher stiffness, even in die presence of openings.
The FEM approach was used in order to evaluate the shear stiffness of the cladding
panel for frame A or which has H/B = 0.8. The mesh is reported in the figure 6.40
and die curve stiffness H/B ratio is in die figure 6.41. The stiffness of die equivalent
diagonal truss has been evaluated according with the procedure presented in a previous
paragraph, and it's value is k = 5.57 kN/mm. Using for the equivalent truss a value of
Young Modulus equal to 210 kN/mm2, the cross sectional area of die truss outcomes
as A = 169.9 mm2 which corresponds approximately to a steel plate of 10x17 mm.
Structural analyses were performed with the aim of evaluate the influence of cladding
action in two cases: full cladding (i.e. cladding panels on each storey) or partial
cladding (i.e., cladding on some storey of the frame).

100
 -I -
U

400

500

6.40. Rearrangement of the mesh.

0.400 0.600 0.S00 1.000 1.200 1.400 1.00 l.*00 zooo

(H/B)

6.41. Curve stiffness H/B ratio.

Full Cladding
First me influence of die geometrical non-linearity was evaluated on the cladding frame
performing 1st and 2nd order analysis in the cases of simple and rigid frames. These
interactions between plasticity and geometrical non linearities are less evident than in
the case of frames without claddings; percentage differences between lateral displace
ments at the service load level (related to die 2nd order analysis) are less tiian 10%.
The simple frame case suggests mat the interaction between plasticity and instability is
not fundamental for die collapse mechanism which is here due to a beam mechanism
for each value of . On the other hand, rigid frames are characterized first by a linear
(elastic) phase and then by a curvilinear (elastoplastic) one. The collapse mechanism is
a mixed one widi plastic hinges bom of beams and columns.

101
The first series of analyses considered each storey as 'stiffened' by die metal sheeting
panel. As it results from the table 6.8, drifts at die service load level ocsjj and ce^, de
termined for die hinged and for die rigid frame neglecting cladding action, vary only
moderately with the type of joint (less man 30% difference between die frames with
hinge and semi-rigid JT13 joints). Moreover, the simple panel considered provides in
many cases sufficient stiffness to make me ultimate limit state govern the design.

Table 6.8. Lateral drift with cladding (SJ. = Sway Index = .

Frame A Frame
SJ. at S.I.h 2k SJ. at SJ. at S.I.h a_ S J. at
Joint as.h Si, j as.j as.h SJ, j s.j
hinge 3659 1.00 2.05 2483 1176 1.00 1.23 943
SI 1
3703 0.99 2.25 1617 1240 0.95 1.46 826
JT12 4013 0.91 2.56 1440 1314 0.90 1.58 820
80
EC3 L.B. 4013 0.91 2.83 1378 1476 0.80 1.75 816
JT13 4761 0.77 2.95 1333 1694 0.69 1.83 771
hinge 1980 1.00 3.11 610 665 1.00 2.09 308
SI 1
2000 0.99 3.53 558 733 0.91 2.36 281
JT12 2091 0.95 3.87 494 885 0.75 2.60 258
20
EC3 L.B. 2091 0.95 4.28 471 826 0.80 2.86 277
JT13 2490 0.79 4.47 461 1081 0.61 2.99 266
hinge 1509 1.00 4.52 326 541 1.00 3.12 166
SI 1
1548 0.95 5.12 291 593 0.91 3.53 153
JT12 1633 0.92 5.60 265 681 0.79 3.84 152
10
EC3 LJ1. 1617 0.93 6.19 256 665 0.81 4.23 150
JT13 1881 0.80 6.45 253 1 1 874 0.63 4.43 149
rigid 1536 1.00 1.00 1528 878 1.00 1.06 878
1
EC3 U.B. 1490 1.03 1.01 1477 834 1.05 1.02 834
80
EPBl-1 1488 1.03 0.98 1460 657 1.34 0.97 657
rigid 674 1.00 1.26 533 296 1.00 1.41 296
1
EC3 U.B. 630 1.07 1.26 508 280 1.06 1.44 280
20
EPBl-1 561 1.20 1.26 501 221 1.25 1.36 221
rigid 506 1.00 1.74 284 334 1.00 1.98 334
1
EC3 U.B. 490 1.03 1.74 271 326 1.03 2.02 326
10
EPBl-1 480 1.05 1.74 259 320 1.04 1.90 320

A great difference characterized the structural response of the sheeted frame with re
spect to die bare one. The linear behaviour of the '-v' curves (see figures 6.42, 6.43,
6.44 and 6.45) underlines tiiat frame response is mainly influenced by the initial rota
tional stiffness of me joint (both for = 0.0125 and = 0.10). There are particularly
differences in the collapse event between flexible and more rigid joints: when die rota
tional stiffness ios increased* the beam collapse mechanism changes to a mixed one
(beam and column) due to die fact diat me load multiplier increases its value by consid
ering the frame as semi-continuous. Thus claddings grant a great stiffness to frames,
preventing instability effects in columns and so reducing tiieir sensitivity to 2nd order
effects.

102
 '3.0

2.5 I FRAME A
V r?n RIGID
EC3 U.B. EpB t_j
2.0 y JT13
EC3 LB.
JT12
1.5 Si:
HINGE

1.0

=0.0125
0.5
Osa

0.0
V [mm] /soo
10 15 20 25

6.42. Response of sheeted frame A for the lower value of (= 0.0125).

3.0
a
EC3 U.B RIGID
2.5 EPB 1-1

2.0 JT13
EC3 LB.
-
JT12
1.5
SI
1 FRAME Bl
1.0
HINGE
r
sa -- V. V' fe -a v. c

0.5 /S=0.0125

500
V [mn]
0.0
0 10 23 30

6.43. Response of sheeted frame for the lower value of (= 0.0125).

Considering the mor flexible joints, both frames A and comply with the limit ser
vice drift of H/500 (Code's values); for the rigid class of joints, for each type of 'rigid'
joints, die maximum horizontal displacemen exceeds me limit Code (see table 6.8).
In otiier terms, the potential strength of the framework can be fully utilized, at least in
die presence of low to moderate horizontal forces. Fairly high horizontal forces ( =
0.10) still make serviceability limits critical. Shear forces in the panel 'in service' are
well widiin die elastic range of its response. It should be also considered diat die stiff
ening action of cladding substantially reduces die geometrical (-) effects, enabling
the designer to use a 1st Order analysis, at least for serviceability checks: lateral drifts
determined by a 1st and 2nd Order analysis in fact differ less tiian 8%. It is interesting
to note that for joint EPBl-1, 2nd order analysis would be required if the analysis
were conducted on die frame skeleton incoporating die joint response.

103
 3.0
a [FRAME Al

2.5

2.0 *
=0.10
1.5

HINGE
SI
EC3 LB.
1.0 JTIS
JT13
EPB 1-1
EC3 OB.
RIGID
0.5

.s* 500
V[n_n]
0.0
20 40 60 80 100

6.44. Response of sheeted frame for the higher value of (= 0.10).

3.0

EC3 .B.
2.5 RIGID

EPB 1-1

2.0 JT13

EC3 LB.
JT12 IFRAME Bl
1.5 SI V
HINGE
/
1.0
"Sj
r \A R , r

0.5 /S=o.io

0.0
/500 V mm ]
0 50 100 151) 200

6.45. Response of sheeted frame for the higher value of (= 0.10).

Finally the minimum shear stiffness Q_ci) required to die cladding to make me frame
model meet serviceability limits was determined by adopting an iterative procedure.
The values are reported in the table 6.9 in the case of frames witi flexible joints. The
values range form 0.04 and 11.8 kN/mm and it dtis range can be easily provided by
metal sheeting panels. A check of die axial loads on me truss and of die values of shear
force trasmitted form the cladding to die frame shows that these values can be provided
by connection vidi a limited number of bolts. Considering die actual bending continu
ity of the more flexible joints (SI, JT12, EC3.LB, JT13) the service load multiplier re
sults higher than die ideal hinged scheme; consequently higher values of the minimum
stiffness are required in order to meet the strain limit but the increase of the axial load
in me truss and of die consequent shear load in me connections is limited.

104
Table 6.9. Minimum values of the required shear stiffnes for cladding; Kci in kN/mm (t the frame
meets serviceability limit without cladding).

Frame A Frame
Joint Kci at ash Kci at as.j Kci at as.h Kci at as.j
hinge 0.38 0.38 1.28 1.28
SI 0.04 0.36 0.77 1.54
1
JT12 0.46 0,51 1.64
80
EC3 L.B. 0.64 . 1.79
JT13 . 0.56 ' 1.68
hinge 1.03 1.03 3.33 3.33
SI 0.36 0.89 2.05 3.07
1
JT12 0.10 1.38 1.28 5.12
20
EC3 L.B. 0.34 2.05 1.79 7.68
JT13 . 2.06 6.66
hinge 1.33 1.33 4.61 4.61
SI 1
0.87 1.23 3.69 4.35
JT12 0.20 1.54 1.83 7.17
10
EC3 L.B. 0.82 2.66 2.46 11.78
JT13 2.46 ' ' 9.73

Partial cladding
A further series of analyses considered me different conditions rising when panels are
not present on all storeys. Taking into account me action of cladding, a notable in
crease in the load capacity and a skilful limitation of the lateral drift, expecially for
hinged frames was observed. Attention was dien focussed on frame A characterized by
flexible joints (hinge, SI, JT12, JT13, EC3.LB), considering partial cladding, i.e.
the cladding only in one or two storeys of die frame. This condition would simulate
me cases in which the panels me presence of openings (doors and windows) exerts a
very weak stiffnening action on the frame and so ds one can be neglected. It has been
die following cases:
- cladding on die first storey;
- cladding on the third storey only;
- cladding on me first and second storey;
- cladding on die second and tiiird storey;
- cladding at all storey;

For each case, using me previous hypotesis, a 2nd Order elastoplastic analysis has
been performed; Figures 6.46 and 6.47 show die response diagram '-V' on me last
storey for an hinged and for a SI jointed frame. The load multiplier has been reported
related to die case of simple frame without cladding action. It is possible to observe
tiiat die number of storeys with cladding panel influence of course the drift of me
frame but also die position of me equivalent truss. In me cases of truss in die last or in
the second and tiiird storey it has been a more stiffness response mat in the cases of
truss on me first and on die first and second storey respectivelly. in the table 6.10 than
value of me relative storey displacemnts are reported for the considered cases.

105
 0.0 100 200

6:46. Response diagram '-V' on the last storey for an hinged frame (type A).

a
'for Jt

0.0 100 200 300

6.47. Response diagram '-V' on the last storey for an S 1 jointed frame (type A).

6.6 Conclusions
In this numerical study, me attention has been focussed on die influence of joint action
and cladding action on the response of frame at the service load level. Neglecting
cladding action, me comparison witii the ideal cases of simple and rigid frames have
shown that, in some cases, me model of semi-cntinous permit to satisfy die deflec
tion limit specified in die Codes. However die service load defined by considering die
ultimate load of die frame is increased also for joints nominally classified as hinge.
Taking into account the cladding action via non sophisticated models (it has been
adopted die model of equivalent truss) die lateral stiffness of die frame increase sub
stantially. The type of cladding used for this study seems sufficient to wash out die in
crease of flexibility associated to die use of semi-rigid joints (EPB 1-1) in lieu of rigid
joints.

106
Table. 6.10. Storey displacement of frame A with partial cladding ($ = evaluated at asj neglecting
cladding action).

Cladding Joint Storey dis Cladding Joint Storey dis

on storey type Storey placement* (mm) on storey type Storey placement* (mm)
I 3.81 I 2,15
I hinge 22.11 I SI 11,78
01 32.52 m
13,86
I 7,36 I 6,46
m hinge 9,72 m si 8,28
HI HI
1,55 1,40
I 1,24 I 1,26
& hinge 1,81 & SI I 1,57
HI DI
5,82 3,87
I 3,76 I 3,10
n&m hinge 2,57 n&m SI 1,50
HI
3,10 m 0,82
I 1,46 I 1,34
.& hinge 1,36 i, n&m si 1,23
HI
3,07 0,67

References
American Institute of Steel Constuction, Load & Resistance Factor Design. Manual of Steel
Construction, First Edition, USA, 986.

D. Anderson, F.S.K. Bijlaard, D.A. Nethercot and R. Zandonini, Analysis and Design of Steel Frames
with Semi-Rigid Connections, International Association for Bridge and Stractural Engineering IABSE
Survey S-39/87, April, 1987.

D. Anderson, Elastic Analysis of Semi-Rigid Steel Frames, Research Report CE/17, Department of
Engineering, University of Warwick, 1985.

A. Astaneh and M.N. Nader, 'Design of Tee Framing Shear Connections', Engineering Journal,
American Institute of Steel Construction, First Quarter, 1989.

J.R. Benjamin and H.A. Wiliam, 'The behaviour of one storey reinforced concrete shear walls',
Proceedings of the American Society of Civil Engineers, 1957, 83 (ST5) Paper 1254.

C. Bemuzzi, R. Zandonini and P. Zanon, 'Rotational behaviour of End Plate Connections',

Costruzioni Metalliche, Vol. 2, 1991, p. 74-103.

E.R Bryan, 27k? stressed skin design of steel buildings,<^sby-Lockwoo-Stapkss, 1972.

Commision of the European Communities, Eurocode 3. Design of Steel Structures. Part 1. General
Rules and Rules for Buildings, 1992.

Commision of the European Communities, Eurocode 4. Design of Steel Concrete Composite

Structures. Parti. General Rules and Rules for Buildings, 1992.

E. Cosenza, A. De Luca and C. Paella, 'Elastic Buckling of Semi-Rigid Sway Frames', Stability and
Strength, Vol. 8, Structural Connections, London 1988, p. 253-296.

107
-, 'Inelastic Buckling of Semi-Rigid Sway Frames', Stability and Strength,
Vol. 8, Structural Connections, London, 1988, p. 297-335.

J.B. Davison, P.A. Kirby and D.A. Nethercot, Rotational Stiffness Characteristics of Steel Beam-to-
Column Connections Joint Flexibility in Steel Frames, London, 1987, p. 17-54.

J.B. Davison, D. Lam and D.A. Nethercot, 'Semi-rigid action of composite joints', The Structural
Engineer,'Vol. 68, No. 24, December 1990, p. 489-499.

N. Kishi and W.F. Chen, 'Moment-Rotation Relations of Semi-Rigid Connections with Angles',
Journal of Structural Engineering, Vol. 116, No. 7, July 1990.

T.C. Liauw, 'Steel Frames with Concrete Infills', Stability and Strength, Vol.: Steel-Concrete
Composite Structures, London, 1988, p. 115-162.

F.M. Mazzolani and V. Piluso, 'Skin effect in pin-jointed steel structures' (in English), Ingegneria
sismica, Anno VH, No. 3, 1990, p. 30-47.

F.M. Mazzolani and F. Sylos Labini, L'effetto pelle nelle strutture sismo-resistenti di acciaio (in
Italian), Proceedings of the IX National Conference on Steel Construction (C.T.A.), Perugia, 1983.

D.A. Nethercot and R. Zandonini, 'Methods of Prediction of Joint Behaviour: Beam-to-Column

Connections', Stability and Strength, Vol. 8, Structural Connections, London, 1988, p. 23-62.

C. Poggi, 'A Finite Element Model for the Analysis of Flexibly Connected Steel Frames',
International Journalfor Numerical methods in Engineering, Vol. 26, 1988.

S.V. Polyakov, On the interaction between masonry filler walls and enclosing frame when loaded in
the plain of the walls, Earthquake Engineering Institute, San Francisco, 1960.

F.G. Thomas, 'The strenght of brickwork', Structural Engineer, 1953, 31 (Feb.), p. 35-46.

R. Zandonini and P. Zanon, Beam design in PR Breced Frames, American Institute of Steel
Construction, Third Quarter, 1991, p. 26-1/26-26.

R.H. Wood, 'Plasticity, composite action and collapse design of unreinforced shear wall panels in
frames', Proceedings of the Institution of Civil Engineers, Part 2, 1978, VoL 65, June 1978, p. 381-
411.

108
 Testing and analysis of a full-scale steel
framed building

This chapter is concerned with the testing and modelling of a full-scale steel-framed
building under service load levels, conducted at TNO-Bouw, Delft. The objectives of
diese tests and models are to observe and predict real values of static lateral stiffness.

7.1 Modelling of a two-storey steel-framed office building

The building investigated is located at TNO, Delft, die Netherlands. It is referred to as
building 1 1, and may be classified as an office building. This structure was chosen as
it is representative of many low-rise steel-framed office buildings in the Neerlands.
General views of the exterior of this structure are shown in figure 7.1 and 7.2. Total
building length in the minor axis is 20.7 m, and in the major axis 30.6 m. Note that me
overall building height is 6.75 m.

I I !
C' c ' ti

<=[|
i
10
i9 ' -
8 7
ill 65
I
4
I
32
I

7. 1. Exterior of the building modelled: (above) minor axis: (under) major axis.

Two finite element models for titis structure were constructed as follows:
- Static model. The initial lateral stiffness of the entire building about the major and
minor axes is obtained. Initial lateral stiffness implies the stiffness of the structure
when subjected to service loads only.
- Dynamic model. The natural frequency of the building about die major and minor
axes is obtained. This implies the determination of me natural frequency at which
side-sway (lateral movements of me entire structure) occur.

109
 SS**

Hi_*f;:-

7.2a. Photo of the buildings exterior: minor axis.

'vW
Ik

-';--

. *T:-

7.2b. Photo of the buildings exterior: major axis.

110
For each model die participation at service load levels of individual building compo
nents against side sway movements is estimated. Bom static and dynamic models were
constructed as a basis of comparison for full-scale test results. In effect, full-scale
testing predicts building natural frequency with precision, while static stiffness must
be estimated.
Structural and non-structural components were included in me finite element analysis.
Non-structural components include die following:
- Cladding. Exterior walls constructed using pre-cast concrete units.
- Partitions. Interior walls constructed of wood and gypsum board diaphragms,
glued togetiier.

Due to the uncomplicated nature of die structural system and die simple building ge
ometry, a two dimensional finite element model was tiiought to be sufficiently accurate
for service load level calculations. Major axis and minor axis analyses are tiius treated
independently. Effectively, structural and non-structural components were treated as
series of parallel frames and shear walls. Each frame and shear wall is connected at die
ground, first floor and roof levels.
The first floor slab consists of lightweight precast concrete units witii a cast-in-place
wearing surface. The roof consists of two dimensional trusses, cross-bracing and a
deep-ribbed tiiin-walled steel decking. Both me first floor slab and roof provide a
substantial degree of in-plane shear stiffness to die structure.
Predictions of lateral movements are made at me first floor and roof levels, in botii me
major and minor axes.

Modelling of structural components

A plan view at die first floor level and at die roof level, taken from die structural
drawings, is shown in figure 7.3. In diese views die general framing plan is indicated
(column locations, floor beams and horizontal bracing systems). Note mat frames in
me minor axis are numbered 1-1 to 1 1-11. Frames in the major axis are numbered A-A
to E-. Elevation views of typical framing schemes in die major axis are shown in fig
ure 7.4. Elevation view of typical framing schemes in the minor axis are shown in fig
ure 7.5.
Stractural calculations indicate that die building was basically designed to resist vertical
dead and live loads. The only lateral load design case considered is a minimum value
equal to 1% of the vertical live load, applied at me roof level. These loads were con
sidered to be resisted by diagonal cross-bracing provided in exterior frames and in die
roof.

One particularity of mis structure is its foundation. The structure occupies die site of a
previous building, of which, a number of existing piles were re-used. Re-use of exist
ing piles required die designer to modify some column emplacements. This may be
noted in the plan view of die ground and first floor structural drawings (see figure 7.3).

111
 I ir ]D C ] S' t E D' D C B B' A
" I I I I I I 1

1 . 1 1

2 2 2
3 3 3
4 4 4
5 5 5
6 6
7 7 7
8 8 8
9 9
10 10 10

'11 a
11 I I 1 1

E D' D C B B' A E D' D C B B' A

7.3. Plan view of structural details at (left) the first floor level and (right) the roof level.

V. D b' e

//ssy?/sy/A\y;/w//sy/Av//.wssy//sy/A<y/sisy/)!sy/Asy,

y/A\y?AsyyA\yy/sy//sy//^/^y/Asy//sy//syyA>^/>syyj^y,

7.4. Elevation view of structural details in th major axis: (above) Frames A- and E-; (under)
Frames B-B to D-D.

1 Z 3 4 5 6 7 8 9 10 .11

/SWAW/AWAV/AW//WSJWWS/&SAWS}WSWSA>?&W//&S/Sy/XWX\

11 10 98 7 6 5 4 3 2 1

_H
//Aws/sy/w/Asys/sysAsysxv/ws/srsAS&Asysiw?ws*sy?/&>/sy/Xv/A'\

7.5. Elevation view of structural details in the minor axis: (above) Frames 1-1 and 11-11; (under)
Frames 2-2 to 10-10.

112
 The building's foundation is typical for the Nemerlands: friction piles, driven to resis
tance at about 20m in sand and sandy-clay sediments. For analysis purposes die foun
dation is assumed to be perfectiy rigid, bodi vertically and horizontally. Actual hori
zontal foundation rigidity is expected to be somewhat less man titis gross simplifica
tion implies.
Column supports consist of base plates bolted to the caps of the concrete pilings.
Typical examples of these details are shown in figure 7.6. The bolt pattern suggests
diat only limited moment connection is provided. Of particular note, however, die
columns were placed in holes in an existing ground slab. These holes were subse-
quendy filled witii concrete. This suggests tiiat some moment resistance may be pre
sent as no concrete cracking at column bases is observed.

--

--
4- --

7.6. Typical column support detail.

___ _

7.7. Typical beam-column connections: (left) half-height connection plate; (right) stub beam de
tail.

In die major axis most beam to column connections are provided using half-height
connection plates. A typical example of this detail is shown in figure 7.7. The resulting
beam to column connection provides litde moment resistance, and is modelled as a pin
ended connection. In die minor axis, most beams are continuous and connected to die
columns using cantilevered stud beams. A typical example of this detail is shown in

113
figure 7.7. Beam are bolted to me stub beams, resulting in a flexible beam to column
connection. These connections are modelled as pin ended.
An endoscope was used to check cross-bracing and cross-bracing to column connec
tions. In some cases cross-bracing the effectiveness of me cross-bracing was ques
tionable.
Each structural model (about the major and minor axes) were analysed using the
DIANA finite element package. Two-dimensional linear calculations in me X-Y plane
were performed using six-node L6BEN beam elements. Nominal section properties
(cross-sectional area and moment of intertia) were used. A modulus of elasticity E of
210,000 N/mm and Poisson's ration equal to 0.3 were used to define basic steel
characteristics. No plastification is expected tiius plastification post yield criteria were
not specified.

Modelling of interior partitions

The locations of interior partitions on die ground and first floors are shown in figure
7.8. All partitions consist of 100 mm thick gypsum board. Partitions were not in
cluded in the static model because of tiieir weak attachment witii die stractural system.

_"o:

L_____r_-_
7.8. Plan view of interior partitions at (left) the ground floor level and (right) the 1st floor level.

Modelling of exterior cladding

A elevation view of the cladding is shown in figures 7.1 and 7.2. Openings between
panels are left for windows and doors. Each panel was measured to be 150 mm thick
and are made of minimally reinforced light-weight concrete.
An upper bound estimation of actual cladding stiffness was assumed in die model. It is
assumed that no movement occurs between individual cladding panels or between
panels and tiieir foundation. Upon visual inspection no evidence of panel cracking,

114
 cracks between panels or cracks at the foundation level were evident Windows and
doors openings are included in the model.
Each model (about the major and minor axes) was analysed using the DIANA finite el
ement package. Two-dimensional linear calculations in the X-Y plane witii eight-node
CQ16M plate elements were performed. A nominal thickness of 15 cm for all elements
was assumed. A modulus of elasticity E of 20,000 N/mm2 and Poisson's ratio equal
to 0.25 were used to define basic concrete characteristics. No concrete cracking or
crushing is expected.

Modelling of attachments
The exterior cladding is connected to the stractural system (exterior columns) using
stractural angels (clips), which are bent to fit A typical detail is shown in figure 7.9.
Upon close inspection no evidence of previous movements between structural frame
and panel was observed. Two stractural models are used to described the behaviour of
the cladding attachment. First, no shear transfer between the structure and cladding
was assumed (no interaction). This lower bound estimate consists of the steel frame
alone. Second, the attachments were assumed to be infinitely rigid (full interaction).
This upper bound estimate consists of the steel frame and the exterior cladding.
The interior partitions are glued to me floor slabs. No attachment is provided, how
ever, at the lateral or top edges of me partitions to the stractural frame. All interior
columns are covered with fire protection materials and enveloped in wooden boxes for
architectural reasons. Partitions are butted against mese boxes. Partitions are, how
ever, butted against exterior columns. All partitions are full height, but do not come
into contact with die adjacent floor beams or slab.

Modelling of the buildings mass

For die buildings structural steel components, a steel density of 7,850 kg/m3 was as
sumed. Manufacturers data was used to define the nominal cross-nominal areas of
each steel section. The weight of each L6BEN element is determined by DIANA using
these two parameters.
A lightweight concrete density, equal to 675 kg/m3, is assumed for the cladding.
Cladding weight is thus equivalent to 10.1 kN/m2.
The floor slab consists of a 240 mm tiiick precast light-weight concrete slab and a 30
mm tiiick normal-weight concrete wearing surface. The density of the lightweight con
crete units was estimated as 675 kg/m3, and of the wearing surface, 2000 kg/m3. The
total unit weight of this slab is thus equal to 222 kg/m2. Interior partition weight was
estimated using a gypsum density of 1100 kg/m3. Partitions on the first floor are 3.13
m high, and have a total approximate length of 145 m. This gives an average first floor
partition weight of 79 kg/m2. The total dead weight of the structure at the first floor
level was thus assumed to be equal to 301 kg/m2. This is in general agreement witii the
building's design calculations, which assume a floor dead weight of 310 kg/m2.

115
 ,
'V

'-* . 'V'.

"f*

...AV.
* '-

Ir
7.9. Photos of a typical attachment between the cladding and an exterior column: (above) bend and
drilled; (under) bent only.

116
The building's design calculations specify diat die maximum first floor live load is
equal to 250 kg/m2. At the time of testing it was estimated tiiat die average live load in
the building on die first floor was 13 kg/m2. Thus the total dead plus live load applied
at the first floor level is 314 kg/m2. Building weight was included in the DIANA
model using PT3T translational point mass elements. These elements were placed at
first floor level beam to column joints.
Roof weight was taken directiy from the design calculations, thus assumed to be equal
to 35 kg/m2. The maximum roof live load is stated to be equal to 90 kg/m2. Actual live
loads on the first floor at die time of testing were minimal (at the time of testing die
roof was dry). The weight of the roof was included in the DIANA model using PT3T
translational point mass elements. These elements were placed at roof level beam to
column joints.
The relative weight of individual building composites are shown in table 7.1.

Table 7.1. Relative weight of individual building components.

Building component Percentage of total building weight

bare steel frame 1
cladding 81
1st floor slab 13
partitions 5
live load (1st floor) <1
roof <1

Modelling of lateral loads

Lateral loads were placed at different heights on the structure to determine the effective
stiffness of the structure against side sway. A nominal value of IN was applied at the
level of the first floor and at the roof level. Parallel to the major axis this load was
placed at beam to column connections on frame 10 (see figure 7.5). The predicted val
ues of horizontal drift at the beam to column connections of frame 7 were used to cal
culated side sway stiffness using the following expression:

K"d
where:
K is the side sway stiffness
F is the applied lateral force at frame 10
d is die predicted lateral displacement at frame 7

Parallel to the minor axis loads are placed at beam to column connections of frame D
(see figure 7.4). Lateral stiffness was calculated in a similar manner.

117
7.2 Static model predictions
A static analysis has been performed to determine the lateral sway stiffness of the
structure described in paragraph 7. 1. This has been done for the following cases:
1 . The entire building, assuming full interaction between cladding and the structural

frame with cross-bracing. All columns are assumed to be fixed-ended ground level.
2. The entire building (cladding and stractural frame) witiiout cross-bracing.
3. The entire building (cladding and structural frame), columns pin-ended at ground
level.
4. The structural frame only (including cross-bracing).

From these four models the relative influence of individual parameters on the lateral
stiffness of die entire building may be estimated. These parameters are the following:
- thecladding
- th structural frame.
- in-plane vertical crss-bracing.
- column base fixity.

Parallel to the major axis

Predicted lateral drifts for case 1 (about the major axis) are shown in figure 7.10. In
the left figure lateral load is applied at the first floor level. The lateral drift resembles
that which may be predicted for a shear diaphragm, clearly illustrating the participation
of die exterior cladding. When die lateral load is applied at roof level (right figure), lat
eral drift more closely resembles that which may be predicted for a moment resisting
frame.
For each case (cases 1 to 4) four values of lateral stiffnesses may be derived. These
values have been calculated and are shown in table 7.2. Using the values several con
clusions can be drawn. These conclusions are, of course, limited in use to die building
investigated, and for the loadings investigated, which producing lateral sway in the di
rection of the major axis:
Relative importance of the cladding vs. the structural frame. The model predicts
that the cladding is suffer than the structural frame at service load levels. The cladding
effectively dominates the stractural response of the building at service load levels. The
relative stiffness of the cladding versus structural components increases when the ap
plied load and measured responce are closer to ground level.
In-plane vertical cross-bracing. The model predicts tiiat cross-bracing in the struc
tural frame of the finished building has a small but detectable influence at service load
levels.
Column base fixity. The model predicts that me influence of column base fixity on
lateral behaviour is of no practical importance.

118
 7000- 7000- A. Height OF
ftpputd Load
/ /
t y
6000- Nnoe Axis t /
/ /
5000- / /
/ /
/ /
~ 4000- / /
r
X
/ /
//
3000- //

2000-

1000-

i r r
0 10 20 0 50 100
LATERAL MOVEMENT FOR 1 kN LATERAL MOVEMENT FOR 1 kN
DF APPLIED LOAD <x IO"" rm) OF APPLIED LOAD <x 10"" iw)

7.10. Predicted deformed shape for the base case 1 (base case): (left) load applied at the 1st floor and
(right) load applied at the roof.

Parallel to the minor axis

Predicted lateral drifts for case 1 about the minor axis are shown in figure 7.10. In die
left figure lateral load is applied at die first floor level. The conclusions that may be
drawn from the shapes of diese curves are similar to tiiose for die major axis.
For each case (cases 1 to 4) four values of lateral stiffnesses may be derived. These
values have been calculated and are shown in table 7.3. Conclusions drawn from an
analysis of these values are die same as mose for the major axis.

Table 7.2. Summary of lateral stiffnesses in the major axis (N/mm).

Lateral deformation Lateral load applied at

recorded at Modeltype 1st floor roof
1st floor basecase 7.2-10^ 4.7-105
base case w/o diagonal bracing 6.0- 105 3.9-105
base case, columns bases pinned 7.1-105 4.6-105
structural system only 9.9-104 9.2-104
roof base easel 5.610^ 1.1105
base case w/o diagonal bracing 4.610^ 8.8-104
base case, columns bases pinned 5.5 1.0105
structural system only 9.5104 2.1 -104
X Summary of base case assumptions:
- Full interaction between structural frame and exterior cladding.
- No participation of interior parutions.
- AU diagonal bracing in the structural frame is included.
- AU columns are fixed ended at their base.

119
 Table 7.3. Summary of lateral stiffnesses in the minor axis (N/mm).

Latnai deformation Lateral load applied at

recorded at Modeltype 1st floor roof

1st floor base easel 6.5-10* 5.3-10*

base case w/o diagonal bracing 5.2-15 4.3-105
base case, columns bases pinned 6.3-10* 5.2-10*
structural system only 1.3-105 1.1-10*

roof base easel 5.2-10* 1.4-10*

base case w/o diagonal bracing 4.2-105 9.614
base case, columns bases pinned 5.1-105 1.4-10*
structural system only 1.1-10* 4.2-104
I. Summary of base case assumptions:
- Full interaction between structural frame and exterior cladding.
- No participation of interior partitions.
- AU diagonal bracing in the structural frame is included.
- All columns are fixed ended at their base.

7.3 Dynamic model predictions

A dynamic analysis has been performed to determine the fundamental lateral sway fre
quency of the model described in paragraph 7.1. This has been done for die following
cases:
1. The entire building, assuming full interaction between cladding and die structural
frame with cross-bracing. All columns are assumed to be fixed-ended ground level.
This is the basic structural system as described in paragraph 7.1.
2. The entire building (cladding and structural frame) without cross-bracing.
3. The entire building (cladding and structural frame), columns pin-ended at ground
level.
4. The structural frame only (including cross-bracing).

Natural frequency tends to increase with lateral stiffness, thus a comparison of relative
stiffnesses can be estimated for the following parameters:
; - the cladding
- die structural frame.
- in-plane vertical cross-bracing.
- column base fixity.

In table 7.4 a summary of the predicted lateral sway natural frequencies parallel to die
major and minor axes are given. The results of die dynamic analysis tend to support
die conclusions of die static analysis given in paragraph 7.2.

120
 Table 7.4. Summary of predicted lateral sway modal frequencies (Hz).

Modeltype Major axis Minor axis

basecase 8.4 0.1
base case w/o diagonal bracing 6.5 6.4
base case, columns bases pinned 7.1 7.1
Structural system only 3.1 3:5
1. Summary of base case assumptions:
- Full interaction between structural frame and exterior cladding.
- No participation of interior partitions.
- All diagonal bracing in the structural frame is included.
- All columns aie fixed ended at their base.

7.4 Summary of model predictions

Model predictions indicate that at service load levels die cladding dominates structural
response in bom the major and minor axes. Vertical cross-bracing has a measurable ef
fect on lateral stiffness. Column base fixity, however, has only a small influence upon
lateral behaviour.
It is of interest to note that the predicted stiffness in the minor axis, when the load is
applied at the roof, are larger than similar values predicted in die major axis. This may
be due to the cladding stiffness in die minor axis. In the minor axis there are fewer
openings at die first floor level than in the major axis. The predicted natural frequency
of the building in die major axis and in the minor axes support the static analysis.
Several parameters that may be of significance, were not investigated. These include
die following:
- The lateral ground support conditions of die columns and cladding.
- The shear connection provided betwn mdividual cMding panels.
- The interaction provided by die attachments between the cladding and the structural
system.

7.5 Testing of a two-storey steel-framed office building

The building investigated is the same as that described in paragraph 7.1. The objective
of these tests is to derive actual values of natural frequency and static lateral stiffness.
This is achieved by exciting the building and measuring die response (for example
displacement, velocity or acceleration). Two means of exciting me structure were used:
a hammer and a shaker (eccentric rotating masses).
Structural properties (static stiffnesses and natural frequencies) can be derived by mea
suring me time function of both excitation and response. Time functions are converted
into frequency response functions. The frequency response function contains informa
tion about the natural frequencies, and damping ratio. Comparing the frequency re
sponse functions with matiiematical models, structural parameters as stiffness, mass
and damping can be derived. The full procedure behind die derivation of structural
properties is explained in paragraph 7.7.

121
Field measurements
Field measurements on March 14, 1992 have been made on building 11 of TNO
Building and Construction Research (see figures 7.1 to 7.9). Table 7.5 lists all field
testing equipment used for these measurements.

Table 7.5. Field testing equipment

Number Description Make Type

6 Acceterometas Sundstrand S-700
6 Conditioner for Sunstrand TNO-Bouw C-S-700
6 Displacement transducers Hottinger B-3
6 Conditioners for Hottinger Hottinger KWS3073
6 Amplifiers Hottinger 3576
1 Data acquisition system Bakker 2570
1 Pulse hammer TNO-Bouw 10 kg
1 Conditioner for hammer B&K 2626
1 Mechanical exciter TNO-Bouw 400 kg

The excitation load was applied by means of a 10 kg instrumented hammer or a 400 kg

mechanical shaker. The structure was excited on the first floor in bom major and minor
axis. The exact locations are shown in figure 7.1 1 and summarized in table 7.6.

Table 7.6. Position of applied excitation.

Hammer position Shaker position Direction

C-6 Major
B-6 Major
B-6 Minor
B-8 Minor
C-7 Major
C-7 Minor [

The response of die structure was simultaneously measured on die ground floor, first
floor and roof level by means of acceleration and displacement transducers in die same
direction as die applied force. These values are given in table 7.7. Typically, displace
ment transducers give better results in die low frequency range (1 < f < 20 Hz) and ac
celeration transducers give better results in the high frequency range (f > 20 Hz).

Table 7.7. Positions of response transducers.

Transducer Major axis Minor axis Level

1 B-10 D-6 roof
2 D-10 D-10 roof
3 B-10 D-6 1st floor
4 D-10 D-10 1st floor
5 B-10 D-6 ground floor
6 D-10 D-10 ground floor

122
 1 2 3 45 6 78 9 10 .11

-* *M-

J5,6
y?^^f^//^y^^/^^//^y^^A^y/^yAsyAc^y>^/x<yyxvyx^^yk<y//s.

litas 2A6|

t3,5

EZ> TRANSDUCER
HAMMER
SHAKER

7.11. Locations of excitation and transducers: (above) elevation view; (under) plan view.

All signals were digitized and recorded simultaneously on die data acquisition system.
Afterwards the signals were converted to the frequency domain. This resulted in a
complex frequency response function and in a coherence function between excitation
and response.

Testing using an impact hammer

An impact hammer was used to excite die structure by providing a pulse load. Under
such a condition, tiieoretically in the frequency domain all frequencies have the same
magnitude. In practice, higher frequencies can be suppressed by mounting a rubber tip
on the hammer. In this manner all frequencies above 500 Hz were suppressed. The
hammer was instrumented witii an accelerometer. Applied load is thus determined us
ing the hammer mass and measured accelerations.
For all locations and directions mentioned in table 7.6 and all positions mentioned in
table 7.7 the natural frequency and stiffness of die building was determined using a
circle fit procedure assuming viscous damping (see paragraph 7.7). In the derivation
of me stiffness the influence of higher order modes has been calculated. Furthermore,
the influence on die stiffness due to damping otiier tiian viscous damping has been cal-

123
culated. The fit procedure was carried out using die acceleration and displacement re
sponse signals.
of an example. In ttis example an excitation is
Signal analysis is illustrated by means
placed at position D-6 and the response is measured at the roof level, D-10 (position
2).

1.00
ilK I \-^
0.8O

0.60 ii
J
0.40
I
0.20

0.00
12 16 20 f [hzl

7.12. Coherence function between force and acceleration, file 2, position 2.

Figure 7.12 shows the coherence function between force and acceleration. In figure
7.13 the frequency response function is presented. Good coherence for die frequency
range between 4 and 12 Hz is observed. Below 3 Hz the coherence is poor. Using
only die frequency range with good coherence, a peak value is observed at approxi
mately 7.5 Hz. This value corresponds with the first natural frequency. Figure 7.14
shows the real and imaginary part of the frequency response function. A circle fit is
shown together with the measured data (indicated by a +). Additional circles can be
observed, which correspond to second and higher order natural frequencies.

0.20

0.15

I
UJ
0.10

o.os : I I , '. j . ; . ; . "-

o.oo
12 16 20 f [hz]

7.13. Frequency response function based on acceleration, file 2, position 2.

124
 1.00

-1.00
-1.00 -.60 -0.20 02.0 0.60 1.00 re
. (E-8)
7.14. Circle fit based on acceleratie, file 2, position 2.

Figures 7.15 and 7.16 show the coherence and circle fits for the same example, but
using displacement measurements. Again a good coherence is observed from 4 to 12
Hz. A slightly better circle fit is obtained in this case than that derived from ac-
celerometer measurements.
The results of all circle fit analyses are summarized in tables 7.8 to 7.15. It may be
concluded that the lateral sway stiffness at the roof and first floor levels can be derived
from both displacement and acceleration measurements. As expected, since the first
natural frequency is in the low range, the results using displacement responce are
slightly better than those obtained using accelerations. The lateral stiffness at the
ground level can only be derived from the measured displacements.

1.00

0.80

0.60

. 0.40

0.20

0.00 ^ ' 1 : 1 . ! j ;

0 -* 8 12 16 20: f [hz]

7.15. Coherence function between force and displacement, file 2, position 2.

125
 1.00

0.60

0.20
"

7
.i iL jrZ
S- L...
-0.20

-0.60
K^_
J

-1.00
-1.00 -.60 -0.20 0.20 0.60 1.00 re
(E-8)
7.16. Circle fit based on displacement, file 2, position 2.

Table 7.8. Natural frequencies in Hz derived from accelerations.

Major axis Minor axis

Position axis C-6 C-6 B-6 B-6 axis B-6 B-6 B-8
roof level B-10 8.0 7.9 8.0 8.0 D-6 8.8 8.4 8.3
D-10 7.4 7.9 7.9 7.9 D-10 8.8 8.4 8.3
1st floor B-10 7.9 7.9 7.9 8.0 D-6 8.8 8.4 7.8
D-10 8.0 7.9 7.9 7.9 D-10 8.8 8.4 7.9
ground floor B-10 7.1 7.1 D-6
D-10 7.8 D-10

Table 7.9. Natural frequencies in Hz derived from displacements.

Viajor axis Minor axis

Position axis C-6 C-6 B-6 B-6 axis B-6 B-6 B-8
roof level B-10 7.7 8.0 7.7 7.7 D-6 8.7 8.4 7.9
D-10 7.3 7.6 7.6 7.6 D-10 8.7 8.4 7.9
1st floor B-10 7.7 8.0 7.6 7.7 D-6 8.7 8.4 8.0
D-10 7.3 7.6 7.6 7.6 D-10 8.7 8.4 8.0
ground floor B-10 7.5 7.1 7.1 7.1 D-6 7.1 8.4 8.0
D-10 7.1 7.1 7.1 7.1 D-10 8.3 8.4 7.1

Table 7.10. Damping ratios derived from accelerations.

viajor axis Minor axis

Position axis C-6 C-6 B-6 B-6 axis B-6 B-6 B-8
roof level B-10 0.15 0.16 0.16 0.15 D-6 0.14 0.13 0.14
D-10 0.14 0.17 0.15 0.16 D-10 0.15 0.14 0.19
1st floor B-10 0.15 0.15 0.15 0.14 D-6 0.13 0.12 0.14
D-10 0.16 0.18 0.15 0.15 D-10 0.15 0.13 0.17
ground flow B40 0.12 0.13 D-6
D-10 0.24 D-10

126
Table 7.11. Damping ratios derived from displacements.

Major axis Minor axis

Position axis C-6 C-6 B-6 B-6 axis B-6 B-6 B-8
roof level B-10 0.14 0.15 0.15 0.15 D-6 0.13 0.12 0.14
D-10 0.14 0.16 0.15 0.16 D-10 0.14 0.14 0.17
1st floor B-10 0.14 0.15 0.14 0.14 D-6 0.13 0.12 0.14
D-10 0.14 0.16 0.15 0.15 D-10 0.14 0.13 0.17
ground floor B-10 0.17 0.17 0.16 0.17 D-6 0.37 0.13 0.39
D-10 0.14 0.16 0.16 0.15 D-10 0.24 0.21 0.52

Table 7.12. Stiffness k [109] derived from accelerations.

Major axis Minor axis

Position axis C-6 C-6 B-6 B-6 axis B-6 B-6 B-8
roof level B-10 0.61 0.65 0.45 0.48 D-6 0.56 0.56 0.17
D-10 0.50 0.53 0.67 0.72 D-10 0.74 0.68 0.48
1st floor B-10 0.78 0.96 0.72 0.76 D-6 0.82 0.74 0.85
D-10 0.58 0.65 0.93 0.95 D-10 1.10 1.00 0.66
ground floor B-10 8.90 D-6
D-10 9.30 D-10 5.80 6.80 1.30

Table 7.13. Stiffness k [109-N/mj derived from displacement.

Majoraxis Minor axis

Position axis C-6 C-6 B-6 B-6 axis B-6 B-6 B3
roof level B-10 0.65 0.69 0.37 0.48 D-6 0.59 0.57 0.66
D-10 0.45 0.36 0.63 0.75 D-10 0.74 0.73 0.64
1st floor B-10 0.98 1.10 0.69 0.83 D-6 0.80 0.80 0.77
D-10 0.63 0.20 0.87 1.00 D-10 1.00 1.00 0.56
ground floor B-10 2.50 1.90 1.00 1.10 D-6 8.60 1.00 7.00
D-10 1.00 1.30 1.60 1.00 D-10 7.80 7.50 1.00

Table 7.14. Coefficient of variation of stiffness derived from accelerations.

Major axis Minor axis

Position axis C-6 C-6 B-6 B-6 axis B-6 B-6 B-8
roof level B40 0.011 0.014 0.009 0.009 D-6 0.011 0.009 0.014
D-10 0.015 0.029 0.009 0.008 D-10 0.019 0.011 0.006
1st floor B-10 0.016 0.021 0.018 0.014 D6 0.016 0.010 0.032
D-10 0.017 0.037 0.016 0.010 D-10 0.030 0.012 0.020
ground floor B-10 0.140 0.270 0.130 0.10 D-6 10.600 10200 10.300
D-10 0.300 0.520 0.320 0.25 D-10 0.780 0.590 0.350

Table 7.15. Coefficient of variation of stiffness derived from displacements.

Major axis Minor axis

Position axis C-6 C-6 B-6 B-6 axis B-6 B-6 B-8
roof level B-10 0.012 0.006 0.017 0.009 D-6 0.009 0.012 0.026
D-10 0.011 0.012 0.023 0.014 D-10 0.014 0.08 0.010
1st floor B-10 0.017 0.005 0.022 0.009 D-6 0.010 0.014 0.020
D-10 0.013 0.006 0.025 0.016 D-10 0.011 0.009 0.007
ground floor B-10 0.012 0.020 0.017 0.018 D-6 0.009 0.009 0.011
D-10 0.016 0.010 0.012 0.028 D-10 0.027 10.500 0.016

127
Summary of tests using an impact hammer
The lateral load response, static stiffness, natural frequency and damping ratio u, of an
entire as-build steel-framed building was measured using a hammer excitation. When
measuring displacements response instead of acceleration responce, the stiffness at
ground floor level can be derived even when exciting at first floor level. The coeffi
cients of variation of the estimation of stiffness per hammer blow are in the order of V
= 0.02. In one direction (major or minor axis) several excitations were applied at each
level (first floor or roof). Averaging all hammer blows on one specific level and in one
direction, me coefficient of variation increases to approximately V = 0.20
In table 7.16 lateral stiffness is presented in the same manner as used in die summary
of the finite element model predictions. The mean value and coefficient of variation has
been translated into values witii a 90% confidence leveL

Table 7.16. Summary of stiffness k [109-N/mj, for load applied at 1st floor.

Major axis Minor axis

Deformation recorded at acceleration displacement acceleratimi displacement
roof 0.42-0.74 0.30-0.80 0.21-0.85 0.54-0.77
1st floor 0.57-1.00 0.30-1.20 0.59-1.10 0.55-1.10
ground floor 0.50-2.30 0.00-1.10

Testing using a shaker

A shaker containing of two eccentric rotating masses (rotating in opposite directions)
was used. The structure was excited several times, each time with the shaker running
at a different frequency. For each shaker frequency only one discrete point on the fre
quency response function can be obtained. By testing at several frequencies, several
points in the frequency response function can be obtained, thus the entire function is
approximated. As shown in paragraph 7.7, starting with a measured frequency re
sponse function, structural characteristics can be derived. In the test, the in frequency
steps used were approximately 1 Hz. The frequency response function was detennined
for both the major and die minor axes, using both accelerometers and displacement
transducers. Test results taken from displacement transducer measurement are shown
in figures 7.17 and 7.18 (major axis) and 7.19 and 7.20 (minor axis). The resulting
frequency response functions using displacement transducers lead to reasonable results
for both major as minor axis.

An analysis of structural parameters using this frequency response function is difficult

to make, since there is very little data available near the first natural frequency. This
analysis has, therefore, been ommited.

128
 I? ai

20 f [Hz]

7.17. Frequency response function for major axis based on displacement

( = hammer 3; - = shaker 3).

0.80

0.70

0.60

0.50
1\
If
I
0.40

0.30

0.20 X
.0.10

0.00
0 12 16 20 f [Hz]
7.18. Frequency response function for major axis based on displacement
(shammer 4; = shaker 4).

0.80
'

0.7

0.60

0.50 v
2
> 00
1

m
0.40 v ,C\ 1

T""
0.30
\
KJ
0.20 r , 7- \
0.10

'" V
\zr^
/^s_'-
^TT^-^
0.00 1

8 12 16 20 f [Hz]

7.19. Frequency response function for minor axis based on displacement

( = hammer 3; =shaker3).

129
 It OJ

20 f [Hz]

7.20. Frequency response function for minor axis based on displacement

( -hammer 4;---= shaker 4).

Summary of tests using shaker

a
When using a shaker consisting of countre-rotating masses, measurements should be
made using small frequency steps near die first natural frequency. Applied force was
not measured directly but was calculated using rotating mass eccentricity and rotation
frequency. The results of the shaker test can be improved by directly measuring the
excitation force.

Comparison of hammer blow and shaker testing

The determination of die natural frequency is potentially as accurate using a shaker as
when using a hammer blow. With a hammer blow, however, measurements are both
accurate and quickly obtained. Using a shaker, many measurements must be made at
different speeds, in an effort to obtain the speed which corresponds to the buildings
maximum response. Good examples of comparisons between natural frequencies for
these two methods may be see in figures 7.18 and 7.20 (die major and minor axes of
die building studied at TNO).
The frequency responses obtained using a shaker were not sufficiently accurate to en
able an estimation of the buildings lateral static stiffness. In contrast, good estimates of
static lateral stiffness were obtained from the frequency responses using a hammer
blow. The frequency responses for the two test methods are compared in figures 7.17
and 7.19 (die major and minor axes of me building studies to TNO).
Irregardless of measurement quality, installation and measurement times for the ham
mer blow technique are much shorter than for the rotating mass. The rotating mass it
self is heavy, requiring special equipment to move it into location. In the case of die
TNO building, it was necessary to install the rotating mass using a fork lift truck
through a first storey window.
The following disadvantages are thus sited when using a shaker, as opposed to a
hammer:

130
 - At low frequencies the excitation force is very small. It is mus vety difficult to es
timate static lateral sliffness.
- The shaker is both large and heavy. In some cases it may be necessary to use a
crane to install.
- Testing time is considerably increased.
Disadvantages of using a hammer blow as opposed to a shaker are as follows:
- As the size of the structure increases, so does die necessary energy input For very
large structures it may become difficult to supply sufficient energy using a hammer.

7.6 Comparison of test results and model predictions

An existing steel framed building was botii modelled and tested under service load
conditions. Static lateral load stiffness and die natural frequency were estimated. A two
dimensional finite element model using beam and pate elements was chosen. All as
sumptions used with this model were intuitive. Only information commonly available
to designers was used. This was done for die following reasons:
- The model is typical of tiiose used to determine die ultimate load carrying capacity
of the structural system. This implies that the designer can add non-structural com
ponents to an existing models.
- Design information on parameters such as column fixity, joint stiffness and the real
interaction between structural and non-structural components is not at present
widely used.

All comparisons given here thus represent the variance (for one particular structure)
between typical design models including non-structural components and real structural
behaviour at service load levels.
Experimental errors at service load levels are generally larger than tiiose that may be
expected at ultimate load levels. This is mainly due to th participation at service load
levels of many building components that noimally fail before ultimate loads are ap
plied.
The inherent variability of service load level measurements implies that a sufficient
number of measurements must be made to estimate the standard distribution of
recorded values. This was done by repeatedly applying load at the same location, and
by applying load at several different locations in the structure. The standard deviation
of measured values in the major and minor axes were thus derived for a confidence
level of 90%.
For the building tested, titis results in the following:
- 10% in die major axis
- 25% in the minor axis.
The larger standard deviation in die minor axis is due to a rotational component of
buildings response. This component was not further investigated as it could not be
predicted using die two dimensional finite element model.

131
Major axis
A comparison of measured and predicted values in the major axis are shown in tabel
7.17. Finite element model (FEM) predictions and test results are seen to differ by
10% to 25%. This difference, however, in negligible when compared to die stiffness
gained by including non-structural components, between 250% and 650%. These
gains can be obtained using existing calculation techniques and without adding material
to the existing structure.

Table 7.17. Comparison of measured and predicted values in the major axis (Lateral load applied at the
lstfloor).
Test FEM FEM
Value (real structure) (real structure) (bare steel frame)

Natural frequency 7.7 Hz 8.4Hz 3.1Hz

1st floor stiffness 640-10* N/m 720-10* N/m 9910* N/m
Roof stiffness 430-10* N/m 56010* N/m 95-10* N/m

Minor axis
A comparison of measured and predicted values in the major axis are shown in table
7.18. Finite element model (FEM) predictions and test results are seen to differ by
10% to 20%. This difference, however, in negligible when compared to die stiffness
gained by including non-structural components, between 270% and 550%. Thse
gains can be obtained using existing calculation techniques and without adding material
to the existing structure.

Table 7.18. Comparison of measured and predicted values in the minor axis (Lateral load applied at the
lstfloor).
Test FEM FEM
Value (real structure) (real structure) fljare steel frame)
Natural frequency 8.4 Hz 7.1Hz 3.1Hz
1st floor stiffness 710-10* N/m 650-10* N/m 13010* N/in
Roof stiffness 420-10* N/m 520-10* N/m 11010* N/m

Conclusions
m In general, measured and predicted values correspond to within reasonable confi
dence limits. This is encouraging as it suggests that linear two dimensional finite el
ement model with simple assumptions for column base fixity, beam to column con
nections and connections between structural and non-structural building components
can give reasonable predictions of real in-service building responses to lateral loads at
service load levels.
The inclusion of non-structural element at service load levels can be done by modi
fying existing structural analyses. This can result in very substantial increases in lateral
stiffness without the need to increase structrual member sizes.

132
7.7 Theoretical principles behind building measurements
When load is applied to a structrue by means of a hammer blow, displacements, ve
locities or accelerations are measured. These measurements lead to time functions of
excitation and response (force and displacements). The time functions for both force
and displacement can be converted into frequency functions by means of a FFT (Fast
Fourier Transfer) routine. The frequency response function () can then be calcu
lated by dividing displacement () and force () frequency functions.

() = ^f(
This measured frequency response function gives information on natural frequencies
and damping ratio. When the mathematical model of the building is known stiffness,
mass and damping can be derived. The mamematical model should represent a multiple
degree of freedom system. For linear behaviour, any multiple degree of freedom sys
tem can be represented as a superposition of single degree of freedom systems. Here,
tiiree classes of system models are described:
- viscously damped
- stractural (or hysterically) damped
- Coulomb friction.

The basics functioning of each system will be explained only for viscous damping.

Viscous damping
Assuming that the building behaves linear and tiiat only viscous damping is present, a
single degree of freedom system can be described mathematically by equation (1):

mx(t) + cx(t)+kx(t) = f(t) ('!)

where:
m is the mass
c is the viscous damping
k is the stiffness
x(t) is the displacement
f(t) is the force

A Laplace transformation of equation (1) leads to:

ms2 + cs + k = f (2)

where:
x(t) =xest
f(t) =fest

Wim no external forcing in equation (2) we obtain the condition:

133
ms2 + cs + k = 0

The solution to this equation is:

-cVc2-4km ,~
s=- 2m (3)
or:

= -^ (4)

where:
(Q = Vk/m the natural frequency in [rad/s]
C G
===_- the damping ratio [-]
CO 2/_ 6

The frequency response function () can be expressed by equation (5):

rT. . k-<j>2m + ica) ,c.

()= 7y ~7y (5)
k - arm + (c(D)x

The magnitude of me frequency response function can tiien be calculated from:

()= . * (6a)
V(k-a>2m)2 + (c>)2
or:
j.
() = k (6b)

A/(i-4)2-(2-^)2
V <"*>

By fitting equation (5) or (6) to the measured frequency response function (), the
model properties coq, , k and m can be derived.
The fit procedure is carried out by means of die circle fit procedure. For dus aim the
frequency response function () from equation (5) is converted in a response func
tion for velocity () (impedance function) and divided in a real part and a imaginary
part.

T ; . <(k-<_m)2 y-v
Im((D) =^ 2 s2 , s2 (7a)
(k-ormr + (cmr

Re(Y)= 2^ -2 (7b)
(k-szmr + (ca>r
Then:

134
(Re(Y)-)2 + Im2(Y) = ()2 (8)

Equation (8) describes a circle with radius and centre (,0). By fitting a circle to the
measured imaginary frequency response function (), the centre and radius can be
calculated.
As mentioned in the introduction die tested building can not be described by a single
degree of freedom system. The frequency response function of a multiple degree of
freedom system, however, can be expressed by a superposition of the frequency re
sponse functions of single degree of freedom systems. When the frequencies of the
other modes are not in the vicinity of the first natural frequency, the contribution to
() of die higher natural frequencies is approximately constant (independent of fre
quency).
The contribution of the other modes is expressed in a constant B. For a multiple degree
of freedom system, instead of equation (6) we get:

() = 1
(k-G)2m)2 + (ca>)2
-
+'B (9)

On die modal circle the effect of is a rotation of the circle and a shift of the circle
centre. The structural parameters of the first mode can tiierefore be derived from a
single degree of freedom system, taking me circle radius as a reference. The influence
of otiier modes is estimated from die shift of die frequency response function.
The derivation of the stractural parameters is explained by figure 7.21.

7.21. Example circle fit

Having a circle as shown in figure 7.16, die angle r can be derived from:

135
 2
1 --
-^
tan|e=2 (10)

Differentiating with respect to we get

-1
dO _ 2
dco 2

1+^
(____)2

When = cq, d_/d<u reaches a maximum value. The term dO/dca corresponds witii me
rate at which die locus sweeps around me circular arc. So die natural frequency coq can
be found at die maximum value of the sweep rate dc/d6.
The damping of die tested structure can be derived supposing:
- 'a' is point on die circle above the natural frequency o
- 'b' is point on the circle below the natural frequency .
Then:

l--2
tan|Ob=-^ (12a)
_____

tan|ea=--^ (12b)
2_

The damping ratio can then be estimated by combining equations (12a) and (12b):

.: <>-< ' (13)

2u30((atanjea + (ptan-jOb)

The stiffness k can now be found via equation (14):

k.- (14)

where 'r' is die radius of die circle.

136
The influence on the stiffness contributed by me other mode shapes is estimated by the
shift of die imaginary of the circle. The contribution to die stiffness of the other modes
is:k'=l/B.

Structural (or hysteric) damping

For multiple of freedom systems, also models using structural damping can be used
[Ewins, 1986]. The damping rate varies inversely with frequency c = h/. Instead of
equation (6b) we get:

() = 2JL f^ (15a)
k-ctrm + i(h)
1
() = - (15b)
1 -( )2+

where me stractural damping loss factor equals

A circle for frequency response function can be described by

dm(H)-)2 + Re(H)2 = ()2 (8)

The term can be estimated by

2(^-;)
^=~2 ^i -
(oitan-jOa + tan|Gb)
(16)

So the stiffness k equals:

k = -i--
2

Coulomb friction
Coulomb or dry friction can occur in combination with viscous damping. The system
then behaves nonlinear. In case Coulomb friction the frequency response function
varies witii me applied force, die damping parameter becomes:

*_. 2R 2
=+:
naar

Where:
R is the colomb friction
a is die exatation amplitude.

137
To derive the parameters of such a model: mass, stiffness, viscous-damping and
colomb-friction is impossible since there are many combinations of parameters which
lead to die same result

Reliability of the derived parameters

When calculating structural parameters experimental error must be estimated. For in
stance die circle will never fit exactly to the measured data. The discrepancy between
fit and measurement can for instance be due to the fact tiiat the model witii viscous
damping is not correct This can be due to coulomb friction and stractural (or hysteric)
damping.
The reliability of the parameters is here expressed by me coefficient of variance of the
parameters.

y __


Where:
V is die coefficient of variance
is die standard deviation
is die mean value.

The standard deviation of the fit is estimated by the discrepancy between fitted and
measured frequency response function. The standard deviation of die flexibility can be
estimated by:

t2
s
f (Hj-H)2
"____ n-1
i=l

References
DJ. Ewins, Modal Testing. Theory and practice. Research Studies Press Ltd, England, 1986.

138
8 Evaluation of the actual structural
behaviour at service load level of full-
scale structures by in situ dynamic tests

A reliable assessment of the stractural response via computer simulation, requires that
analytical models capable of approximating die main phenomena affecting the system
performance are available. This is particulariy true with reference to die analyses aimed
at evaluating the response under service loads. Serviceability is in general a critical
limit state governing the design of steel stractures.
As outlined in [Castoldi et al. 1987; Kobayashi, 1987], dynamic in-situ testing repre
sents a viable approach to an appraisal of the influence of the various factors affecting
tiie in service behaviour. It was then selected, in order to assess the influence of fram
ing continuity on the structural stiffness and to consequently provide indications useful
for design analyses.
This chapter reports die results of in-situ dynamic tests and numerical analyses carried
out at the University of Trento on two different types of new buildings: die first one
constructed with steel columns and steel-concrete floors, die second one built entirely
out of steel.
Two dynamic exciters were used witii regard to the different masses of the two build
ings. Physical results (in terms of acceleration and power spectrum) are compared with
those of some numerical models obtained assuming different boundary conditions and
internal continuity between die elements.
Conclusions are drawn concerning the capability of in-situ dynamic test memods to
correctly estimate the stractural behaviour at service load levels.

8.1 Structures investigated

Composite structure
The first structure investigated, briefly described in figures 8. 1, 8.2 and 8.3, is the
new domestic arrivals hall of the passengers aerostation at Milan-Linate Airport The
building, named 'Corpo-Ovest', has been designed to bear the nominal loads of aero
station halls on the first floor and of parking places on the second. It is essentially a
rectangular building with a basement and one above-ground floor; the totl building
dimensions are 64.96 by 36.96 by 7.89 m (only 5 m above ground).
The floor structure consists of precast concrete units witii load-reducing polystyrene
blocks and with a cast-in-plac concrete wearing surface. The floor thicknesses are
0.45 and 0.50 m respectively for spans of 8 and 12 m.
In correspondance with die columns lines the floors are built with no load-reducing
polystyrene blocks.

139
 1st and 2nd floor building plan
7.5
gq1 -' V- /

8
-
\ Building

8
\ !
36.96 .

bT 12
\ Tb

8 I 8 J 8 I 8

64.96m

foundatiouns plan

a a a s a m

37.48 a a a aa a a

aaaaaaaa
bT

8.1. Plans of composite structure.

Section A-A Detail A

36.96 / Detail B

0.4 0.6 0.5 0.6 0.5, 0.5

__9 ,0.5 ' ,0.45 " /"
+3.89 ,0.5 ,0.45
^77777777?
4 \- 4
_s _k_ a
37-46

Section B-B
: ___ SiAS _ .. : .

1 , +7.89

:j + +-<"> I
I
I

'///,

.+ + I

elel-lelelelele
I
1
64.73

8.2. Sections of composite structure.

140
 +7.89

c
o
o
o

Detail A
Side Columns .

.00

0.48
"""1 +7.89
U5T

sE

**t
_

Detail
Inner Columns -%-z%-

8.3. Beam-to-beam and beam-to-column connection details.

The floors are supported by die bottom flange of longitudinal 8 m-long steel beams
which are included into die final layer of concrete. The beams are made with HE 400B
and HE 450B steel shapes and are simply bolted to die columns or to stub beams
welded to the columns; at the column lines there is no continuous steel reinforcement
on the beams.
The columns are continuous from the foundations to the second floor and are con
structed with HE 400B and HE 450B steel shapes. The loads are transmitted to the
columns by means of welded stub beams on the first floor and on the internal columns
of the second one; on the external columns of the second floor the beams are continu
ous over the columns so that the loads are directly trasmitted (see figure 8.3).

141
The foundations are made with bearing micropiles connected to independent footings
for the internal columns and to a grade beam for the external ones; at the same time the
grade beam is connected to a retaining wall which reaches the first floor (see figure
8.2).

Steel framework
The second structure investigated, briefly described in figures 8.4, 8.5 and 8.6, is a
part of the steel roofing structure of the new parking area of the passengers aerostation
at Milan-Linate Airport. The steel roofing structure, whose total dimensions are 48 by
36 by 3.87 m, is based on a part of a composite reinforced concrete-steel building
named 'Corpo-Sud'.
The structure consists of six five-bay portal frames connected with secondary beams.
The portal frames are constructed witii longitudinal HE 300B and 400 main
beams for the 8 m spans, and with welded beams for the 12 m spans. The secondary
beams are 8 m long IPE 270. The columns are constructed with HE 240A and HE
400B shapes.

4. m

8. m

8. m

8. m

L
8. m
-
J'
12. m

I 1 1

8. m 4. m 8. m 2. m

8.4. Plan of steel roofing structure.

142
 Section A-A

Dfttr tAifj

- s
Section B-B

03

1 12. 1 " 8 I 8 8 . ? 1 4

8.5, Sections of the steel roof structure.

4- - f -f HEB 300 or IPE 400 or -welded beam

X
IPE 270 370

+ > f -- (b)

(a)

(c)

8.6. Details of: (a) column-base joints; (b) rubber joints; (c) longitudinal beam-to-column joints.

In die transversal plane, normal to the plane of the main portal frames, the horizontal
deflections are limited by six single-bay, 'nominally' pinned, portal frames constructed
witii HE 260B beams and welded columns. In the longitudinal plane die horizontal
deflections are limited by the main portal frames and by a brick wall along die first two
bays of the A alignment (see figure 8.4).
In die roof plane a cross-bracing system, made with 28 mm diameter steel bars, pro
vides the necessary in-plane stiffness (see figure 8.4). The roofing consists in deep-
ribbed tiiin-walled steel decks riveted to me structure.

8.2 Numerical models and results

Numerical model and results for the composite building
A 3-D finite element model was set up to evaluate the natural vibration frequencies of
the structure as a whole. The model is made with beam type elements only and is
shown in figure 8.7.

143
 2

t^j>

8.7. Finite element model of composite structure.

Due to die presence of the concrete retaining wall, connected to the grade beam and to
the first floor, die basement of the building was considered perfectly rigid. By this as
sumption die displacements of the first floor were neglected and only the second story
columns and the second floor were modeled. In addition, die connection between the
second floor and the existing buiding was considered rigid respect to horizontal dis
placements. For the purpose of die analyses, the slab was modeled using cross-brac
ings connected to die top end of die columns.
The mass of the building elements was computed using die shop drawings. For the
steel elements a density of 7850 kg/m3 and manufacturers' cross-nominal areas were
assumed. For me slab and the reinforced concrete beams a density of 2500 kg/m3 was
used. At the time of testing no live or dead weights were present
With reference to the internal constraints between the elements, the shop drawings
show diat the beams and the top end of die columns were designed as pin-ended. In
spite of this, die presence of the concrete wearing around the beams and of bolted con
nections between beams and columns (see figure 8.3) suggests different end con
straints between die elements. With reference to the column-bases rotational con
straints at first floor, die presence of the concrete suggests an increase in the theoretical
elastic stiffness.
For these reasons, die internal continuity degree between the elements and the rota
tional constraints at the column-bases were assumed as unknown parameters.
Numerical analyses were earned out for die following cases:
- base-pinned columns and full continuity in beam-to-beam and beam-to-column
joints (case 1);
- base-fixed columns and no continuity in beam-to-beam and beam-to-column joints
(case 2);
- base-fixed columns and full continuity in beam-to-beam joints (case 3);
- elastically base-restrained columns and full continuity in beam-to-beam and beam-
to-column joints (case 4);
- base-fixed columns and full continuity in beam-to-beam and beam-to-column joints
(case 5);

144
The results of the analyses performed in terms of natural vibrating frequencies are
summarized in table 8.1.

Table 8.1. Vibrating frequences.

Natural Frequencies [Hz]

Case model mode 2 mode 3
1 1.87 4.37 4.96
2 2.10 4.18 4.79
3 2.16 4.42 5.05
4 3.04 4.70 5.51
5 3.48 5.08 .. 5.81. . fo

The numerical vibrating modes obtained from the analyses are the same for the five
cases considered and are essentially horizontal ones. For case 5 they are shown in fig
ure 8.8.

U
(a)

(b) (c)

8.8. Composite structure: (a) 1st fundamental mode; (b) 2st fundamental mode; (c) 3rd fundamental
mode.

Numerical model and results for the steel framework

A 3-D finite element model, which also makes use of beam type elements only, was
adopted for the numerical study of the steel roofing structure; it is shown in figure 8.9.
Due to the fact that the structure rests on a very large and heavy reinforced concrete
building, the boundary constraints were assumed to be perfectly rigid with respect to
horizontal movements. In the plane of longitudinal multi-bay portal frames and of
transversal single-bay ones, full continuity was considered at the top end of the
columns and at the ends of beams due to die presence of moment resisting connec
tions. In contrast, the secondary beams were considered as simply supported

145
 Y^s

8.9. Finite element model of steel structure.

For the purpose of the analyses die brick wall (along the A alignment in the first two
bays) and die profiled steel sheeting were modeled using cross-bracings connected to
the nodes of die model.
As regard the above structure, the masses were computed assuming a density of
7850 kg/m3 for the steel elements. For the brick wall a density of 1850 kg/m3 and
manufacturers' cross-nominal areas were assumed.
At die time of testing no live or dead weights were present. The numerical analyses
were carried out considering die contribution of the bare structure, of the non-struc
tural components and of the column-bases constraints. The following cases were con
sidered (table 8.2):
- die bare structure with base-pinned columns (case 1);
- as above with the contributions of the brick wall and of the steel sheeting (case 2);
- as above but witii base fixed-end columns (case 3).

Table 8.2. Analyses results.

Natural Frequencies [Hz]

Case model mode 2 1 mode 3
1 2.27 3.08 j 4.27
2 2.44 3.32 4.58
3 | 4.39 5.95 il 6.62

The first two modes obtained from numerical analyses are the same for the three cases
considered and are essentially horizontal vibrations along the transversal direction; for
case 3 they are shown in figure 8.10. The third mode is again an horizontal vibrating
mode for cases 1 and 2, while it is a vertical one,
of the cantilever part of the structure,
for case 3.
Comparing the results of cases 1 and 2 it is possible to note that for the investigated
structure the brick wall does not play an important role on the natural frequencies: this
is due to the fact that vibrations related to the natural frequencies are essentially or
thogonal to die plane of die wall.

146
 (a)
(b)

8.10. Steel structure: (a)Tst fundamental mode; (b) 2st fundamental mode.

8.3 Physical tests

Tests description
Physical tests on botii stractures were performed to derive the actual values of the nat
ural frequencies. The tests were carried out by applying a sine force at pre-established
points by means of a mechanical exciter, and by recording die accelerations at signifi
cant points of the structures by piezometric accelerometers. The data was collected by a
recording system linked to a function analyser allowing an on-line check of stractural
responses. Accelerations were recorded in relation to the stationary response of the
stractures for several forcing frequencies and for the free vibrating transients.
For the composite structure, a 200 kg counterrotating masses dynamic exciter was
used. The tests were performed by arranging the dynamic exciter so as to provide a
sine horizontal force in the plane of the second floor and by recording the horizontal
accelerations at the top of the columns. The location of the exciter and of the ac
celerometers is shown in figure 8.11.
7.5

M8 6? Existing
Building

\
36.96

\
12
,
\ 4
5_

8 | 8 | 8 | 8

> 64.96m

- Accelerometer and direction of measurement

m Dynamic exciter and direction of excitation

8.11. Location of exciter and accelerometers in the composite structure.

147
For the steel roofing structure a 50 kg dynamic exciter was used due to the lower total
mass. The tests were performed by arranging the dynamic exciter so as to provide a
sine horizontal force near die roof plane, in both directions separately, and by record
ing the horizontal accelerations at the top of the columns. Figure 8. 12 shows the loca
tion of the exciter and of die accelerometers. The forcing frequency ranges from 1 to 8
Hz for both stractures.

- Accelerometer and direction of measurement

Cp Dynamic exciter and direction of excitation

8.12. Location of exciter and accelerometers in the steel structure.

Results for the composite structure

in tenns of natural frequencies, the results are shown in figure 8.13 which illustrates
the power spectra of accelerations recorded by accelerometers 1-8 during the free vi
bration transients. From figure 8.13, it can be observed mat die actual first two natural
frequencies are respectively 3.25 and 6.1 Hz. Moreover, it can be seen from the
magnitude of die natural frequencies, that accelerometers 4, 6 and 7 are influenced
more considerably by die second mode dian by the first one.
Figures 8.14 and 8.15 show the accelerations recorded during the stationary response
of the structure for forcing frequencies of 3.3 and 6.45 Hz. By analyzing the ampli
tudes and die phases of each set of data a cantilever movement of the structure along
the y axis at 3.3 Hz and a torsional movement at 6.45 Hz can be detected.

148
 PUR SP A han LIN PMR SP8 han UK 20Hz
X 3,230Hz Y: X: 3.300Hz Y: .802E-3 G

PUB SP B han LIN 20Hz

X: 3.250Hz Y: .59BE-3 G

.864 .435
-3 E-3

UwJwN"*-KjiffiTff-t-tni M
PUB SP A han LIN
m ri.i r
20Hz
ivi 0 _
0
^^JS^itUJ^JL^
PKB SP B han LIN 20Hz
X: 6.100Hz Y: .2S8E-3 G X: 6.100Hz Y: .131E-3 G

.891
E-3

HAG
G

jUfelltf
PMR SP A han LIN 20Hz
X: 3.250Hz Y: .386E-3 G

8.13. Free accelerations power spectra for the composite structure.

Results for the steel framework

In terms of natural frequencies, the results are shown in figure 8.16 which illustrates
the power spectra of accelerations recorded by accelerometers 1-8 during the free vi
bration transients. From the figure, it can be observed that the actual first two natural
frequencies are respectively 4.45 and 5.45 Hz.
Moreover, it can be seen from the magnitude of the natural frequencies, that ac
celerometers 5 and 6 are influenced by the first mode and that accelerometers 1, 2, 3
and 8 are influenced by the second one. The responses of accelerometers 4 and 7
clearly indicate that they are placed in correspondance of nodal points of the lowest
natural vibrating modes. Figures 8.17 and 8.18 show die accelerations recorded dur
ing the stationary response of the structure for forcing frequencies of 4.5 and 5.5 Hz.
By analyzing the amplitudes and the phases of each set of data a cantilever movement
along die axis Of die bottom part of the structure at 4.5 Hz and of the upper part of
the structure at 5.5 Hz can be recognized.

149
 TIME A LIN

TIME A LIN

.348
E-l

REAL W^^^^>a^*s/>*%A^*^^*^^^W^*^'Wl'^^*^^,^'
G

-.348
E-l
TIME A LIN TIME 8 LIN

8.14. Recorded data at a forcing frequency of 3.3 Hz for the composite structure.

8.4 Comparisons
Comparisons for the composite structure
Comparing the actual first natural frequency and die vibrating mode (figures 8. 13 and
8.14) with the results of numerical models (table 8.1 and figure 8.8a) it can be seen
diat the actual natural frequency is contained between die values of cases 4 and 5 and
that me actual vibrating mode is fully consistent witii tiiose obtained by die same mod
els.
Comparing the results relative to the actual second natural frequency and the vibrating
mode (figures 8.13 and 8.15) with those of numerical models (table 1 and figure 8.8b)
it can be seen that this frequency is greater than numerical ones and also mat the actual
vibrating mode is not consistent witii mose of me same models.

150
 .349
E-l

REAL
G

-.349
-i
TIME A LIN eSEC TIME B LIN

.357 .354
E-l E-l

REAL
G

-.357 -.354
E-l E-l
TIME A LIN

.346
E-l

REAL
S

-.348
E-l

^wmmmmmmmm
-.357
E-l.

8.15. Recorded data at a forcing frequency of 6.45 Hz for the composite structure.

These results seem to state diat the structure behaves like a frame with full continuity in
beam-to-beam and beam-to-column connections while die column-bases are elastically
restrained with springs whose rotational stiffness are greater than the nominally elastic
ones. Indeed, the lack of consistence founded in the second mode is probably due to
the difficulties in evaluating some boundary restraints (i.e. the actual restraint of the
connection with the existing building) which play in effect a more important role in the
second mode dian in the first one.

Comparisons for the steel structure

Comparing the actual first natural frequency and the vibrating mode (figures 8.16 and
8.17) with the results of numerical models (table 8.2 and figure 8.10a) it can be seen
the actual natural frequency is satisfactorily consistent with die value obtained from
case 3 and the actual vibrating mode, recorded at 4.5 Hz, is fully consistent witii that
obtained from the same model.

151
 .442 .447
E-3 E-3

HAG HAG
G G

PHR SP A han LIN PHR SP B han LIN

X: 5.450Hz Y: X: 5.450Hz Y:

PHR SP A han LIN PHR SP B han

X: S. 450Hz Y: X: S.4S0HZ

PHR SP B han
X: 4.450Hz

.442 .460
E-3 E-3

MAG
MAG
G
6

PMR SP A han LIN PHR SP B han LIN

5.450Hz Y: X: S. 450Hz Y:

8.16. Free accelerations power spectra for the steel structure.

Comparing the results relative to the actual second natural frequency and the vibrating
mode (figures 8.16 and 8.18) widi those of the case 3 numerical model (table 8.2 and
figure 8. 10b) it can be observed tiiat die actual frequency is slightly overrated but die
actual vibrating mode is fully consistent widi tiiat obtained from the same model.
The full consistency between numerical and experimental results states that the struc
ture behaves like a frame with full continuity between the elements of portal frame
structures and also in the connections at die base of die columns.

8.5 Conclusions
An investigation was carried out into whether die actual stractural behaviour of
buildings at service load levels can be evaluated by means of simple in-situ dynamic
tests. The comparison between experimental and numerical results at service loads
shows tiiat almost full continuity can be considered in the in service analytical models

152
 .434 .447
E-2 E-2

REAL REAL
G
>m<gt>IMWIW>HW il'*"! HvMlllillHim
S m+W'trmrmtti*mmr^m&i^^tyH*i*
*<W*i

-.434 -.447
-2 E-2
TIME A LIN TIME B LIN

.436
E-2

REAL SEAL
G
'II' "Hi UH Htr** ' W> lWMHM'M> ..WWWW S

-.436 -.432
E-2 E-2
TIME A LIN TIME . LIN

.435 .452
E-2 E-2

REAL REAL
G

-.435 -.452
E-2 E-2
TIME A LIN TIME 8 LIN

ACC. 7

E-2
'.
REAL
G

-.442
E-2
TIME A LIN TIME B LIN

8.17. Recorded data at a forcing frequency of 3.3 Hz for the steel structure

of both structures. This information, if properly recognized in numerical models,

enables the global stractural behaviour for 'service load levels' to be evaluated without
an excessive design burden.
The results obtained provide useful design indications, though it should be stressed
tiiat they are correlated to die type of tested stractures.
Moreover, the results are related to the global behaviour of the building and not to the
behaviour of single connections; consequently me detected qualitative behaviour of the
joints can be used only for the evaluation of the global behaviour of the building, like
side-way displacements, and not for the evaluation of the behaviour of single ele
ments.
Finally, it is important to underline that numerical models must be able to decribe
the full set of nodal displacements and correctly evaluate the mass distribution on the
structure: therefore, generally, 3-D numerical models are necessary.

153
 .434 .447
E-2 E-2

REAL REAL
G G

-.434 -.447
E-2 E-2
TIME A LIN 20SEC TIME 8 LIN

.436 .432
E-2 E-2

REAL REAL
G G

-.436 -.432
-2 E-2
TIME A LIN TIME 8 LIN

.435 .452
E-2 E-2

REAL REAL
G G **"***>>

-.435 -.452
E-2 e-2
TIME A LIN 20S EC TIME 8 LIN
X: 2.051 SEC Y: .S31E-5 G

.442 .460
e-2 E-2.

REAL REAL,
G

-.442 -.460
E-2 E-2
time A LIN TIME 8 LIN

8.18. Recorded data at a forcing frequency of 6.45 Hz for the steel structure.

The authors wish to express their deep thanks to Fratelli Dioguardi S.p.A., die finn
responsible for die construction work, whose availability and kind collaboration has
made it possible to carry out die tests.

References
A. Castoldi, A. Chiarugi, G. Giuseppetti and M. Fanelli, In-Situ Dynamic Tests on Ancient
Monuments, IABSE Colloquium on Monitoring of Large Structures and Assessment of their Safety,
Vol. 56, 1987, p. 131-141.

A.S. Kobayashi (editor), Handbook on experimental mechanics, Englewood Cliffs, NJ., 1987.

154
9 Summary of major findings

The major findings of this report have been summarised. The purpose of this summary
is to give die reader a clear view of the work that has been completed and its practical
implications. This summary does not include conclusions and recommendations for
further research. These are presented in chapter 10.

9.1 Design codes and serviceability problems in steel-framed buildings

The literature survey conducted on serviceability showed tiiat the issue is recognised to
be an important design consideration most national and international standards and
codes of practice. This reflects increasing concerns about building quality as opposed
to stractural safety. Design rules for serviceability, however, are not based upon sys
tematic research. They represent common beliefs based upon past experience and
practice and are thus written to give designers as much freedom of action as possible.
A systematic compilation of serviceability related problems is not available. This is
partly due to legal considerations The main problem, however, is the complexity and
magnitude of work that is needed. Tools for measuring long term effects such as
foundation movements, creep, etc. are well understood. Using them, however, re
quired die systematic observation of a large number of structures over a period of sev
eral years to obtain useful data. Only recently have tools been developed which enable
researchers to measure the instantaneous behaviour of a real structure. Both types of
tools must be used to separate structurally related serviceability problems non-strac
tural serviceability problems (such as the effects of water penetration, chemical attack;
etc.).
The review conducted throughout the project showed a rich literature but rather patchy.
More and more computer software are including serviceability aspects in their struc
tural analyses. However, further testing of real stractures in needed in order to include
die correct parameters of me non-stractural element in the analysis.

9.2 Measured in-service behaviour of steel-framed buildings

The real behaviour of steel-framed buildings at serviceability load levels can be quite
different from that predicted by ultimate limit state design models. This is illustrated by
figure 9. 1 for a typical low rise office building. Ultimate limit state models normally
include only stractural components (die bare steel frame). These models may be linear
(first order) or non-linear (higher order). A substantial amount of research has been
conducted concerning model predictions and building response to loads at ultimate

155
limit states. It is thus not surprising that the ultimate limit state of a real structure
should be relatively well predicted.
At service load levels, however, such models can be excessively conservative. This
project has shown that real building stiffness can exceed die bare steel frame stiffness a
factor of 6. As shown in figure 9.1, a first-order bare steel frame structural model,
modified to include non-structural components (exterior cladding and interior parti
tions), can reasonable predict measured building response.

ln ORDER ANALYSIS
INCLUDING STRUCTURAL AND
NON-STRUCTURAL CQHPDCNTS

ln ORDER ANALYSIS OF
THE BARE STEEL FRAME

HIGHER ORDER ANALYSIS

OF THE BARE STEEL
FRAME

LATERAL SWAY
9.1. A first-order bare steel frame structural model can reasonable predict measured building re
sponse.

The effects of ignoring non-stractural components at service load levels have several
implications. These are listed in table 9. 1. In some cases ignoring non-structural com
ponents can lead to an overly conservative design. In other cases, ignoring non-struc
tural components can be the root cause of serviceability problems.
Two rational alternatives are possible. First, non-stractural components may be in
cluded in design models at service load levels. To do this the initial in-plane stiffness
of the cladding, partitions and attachments between stractural and non-stractural com
ponents must be know. The designer then must check model predictions for each
cladding, partition and attachment to ensure that imposed forces (or deformations) re
main within reasonable limits.
Secondly, the designer may choose to isolate structural and non-stractural compo
nents. This can be done by providing attachments which transfer minimal force until
allowable lateral drift have been exceeded. This approach is typically used for glass
curtain walls in high-rise office buildings.

156
 Table 9.1. The effects of ignoring non-structural component participation.
Effect Result
increased lateral stiffness As the importance of nn-structural components increases, the force that
they are subjected to increase accordingly. High forces may result in un
foreseen serviceability problems.
structural response An apparently symmetric structure may in reality behave non-symmetri-
cally. The resulting torsional response may cause serviceability prob
lems.
decreased lateral drift Steel members may be oversized to limit lateral drifts that do not occur.
Ibis increases building costs for no gain. Oversizing steel members may
not eliminate serviceability problems as non-structural components are,
m general, much more rigid than the steel frame.

The idealised curves shown in figure 9.1 are not typical for all steel framed buildings.
In table 9.2 die service load behaviour of different types of steel framed buildings and
the economic importance of serviceability considerations are shown. Serviceability is
indicated to be of economic importance if existing design rules require increasing steel
frame member sizes. Serviceability is Of least economic importance if existing design
rales require no action to be taken.

Table 9.2. Serviceability considerations for different steel framed building types.
Building type Behaviour under service load Economic importance of
conditions serviceability requirements
low or medium-rise office or dominated by non-structural low
residential buildings components
high-rise office or residential dominated by structural com high
buildings ponents
long span commercial and variable high
industrial buildings

From table 9.2 we may conclude that even though non-structural components are ca
pable of vastly reducing lateral drift in low or medium rise office and residential
buildings, existing serviceability requirements do not penalise steel construction. No
reduction in building cost would thus be expected. For long-span and high-rise struc
tures the cost reduction potential is high. The potential to use existing cladding to re
duce costs in long-span stractures is thought to be the greatest Use can be made Of me
well known principles underpining stressed skin design. For high-rise stractures exist
ing cladding can not in most cases be used. However, modified cladding has recently
been designed to control lateral drift in an existing high rise office building.

9.3 Design models at serviceability load levels

In the past, deflection limits recommended by codes were empirically established, and
by no means, represent an indication of die actual in-service performance. Their effec
tiveness and reliability is heavily dependent upon die structural model adopted by the

157
designer. Economically it is of great importance that a designer can re-use ultimate
limit state models to perform service limit state calculations. Therefore, serviceability
limit models are traditionally defined using elastic models including of the bare steel
frame with ideal restraint conditions.
The numerical analyses presented in the previous chapters, however, clearly indicate
that joint, cladding and panel action can have a substantial influence on structural re
sponse at service load levels. Incorporating of these aspects in the design process may
provide substantial economic benefits at serviceability and ultimate strength load lev
els. A good example of such effects is light steel cladding. Significant advantage may
be achieved just by accounting for its presence. These advantages often overshadow
increases in frame flexibility associated witii the use of semi-rigid joints in lieu of rigid
joints. A first-order elastic analysis would be adequate to justify such advantages.
An area requiring more investigation is including the effects of cladding and panels
widi the bare steel frame. Two ways of doing this have been explored, each witii par
ticular advantages and disadvantages, as indicated in table 9.3.

Table 9.3. Advantages and disadvantages of different calculation methods including the effects of
cladding at service load levels.

Method Advantage Disadvantage

Diagonal truss Simple Not flexible. Difficult to account for
elements geometric irregularities such as panel
aspect ratio, doors and windows.
Plate elements More complicated. Not all designers have Flexible. Geometric irregularities are
access to programs witii plate elements. easily accounted for.

The basic behavioural feature to be satisfactorily approximated is die in-plane shear

stiffness of the cladding and panel at service load levels. This is a radier simple task,
due to die predominandy linear elastic response. In fact concrete cracking or crushing,
steel plastification, tearing or buckling and attachment movements etc. should not oc
cur at service load levels. These represent the principle phenomena inducing non-lin
earity, and thus, may be neglected.
It has been observed that the in-plane shear stiffness of cladding and panels is strongly
related to the shear stiffness of their base material(s). Models adopting a F.E. mesh
witii plate elements possess a distinct advantage due to their adaptability: variations in
in-plane shear stiffness due to die panel aspect ratio, small or large openings are auto
matically included. Potential problems arise mainly when modelling the interconnec
tion between cladding, panels and die bare steel frame.
The reliability of diagonal truss models is hampered by two problems: the limited
number of existing guide-lines, and by uncertainties of how to incorporating the ef
fects of aspect ratio, openings and geometric irregularities into their formulation. These
two factors are in fact inter-dependent On the other hand, it should be emphasized tiiat
in many instances the effect of me cladding or panel action is so large tiiat even a rather
inaccurate model may be sufficient for design purposes.

158
Furthermore, in situ dynamic tests confirm the important role played even by the
'single' framing connections. A satisfactory appraisal of the stractural stiffness was in
fact achieved when full continuity was considered within the framework and at the
baseplates.

As concluding remarks, it should be stressed diat the present state-of-the art concern
ing the behaviour of buildings at service load levels is rapidly improving. At me same
time, numerical models available to designers are rapidly increasing in sophistication
and decreasing in cost These simultaneous developments will have a significant im
pact on future serviceability assessments. It thus seems that the time has arrived to re
vise present serviceability limit state criteria. The first step in dus direction should be to
provide a direct link between design models and a steel framed buildings real be
haviour at service load levels.

159
10 Conclusions and recommendations

Major conclusions from this study are listed as follows:

There is little useful information contained in existing literature concerning the fol
lowing subjects:
- Serviceability problems in existing steel framed buildings.
- The basis of existing serviceability limits for the design of steel framed buildings.
- Serviceability design models for steel framed buildings.
- The initial stiffness of non-stractural elements and connections between structural
and non-stractural elements commonly used with steel framed construction.
- Serviceability limits to the strength or deformation capacity of non-stractural ele
ment and connections between structural and non-structural elements commonly
used with steel framed construction.
Non-stractural components dominate the static structural response to lateral loads
at service load levels for many types of common steel frames buildings. This was ob
served by dynamic testing and predicted by modelling.
Existing code provisions for serviceability are inadequate. The relationship be
tween actual building behaviour at service load levels, design models and serviceability
limits remains unknown.

10.1 Recommendations
The following recommendations are proposed:
Full-scale buildings tests should be continued to estimate the actual static response
to lateral loads of a variety of common steel framed buildings.
Literature research, and the compilation of data sheet information, should be con
tinued for the initial stiffness of common non-structural elements and connections be
tween structural and non-stractural elements. This search should also include strength
or deformation limits to serviceability for each component investigated.
The relationship between actual building behaviour at service load levels and de
sign models should be established. Design models should include both stractural and
non-structural components. A wide range of common steel frame building types and
uses should be included in this study.

160
 New code provisions for serviceability should be proposed for use in EC3: Part I:
General rules and rules for buildings. These rules should included the following:
- Acceptable types of analysis. These models should be as similar as possible to
those used for ultimate load calculations.
- Mormation on the modelmg of non-stractural components.
- Load levels to be used for serviceability calculations.
- New serviceability limits.

161
 11 Literature list for all related references

This chapter contains a list of all references that were reviewed during the redaction of
this report References used in individual chapters) are identified by brackets. The
numbers in the brackets correspond to the chapters in which the reference appears.
References that are of interest but that do not appear in the text are included here for
completness. For mese references no chapter number is given.

ACI Committee 435, 'Allowable deflections', Journal of the American Concrete Institute, 1968, p.
433-444.

(2) ACI Committee 442, 'Response of buildings to lateral forces', Journal of the American Concrte
Institute, 1971, p. 81-106.

(2, 3) Ad Hoc Committee on Serviceability Research, Committee on Research of the Structural Division,
'Structural Serviceability: A critical appraisal and research needs', Journal of Structural Division,
ASCE. Vol. 112, No.12, December 1986, p. 2646-2664.

D.E. Allen and W.R. Schriever, Structural failures. Modes, causes, responsibilities (ASCE publica
tion), ASCE National meeting on Structrual Engineering, Cleveland, Ohio, April 1972, p. 21-47.

D.E. Allen, Serviceability and structural standards, Institute for Research in Construction, National
Research Council of Canada, Ottawa, Ontario, Canada (CIB-W85).

DJB. Allen, Floor vibrations from aerobics, Institute for Research in Construction, National Research
Council of Canada, Ottawa, Ontario, Canada, 1990, (CIB-W85).

(6) American Institute of Steel Constuction, Load & Resistance Factor Design. Manual of Steel
Construction, First Edition, USA, 1986.

(5) J.E. Amrhein, R.H. Hatch and M.W. Merrigan, Anchor connections of stone slabs, Fifth North
American Masonry Conference, June 1990, p. 1417-1424.

(6) D. Anderson, Elastic Analysis of Semi-Rigid Steel Frames, Research Report CE/17, Department of
Engineering, University of Warwick, 1985.

(6) D. Anderson, F.S.K. Bijlaard, D.A. Nethercot and R. Zandonini, Analysis and Design of Steel Frames
with Semi-Rigid Connections, International Association for Bridge and Structural Engineering IABSE
Survey S-39/87, April, 1987.

G. Andreaus, M. Ceradini, M. Cerone and P. D'Asdia, Masonry columns under horizontal loads: A
comparison between finite element modelling and experimental results, Proceedings of the 7th interna
tional brick masonry conference, Melbourne Australia, February 1985, p. 469-478.

G. Andreaus, M. Ceradini, M. Cerone and P. D'Asdia, A finite element model for the analysis of ma
sonry structures under cyclic actions, Proceedings of the 7th international brick masonry conference,
Melbourne Australia, February 1985, p.479-488.

162
(2) ASCE Committee on Design of Steel Building Structures, 'Wind drift design of steel-framed build
ings. State-of-the-art report', Journal of the Structural Division, ASCE, Vol. 1 14, No.9, Sep.1988, p.
2085-2108.

(6) A. Astaneh and MJi. Nader, 'Design of Tee Framing Shear Connections', Engineering Journal,
American Institute of Steel Construction, First Quarter, 1989.

(5) R.H. Atkinson, B.P. Amadi, S. Sb and S. Sture, 'Response of Masonry Bed Joints in Direct
Shear', Journal of the Structural Division, ASCE. Vol. 115, No. 9, February 1989, p. 2276-2296.

R.H. Atkinson, JJL. Noland and G.C. Hart, Properties of masonry materials of limit states design,
Proceedings of the 9m international brick masonry conference, Berlin, October 1990, p. 678-685.

(3) Australia Standards AS4100.I-1990. Steel Structures Code, 1990.

(2) W.F. Baker, Sizing techniques for lateral systems in multi-storey buildings, Proceedings Tall
Buildings: 2000 and beyond, p. 454-554, November 1990.

R. Bares, Loading tests of building structures, ACSE National Structural Engineering Meeting,
Cleveland, Ohio, April 1974.

Bases for design of structures. Serviceability of buildings against vibration, ISO/DIS 10137, ISO/TC
98/SC2, 1990.

Bases for the design of structures: Deformations of buildings at the serviceability limit states,
International Standard ISO 4356, First edition, 1977-11-15.

(2) A.A. Bat, V.A. Otstavnov and L.L. Lemysh, On deflections and displacements, CIB W85 Reference
Document R91: IB, 1991 (Translated from Russian).

(6) J.R. Benjamin and H.A. Wiliam, 'The behaviour of one storey reinforced concrete shear walls',
Proceedings of the American Society of Civil Engineers, 1957, 83 (ST5) Paper 1254.

(2) R. Bergmann, Structural serviceability aspects of building envelopes in tall buildings, Proceedings of
the symposium/workshop on serviceability of buildings, National Research Council, Canada, p. 293-
322, May 1988.

(6) C. Bernuzzi, R. Zandonini and P. Zanon, 'Rotational behaviour of End Plate Connections',
Costruzioni Metalliche, Vol. 2, 1991, p. 74-103.

F. Bljuger, On serviceability requirements regarding deformations, National Building Research

Institute, Technion - Israel Institute of Technology, Haifa, Israel (CD3-W85).

(6) E.R. Bryan, The stressed skin design of steel buildings, Crosby-Lockwood-Staples, 1972.

(2) E.R. Bryan and JJM, Davies, Stiffening effect of light cladding, Proceedings of the International Cof.
on Planning and Design of Tall Buildings. Lehigh University, Beehem, Pennsylvania, August 21-
26, 1972, p. 643-651.

(3) BS 5950, The use of structural steel in buildings (1990). British Standard Institution.

B. Calderoni, Lab tests on small scale brick masonry buildings, Proceedings of the 9th international
brick masonry conference, Berlin, October 1991, p. 600-611.

163
 GJM. Calvi, G. Macchi and P. Zanon, Random cyclic behaviour of reinforced masonry under shear ac
tion, Proceedings of the 7th international brick masonry conference, Melbourne Australia, Febuary
1985, p. 1249-1256A.

(2) Canadian Institute for Scientific and Technical Information, Deformation requirements for buildings,
joint publication of die Steelbuilding Association, Rotterdam and die Concrete Association,
Zoetenneer. Report NRC/ CNR TT-1969. Canada, Ottawa, 1980.

(5) J.M. Carney, 'Bibliography on Wood and Plywood Diaphragms', Journal of the Structural Division,
ASCE. Vol. 101, No. ST11, February 1975, p. 2423-2436.

(8) A. Castoldi, A. Chiarugi, G. Giuseppetti and M. Fanelli, In-Situ Dynamic Tests on Ancient
Monuments, IABSE Colloquium on Monitoring of Large Structures and Assessment of their Safety,
Vol. 56, 1987, p. 131-141.

CEB, Manuel de calcul 'Flches'. 2e Partie. Limitation desflches, 1973.

LE. Chandler, Defects and Repairs to Brick Panels Walls in High Rise Buildings, Fifth North
American Masonry Conference, June 1990, p. 1397-1416.

(2) F. Charney, DISPAR for SAP. A post processor for the SAP90 Finite Element Analysis Program,
Advanced Structural Concepts Division, J.R. Harris and Company, Denver, Colorado, 1990.

(2) F. Charney, Sources of elastic deformation in laterally loaded steel frame and tube structures. Design
methods based on stiffness, J.R. Harris and Company, Denver, Colorado, 1990.

(4) M.S. Cheung and J. Khan, Premature Deterioration of Building Enclosures in High-Rise Structures,
Council on Tall buildings and Urban Habitat, Tall Buildings: 2000 and Beyond. Fourth World
Conference, Hong Kong, November 1990, p. 201-214.

(2) C.-K. Choi and M.S. Bang, 'Plate Element with Cut-out for Perforated Shear Wall', Journal of the
Structural Division, ASCE, Vol. 113, No. 2, Februaiy 1987.

(3, 6) Commission of the European Communities, Eurocode 3. Design of steel structures. Part 1. General
Rules and Rules for Buildings, 1991.

(6) Commision of die European Communities, Eurocode 4. Design of Steel Concrete Composite
Structures. Parti. General Rules and Rules for Buildings, 1992.2

(2) R.C. Cooney and A.B. King, Serviceability criteria for &HiWmgs, Building Research Association of
New Zealand. BRANZ study report. Report SR14, 1988.

(6) E. Cosenza, A. De Luca and C. Faella, 'Elastic Buckling of Semi-Rigid Sway Frames', Stability and
Strength, Vol. 8, Structural Connections, London 1988, p. 253-296.

(6) E. Cosenza, A. De Luca and C. Faella, 'Inelastic Buckling of Semi-Rigid Swy Frames', Stability
and Strength, Vol. 8, Structural Connections, London, 1988, p. 297-335.

E. Cosenza, C. Greco and M. Pecce, 'Nonlinear Design of Reinforced Concrete Continuous Beams',
IABSE, Structural Engineering International 1/91, 1991, pp. 19-27.

(5) J.W. Cowie and P. Ameny, Replacement ofdefective precast concrete wall panels with brick masonry
veneer without relocating building tenants, Fifth North American Masonry Conference, June 1990, p.
1371-1384.

164
 R.A. Crist and JJt Shaver, Deflection Performance Criteria for Floors, National Bureau of Standards,
Report PB-252 170, April, 1976.

G. Dallaire, 'Inspecting Buildings: Are we doing a good enough job?', Civil Engineering, January
1983.

(6) J.B. Davison, P.A. Kirby and D.A. Nethercot, Rotational Stiffness Characteristics of Steel Beam-to-
Column Connections Joint Flexibility in Steel Frames, London, 1987, p. 17-54.

(6) J.B. Davison, D. Lam and D.A. Nethercot, 'Semi-rigid action of composite joints', The Structural
Engineer, Vol. 68, No. 24, December 1990, p. 489-499.

(5) Jl. Dawe and R.T. McBride, Experimental investigation of the shear resistance ofmasonry panels in
steel frames, Proceedings of the 7th international brick masonry conference, Melbourne Australia.,
February 1985, p. 791-801.

(5) Jl. Dawe and T.C. Yong, An investigation offactors influencing the behaviour of masonry infill in
steelframes subjected to in-plane shear, Proceedings f the 7th international brick masonry conference,
Melbourne Australia, February 1985, p. 803-814.

Deformations in multi-storey steel structures. Gravitational and Temperature Deformations, European

Convention for Constructie Steelwork (ECCS), 1981.

(5) Dhana, Page and Kleeman, The behaviour of'brick masonry under biaxial stress with particular refer
ence to infilled frames, Proceedings of the 7th international brick masonry conference, Melbourne
Australia, February 1985, p. 815-824.

(3, 5) L. Douhan, Deformation limits for roofs and walls for profiled sheeting, Swedish Council for
Building Research, Document D32: 1980, Stockholm, Sweden, 1980.

(2) P. Dubas, Interaction ofstructural elements with cladng, Proceedings of the International Conference
on Planning and Design of Tall Buildings. Lehigh University, Bethlehem, Pennsylvania, p. 675-683,
August 21-26, 1972.

(5) J.T. Easley, M. Foomani and R.H. Dodds, 'Formulas for Wood Shear Walls', Journal of the
Structural Division, ASCE. Vol. 108, No. ST11, November 1982, p. 2460-2478.

J.Eibl and R. Rsch, 'Schwingungsprobleme in einem FuszbaUstadion', Bauingenieur (1990) 65, p.

307-311.

B. Ellingwood, Serviceability Guide-lines for Steel Structures, National Steel Construction

Conference, Florida, USA, June 1988.

B. Ellingwood and A. Tallin, 'Structural Serviceability. Floor Vibrations', Journal of Structural

Engineering, ASCE. Vol. 110, No. 2, Febuary, 1984, p. 401-418.

P.E. Eriksson, Vibration of low-frequency floors - Office buildings and shopping centres, Chalmers
University of Technology, Gteborg, Sweden, 1990, (CIB-W85).

(7) DJ. Ewins, Modal Testing. Theory and practice, Research Studies Press Ltd, England, 1986.

(5) J.B. Fallon, 'Stainless steel in building', Steel Construction Today, VoL 6, No.l, January 1992, p.
7-11.

(5) S.G. Fattal and L. Cattaneo, Evaluation of Structural Properties of Masonry in Existing Buildings,
National Bureau of Standards, Building Science Series 62, 1977, p. 1-117.

165
(5) A. Filiatrault, 'Static and Dynamic Analysis of Timber Shear Walls', Canadian Journal of Civil.
Engineering, Vol. 17, No. 4, August 1990, p. 43-651.

M. Fintei, 'Deflections of High-Rise Concrete Buildings', American Concrete Institute Journal, ACI.
Vol. 72, No. 7, July 1975, p. 1637-1652.

(3) JJM. Fisher and M-. West, Serviceability design considerations for low-rise buildings, American
Institute of Steel Construction (AISC), 1990.

N. FitzSimons and A. Longinow, Commentary on recommended practice for load tests on existing
buildings andfull-scale prototypes, ACSE National structural engineering meeting, Cleveland, Ohio,
April 1974.

F.R. Fossas, Implications of structural characteristics on failure, Proceeding of the International

Conference: The Mexico Earthquake - 1985, Mexico City, 19-21 September 1986.

(2) R. Gaiotti and B.S. Smith, 'Stiffening of moment-resisting frame by precast amerete cladding',
Precast/Prestressed Concrete Institute (PCI) Journal, September/October 1992, Vol.37, No.5, p. 80-
92.

(3) T.V. Galambos and B. Ellingwood, 'Serviceability limit states. Deflections', Journal of Structural
Engineer, ASCE, No. 112, Jan.-Apr., 1986, p. 67-84.

(3) T.V. Galambos, P.L. Gould, M.K. Ravindra, H. Suiyoutomo and R.A. Crist, Structural deflections.
A Uterature and state-of-the-art survey, U.S. Department of Commerce, National Bureau of Standards.
Building Science Series 47, October 1973.

T.V. Galambos, T.A. Reinhold and . Ellingwood, 'Serviceability limit states: Connection slip',
Journal of the Structural Division, ASCE. Vol. 108, no. ST12, December 1983.

F.K. Garas, JX. Clarke and G.S.T. Armer (editors), Structural Assessment. The use offull and large
scale testing, London, 1987.

(5) F. Gartner, Design Aspects of Curtain Walls of High-Rise Buildings, Tall Buildings: 2000 and
Beyond, Fourth World Congress, Hong Kong, November 1990, p. 307-316.

(2) D.A. Gasparini, L.W. Cuny and A. DebChaudbury, 'Damping of Frames with Visco-elastic In-fill
Panels', Journal of the Structural Division, ASCE, May 1981, Vol. 107, No. ST5, p. 889-905.

WJ. Gerasch, 'Reduzierung von beim Tanzen erzeugten Bauwerksschwingungen durch

Schwingungstilger', Bauingenieur (1990) 65, p. 313-318.

(5) Glass in Building. Design and Cb/isTracion, Building Research, May-June 1967.

(2,3) D. Golembiewski, On the limitation of lateral displacement in designing steel frames,

Seminar/Workshop on serviceability limit states for steel buildings. IABSE Woiking commission ,
Zurich, April 1, 1989, p. 8-15.

(5) C.T. Grimm, 'Strength and Related Properties of Brick Masonry', Journal of the Structural Division,
ASCE. Vol. 101, No. STI, January 1975, p. 217-233.

(5) CT. Grimm, 'Metal Ties and Anchors for Brick Walls', Journal of the Structural Division, ASCE.
Vol. 102, No. ST4, April 1976, p. 839-858.

(5) AK. Gupta and G.P. Kuo, 'Modelling of a Wood-Framed House', Journal of the Structural Division,
ASCE. Vol. 113, No. 2, February 1987, p. 241-259.

166
(5) A.K. Gupta and G.P. Kuo, 'Wood-Framed Shear Walls with Uplift', Journal of the Structural
Division, ASCE. Vol. 113, No. 2, February 1987, p. 260-278.

(5) RJvI. Gutkowski and AX. Castiflo, 'Single- and Double Sheathed Wood Shear Wall Study', Journal
of the Structural Division, ASCE. Vol.114, No. 6, February 1988, p. 1268-1284.

G.C. Hart, R.E. Englekirk and T.A. Sabbi, Limit state design of masonry in the United States,
Proceedings of the 7th international brick masonry conference, Melbourne Australia, Febuary 1985, p.
1417-1425.

(5) G. A. Hegemier and G. Krishnamoorty, An Experimental Study of Concrete Masonry Under Seismic
Type Loading, National Bureau of Standards. Building Science Series 106, September 1977, p. 114-
154.

A.C. Heidebracht and S.B. Smith, 'Approximate Analysis of Tall Wall-Frame Stractures', Journal of
the Structural Division, ASCE. Vol. 105, No. 578, August 1979, p. 637-652.

(2) M. Holicky and G. Deak, Basic principles of the control of serviceability, Draft for CIB Programme
W85, February 1989.

M. Holicky, Fuzzy criteria ofserviceability limit states, (CB3-W85).

(2) W.M.G. Ho and SX. Chan, On the effect of non linearity and joint flexibility in lateral drift determi
nation of steel buildings, Hong Kong Polytechnic, Department Of Civil and Structural Engineering,
Research Report CE/010491, April 1991.

(2, 4) M.W. Huggins and JJD. Barber, 'Building deflection, distortions and vibrations. A survey', Canadian
Journal of Civil Engineering, March 1982, VoL 9, No.l, p. 133-137.

Institute of Structural Engineers, Conference of the correlation between calculated and observed
stresses and displacements in racHre, London, Sqrtembr 1955.

Institute of Structural Engineers, Report of a committee on the testing of structures, London,

September 1964.

(5) Italsider, Indagine sperimentale sulla duttilit di pannelliparete in lamiera grecta. Il comportamento
delle strutture portentidi acciaio dteazoinisismiche,Monograpayno.7.

(4) J.R. Janey, ACSE: Structural failures: modes, causes, responsibilities, ACSE National meeting on
structural engineering, Qeveland, Ohio, April 1972, p. 11-20.

(2) A.P. Jeary, BJE. Lee and P.R. Sparks, The determination of Modal Wind Loads from Full-Scale
Building Response Measurement, The International Conference on environmental forces on enginer-
ing buildings, held at Imperial College, London, July 1979.

(2) A.P. Jeary and B.RJEUis, A Study of the Measured and Predicted Behaviour of a 46-Storey Building,
The International Conference on environmental forces on engineering buildings, held at Imperial
College, London, July 1979.

(2) RJP. Johnson, Deflection of continuous beams in steel frames with composite floors of steel and con
crete, Seminar/Workshop on serviceability limit states for steel buildings. IABSE Working commis
sion , Zurich, p. 3, April 1, 1989.

J.C. Kariotis, R.D. Ewing and A.W. Johnson, Strength determination and shearfailure modes ofunre-
inforced brick masonry with low strength mortar, Proceedings of the 7th international brick masonry
conference, Melbourne Australia, February 1985, p. 1327-1337.

167
 H.-G. Kessler, Limits of the Shear Carrying Capacity of Masonry, Proceedings of the 9th interna
tional brick masonry conference, Berlin, October 1990, p. 790-799.

(6) N. Kishi and W.F. Chen, 'Moment-Rotation Relations of Semi-Rigid Connections with Angles',
Journal of Structural Engineering, Vol. 1 16, No. 7, July 1990.

(8) A.S. Kobayashi (editor), Handbook on experimental mechanics, Englewood Cliffs, NJ., 1987.

T. Kubota, S. Okamoto, Y. Nishitani and S. Kato, A study of structural behaviour of brick panel
walls, Proceedings of the 7ui international brick masonry conference, Melbourne Australia, February
1985, p. 1281-1290.

(2) U. Kuhlmann, Slender footbridges, Seminar/Workshop on serviceability limit states for steel build
ings, IABSE Working commission , Zurich, April 1, 1989, p. 55.

RJL Leicester and L. Pham, On developing an Australian limit states code, CSIRO Division of
Building, Construction and Engineering, Melbourne, Victoria, Australia, 1990, (CB3-W85).

(4) RJL Leicester and L. Pham, Serviceability limits, The Institution of Engineers, Australia, 1987
(CIB-W85).

R.H. Leicester and F.D. Beresford, A probabilistic model for serviceability specifications, CSIRO
Division of Building, Construction and Engineering, Melbourne, Victoria, Australia, 1990.

(2) T.C. Liauw and S.W. Lee, On the behaviour and analysis of multi-storey infilled frames subject to
lateral loading, Proceedings of the Institution of Ovil Engineers, Part 2, Vol. 63, September 1977, p.
641-656.

T.C. Liauw and K.H. Kwan, 'Unified Plastic Analysis for Infilled Frames', Journal of the Structural
Division, ASCE. Vol. Ill, No. 7, July 1985, p. 1427-1448.

(6) T.C. Liauw, 'Steel Frames with Concrete Infills', Stability and Strength, Vol.: Steel-Concrete
Composite Structures, London, 1988, p. 115462.

RJ. Mainstone, 'On the stiffness and strengths of infilled frames', Proceedings of the Institute of
Civil Engineers, 1971, Supplement P^er 7360 S, 1971, p. 57-90.

(2) B.F. Maison and CE. Ventura, 'Dynamic Analysis of Thirteen-Story Building', Journal of the
Structural Division, ASCE, Vol. 117, No. 12, December 1991, p. 3783-3803.

S.A. Magd, El. Bbu and I.A. MacLeod, 'Experimental Tests on Brick Beams Under In-Plane Bending
Conditions', The Structural Engineer, Vol. 58B, No. 3, September 1980, p. 62-66.

D.V. Maffick and R.T. Severn, 'The behaviour of infilled frames under static loading', The institute of
Civil Engineers, Vol. 38, December 1967.

P. Marek and WJ. Venuti, On Combinations of Load Effects, San Jose State University, San Jose,
California, USA, 1990.

J. Mathez, Recommendations relative to permissible deformations and deflections of structures, CIB

Committee W23.

(2) MIS. May, 'Determination of collapse loads with and without openings', Proceedings of the Institute
of Civil Engineers, Part 2, No. 71, March 1981, p. 215-233.

168
 (6) FJM. Mazzolani and F. Sylos Labini, L'effetto pelle nelle strutture sismo-resistenti di acciaio (in
Italian), Proceedings of the DC National Conference on Steel Construction (C.T.A.), Perugia, 1983.

(6) F.M. Mazzolani and V. Piluso, 'Skin effect in pin-jointed steel structures' (in English), Ingegneria
sismica, Anno VH, No. 3, 1990, p. 30-47.

(5) WJ. McCulcheon, 'Racking Deformations in Wood Shear Walls', Journal of the Structural Division,
ASCE. Vol. Ill, No. 2, February 1985, p. 257-269.

P. AD. MilL Fundamental Considerations Prior to Diagnosing Total Building Performance.

(4) JJS. Minor, K.C. Mehta and J.R. McDonald, ACSE: Structural failures: modes, causes, responsibili
ties, ACSE National meeting on structural engineering, Qeveland, Ohio, April 1972, p. 49-73.

(8) Monitoring ofLarge Structures and Assessment of their Safety, IABSE, vol 56, 1987.

(6) D.A. Nethercot and R. Zandonini, 'Methods of Prediction of Joint Behaviour: Beam-to-Column
Connections', Stability and Strength, Vol. 8, Structural Connections, London, 1988, p. 23-62.

D . Newland, An introudction to random vibrations and spectoral analysis, New York, 1984.

S. Ohlsson, Springiness and human-inducedfloor vibrations, Swedish Council for Building Research,
Stockholm, Sweden, 1988.

S. Ohlsson, Stiffness Criteria and Dynamic Serviceability of Light-Weight Steel Floors, Chalmers
University of Technology, Gothenburg, Sweden (CIB-W85).

S. Ohlsson, Serviceability criteria, especially floor vibration criteria, 1991 International timber engi
neering conference Londen, (CIB-W85), 1991.

S. Ohlsson and PJE. Eriksson, Structural Serviceability of commercial steel buildings. General as
pects and vibrations, 1991 Nordic Steel CoUoquium, Odense (CIB-W85), 1991.

L. Ostlund, Structural performance criteria, Joint Committee on Structural Safety (CIB-W85), 1990.

P.C. Perdikaris and R.N. White, 'Shear Modulous of Precracked R/C Panels', Journal of the
Structural Division, ASCE. Vol. 1 1 1, No. 2, February 1985, p. 270-289.

G. Peraica, 'Dynamic load factors for pedestrian movements and rhythmic exercises', Canadian
Acoustics 18(2)3-18(1990) (CIB-W85), 1990.

O. Pfefferman, Fissuration de maonneries, Centre scientifique et technique de la construction,

Technical Information Note No. 65, 1967.

O. Pfefferman, 'Les fissures dans les constnictions', Annales de l'Institute technique du btiment et des
traveauxpublics,No.250, 196%, p. 1453-14S2.

E.O. Pfiang and F.Y. Yokel, Structural evaluation of a building system, Building Science Series 25,
Building Research Division, National Bureau of Standards.

(6) S.V. Ployakov, On the interaction between masonry filler walls and enclosing frame when loaded in
the plain of the walls, Earthquake Engineering Institute, San Francisco, 1960.

(6) C Poggi, 'A Finite Element Model for the Analysis of Flexibly Connected Steel Frames',
International Journal for Numerical methods in Engineering, Vol. 26, 1988.

169
(6) S.V. Polyakov, On the interaction between masonry filler walls and enclosing frame when loaded in
the plain of the walls, Earthquake Engineering Institute, San Francisco, 1960.

MX. Porta-, A.M. Wolde-Tinsae and M.H. Ahmed, Behaviour of composite brick walls, Proceedings
of the 7th international brick masonry conference, Melbourne Australia, February 1985, p. 876-887.

(2) M. Qamaruddin, A.S. Araya and B. Chandra, 'Dynamic Response of Multistory Brick Buildings',
Earthquake Engineering & Structural Dynamics, Vol. 13, Mar.-April 1985, p. 135-150.

(5) J. A Richards, RP. Kerfoot and Drueger, "Wood shear panels bonded with flexible adhesives', ASCE,
Journal of the Structural Division, VoLlOl, No, Stl, January 1975, P 131-149.

(5) J.R. Riddington and M.Z. Ghazili, Hypothesis for shear Failure in Masonry Joints, Proceedings,
Institute of Qvil Engineers. Vol. 89, Part 2, March 1990, p. 89-102.

(5) W. Samarasinge, A.W. Page, and A.W. Hendry, 'Behaviour of Brick Masonry Shear Walls', The
Structural Engineer, Vol. 59B, No. 3,September 1981, p. 42-48.

(2) K. Sato and M. Yoshida, Predicted and observed vibration of high-rise buildings and related human
sensitivity during typhoon, Seminar/Workshop on serviceability limit states for steel buildings,
IABSE Working commission , Zurich, April 1, 1989, p. 18-25.

(2) JJ3 . Scalzi and A. Arndt, Plate wall cladding, Proceedings of the International Conference on Planning
and Design of Tall Buildings, August 21-26, 1972, Lehigh University, Bethlehem, Pennsylvania, p.
653-665.

(2) RJ. Schmidt and R.C. Moody, 'Modelling Laterally Loaded Light-Frame Buildings', Journal of the
Structural Division, ASCE, Vol. 115, No. 1, January 1989, p. 201-217.

(2) AJi. Schultz, 'Approximating Lateral Stiffness of Stories in Elastic Frames' Journal of the Structural
Division, ASCE, Vol. 118, No. 1, January 1992, p. 243-263.

(3) SIA 161, Stahlbauten, Edition 1979.

(2) B.P. Sinha and A.W. Hendry, 'Structural testing of brickwork in a disused quarry', Proceedings,
Institute of Civil Engineers, Part L Vol. 60, February 1976, p. 153-162.

BP. Sinha and A.W. Hendry, 'The stability of a five storey brickwork cross-wall structure following
the removal of a section of a main load bearing wall', Proceedings of the Institute of Structural
Engineers, October 1971, p. 467-474.

(2,4) A.W. Skempton and D.H. MacDonald, 'The allowable settlements of buildings', Journal of the
Structural Division, ASCE, May 1956, p. 727-784.

(2) B.S. Smith and C. Carter, 'A method of analysis for infilled frames', Proceengsofthe Institution of
Civil Engineers, Vol. 43, September 1969, p. 31-48.

State Committee of the USSR for Construction, Loads and effects, Supplements, section 10,
Deflections and Displacements, Moscow (CB3-W85), 1989.

(2) B.S. Stafford Smith and B. Carter, 'A method of Analysis for infilled frames', Proceedings of the
Institute of Civil Engineers, Vol. 64, 1976, p. 31-48.

(5) B.S. Stratford Smith and R. Gaiotti, Effect of Precast Concrete Cladding of the Behaviour of Tall
Buildings, Tall Buildings: 2000 and Beyond. Fourth World Congress, Hong Kong, November 1990,
p. 317-332.

170
(3) Steel Construction Today, The Steel Construction Institute (SCI), Advisory Desk, Vol. 5, No. 4, p.
203-206, July 1991.

(2) M. Strand and M. Pirner, 'Static and dynamic full-scale tests on a portal structure', The Structural
Engineer, September 1978, Vol. 56b, No. 3, p. 45-52.

A. Tallin and B. Ellingwood, 'Serviceability limit states: Wind induced vibrations', Journal of
Structural Engineering, ASCE. Vol. 110 No. 10, October 1984, p. 2424-2438.

(5) J.G. Tawressey, Factors of safety for thin stone in building curtain walls, Fifth North American
Masonry Conference, June 1990, p. 1417-1424.

(6) F.G. Thomas, 'The strenght of brickwork', Structural Engineer, 1953, 31 (Feb.), p. 35-46.

(2. 5) RX. Tomasetti, A. Gutman, IPXew and LJVl. Joseph, Development of thin wall cladding to reduce
drift in high-rise buildings, IABSE colloquium on thin-walled metal structures. Stokholm, 1986, p.
239-246.

(2) F. Tschemmeniegg, On the limitation of deflection of sway-frames, Seminar/Workshop on service

ability limit states for steel buildings, IABSE Working commission , Zurich, April 1, 1989, p. 16-
17.

A.I. Tseitlin, E.A. Neustroev and M.O. Kobakhidze, The dynamic stiffness of interstorey floors (CIB-
W85), 1988.

RX. Tuomi and W.J. McCutcheon, Testing of a Full-Scale House under Simulated Snow loads and
Wind Loads, Research Paper FPL 234, U.S. Department of Agriculture, Forest Products Laboratory,
Madison Wise., 1974.

(5) RX. Tuomi and W.J. McCutcheon, 'Racking Strength of Light-Frame Nailed Walls', ASCE, Journal
of the Structural Division, Vol. 104, No. ST7, July 1978, p. 1131-1140.

(2) N. Uchida, M. Kawamura, T. Aoyagi and H. Kirihara, Vibration offloor beams due to walking occu
pants. Study on measured data, Seminar/Workshop on serviceability limit states for steel buildings,
IABSE Working commission , Zurich, April 1, 1989, p. 31-52.

Vervormingseisen voor bouwconstructies, Steelbuilding Association (SG), Rotterdami and the

Concrete Association (CUR), Zoetermeer SG-CUR Report No.2, Delft, 1975.

(2) E. Wilson and A. Habibullah, SAP90 Finite Element Analysis Progra. Computers and Structures,
In., Berkeley, California, 1989.

(2) R.H. Wood, 'Efffective lengths of columns in multi-storey buildings. Parts 1, 2 and 3', The
Structrual Engineer, Vol.52, No. 9, September 1974, p. 235-244, 295-302, 341-346.

(2) RJL Wood and E.H. Roberts, 'A graphical method of predicting sideways in the design of multistorey
buildings', Proceedings of the Institution of Civil Engineers, June 1975, Vol. 59, Part 2, p. 353-372.

(2. 6) R.H. Wood, 'Plasticity. Composite action and collapse design of unreinforced shear wall panels in
frames', Proceedings of the Institute of Civil Engineers, June 1978, Part 2 No. 65, p. 381-411.

(2) SX. Wood, R, Stark and S.A. Green, 'Collapse of Eight-Storey RC Building During 1985 Chile
Earthquake', Journal of the Structural Division, ASCE, February 1991, Vol. 117, No. 2, p. 600-619.

(2) S.T. Woolcock and S. Kitipornchai, 'Deflection limits for portal fimes. Steel Construction', Journal
of the Australian Institute of Steel Construction, Vol. 20, November 1986, No.3, p.1-12.

171
 S.T. Woolcock and S. Kitipornchai, 'Survey of Deflection Limits for Portal Frames in Australia',
Journal of Constructional Steel Research, 1987.

T.A. Wyatt, Design guide on the Vibration of Floors, SQ publication 076, The Steel Construction
Institute, Ascot, 1989.

CW. Yancey and NF. Somes, Structural Tests of a Wood Framed Housing Module, report NBSIR
73-121, National Bureau of Standards, Washington D.C., 1973.

F.Y. Yokel, G. Hsi and NJ\ Somes, Full Scale Test on a Two-Story House Subjected to Lateral
Load (Building Science Series 44), National Bureau of Standards, Washington D.C, March 1973.

(5) F.Y. Yokel and S.G. FattaL 'Failure Hypothesis for Masonry Shear Walls', Journal of the Structural
Division, ASCE. Vol. 102, No. St3, March 1976, p. 515-532.

(2) M. Yoshida and K. Sato, Predicted and observed vibration of high-rise buildings and related human
sensitivity during typhoon, Seminar/Workshop on serviceability limit states for steel buildings,
IABSE Working commission , Zurich, April 1, 1989, p.26-27.

(6) R. Zandonini and P. Zanon, Beam design in PR Breced Frames, American Institute of Steel
Construction, Third Quarter, 1991, p. 26-1/26-26.

C Zeiger, Proposal for masonry design in common action with concrete and reinforcement,
Proceedings of the 7th international brick masonry conference, Melbourne Australia, February 1985,
p. 1145-1152.

172
 o
CORDIS
The Community Research and Development Information Service

Your European R&D Information Source

CORDIS represents a central source of information crucial for any organisation - be it industry, small and
medium-sized enterprises, research organisations or universities - wishing to participate in the
exploitation of research results, participate in EU funded science and technology programmes and/or seek
partnerships.
CORDIS makes information available to the public through a collection of databases. The databases cover
research programmes and projects from their preparatory stages through to their execution and final
publication of results. A daily news service provides up-to-date information on EU research activities
including calls for proposals, events, publications and tenders as well as progress and results of research
and development programmes. A partner search facility allows users to register their own details on the
database as well as search for potential partners. Other databases cover Commission documents, contact
infoimation and relevant publications as well as acronyms and abbreviations.

By becoming a user of CORDIS you have the possibility to:

Identify opportunities to manufacture and market new products
Identify partnerships for research and development
Identify major players in research projects
Review research completed and in progress in areas of your interest

The databases - nine in total - are accessible on-line free of charge. As a user-friendly aid for on-line
searching, Watch-CORDIS, a Windows-based interface, is available on request. The databases are also
available on a CD-ROM. The current databases are:

News (English, German and French version) - Results -

Partners - Projects - Programmes - Publications -
Acronyms - Comdocuments - Contacts

CORDIS on World Wide Web

The CORDIS service was extended in September 1994 to include the CORDIS World Wide Web (WWW)
server on hitemet. This service provides information on CORDIS and the CORDIS databases, various
software products, which can be downloaded (including the above mentioned Watch-CORDIS) and the
possibility of downloading full text documents including the work programmes and information packages
for all the research programmes in the Fourth Framework and calls for proposals.
The CORDIS WWW service can be accessed on the Internet using browser software (e.g. Netscape) and
the address is: htip://www.cordis.lu/
The CORDIS News database can be accessed through the WWW.

Contact details for further Information

If you would like further information on the CORDIS services, publications and products, please contact
the CORDIS Help Desk:

CORDIS Customer Service Telephone: +352-401162-240

B.P. 2373 Fax: +352-401162-248
L-1023 Luxembourg E-mail: helpdesk@cordis.lu
WWW: http://www.cordis.lu/
European Commission

EUR 15819 Properties and service performance

Serviceability deflections and displacements
in steel-framed structures

C. Bijl, F. Bijaard, ft Zandonini, D. Nethercot

Luxembourg: Office for Officiai Publications of the European Communities

1 997 1 72 pp. 21 .0 29.7 cm

Technical steel research series

ISBN 92-828-0167-5

Price (excluding VAT) in Luxembourg: ECU 30

This report contains state-of-the art reviews of present serviceability

design methods for steel-framed buildings, details and results of full-scale
dynamic testing of steel-framed structures and presents investigations of
new serviceability modelling methods. State-of-the art reviews were
performed to investigate the present status of serviceable design as
reflected in code requirements around the world. This involved an
investigation of serviceability load levels and combinations, deformation
types and limits. Previous general studies in the area of serviceability were
reviewed, uterature reviews were also made of th in-plane stiffness and
limits to serviceability of non-structural components (cladding, partitions
and connections between cladding, partitions and the structural steel
frame).
Two full-scale steel-framed buildings were tested. The objective of these
tests was to use dynamic techniques to observe the real in-service lateral
stiffness of steel-framed structures. Two testing techniques were used:
counterbalanced rotating masses and impulse loading using a hammer
blow. Both investigations concluded that structural response at service
load levels is dominated by the behaviour of non-structural components.
 Venta Salg Verkauf Sales Vente Vendita Verkoop Venda Myynti Frsljning
BELGIQUE/BELGIE GRAND-DUCHE DE LUXEMBOURG SCHWEIZ-SUISSE/SVIZZERA ISRAEL

Moniteur beige/Belgisch Staatsblad Messageries du livre Sari OSEC R.O.Y. International

Rue de Louvain 40-42/ 5. rue Raiffeisen StampfenbachstraBe 85 17, Shimon Hatarssi Street
Leuvenseweg 40-42 L-2411 Luxembourg CH-8035 Zrich PO Box 13056
B-1 000 Bruxelles/Brussel Tl. (352) 40 10 20 Tel. (41-1)365 53 15 61 130 Tel Aviv
Tl. (32-2) 552 22 11 Fax (352) 490 661 Fax (41-1) 365 54 11 Tel. (972-3) 546 14 23
Fax(32-2)5110184 E-mail: mdl@pt.lu E-mail: urs.leimbacher@ecs.osec.ineLch Fax (972-3) 546 14 42
URL: www.osec.ch E-mail: royii@netvision.net.il
Jean De Lannoy Abonnements:
Sub-agent for the Palestinian Authority:
Avenue du Ro! 202/ CESKA REPUBLIKA
Koningslaan 202
Messageries Paul Kraus Index Information Services
MS CR prodejna
B-1060 Bruxelles/Brussel 1 1 , rue Christophe Plantin PO Box 19502
Tl. (32-2) 538 51 69 L-2339 Luxembourg Konviktsk 5 Jerusalem
Fax (32-2) 538 08 41 Tl. (352) 499 88 88 CZ-113 57Praha1 Tel. (972-2) 27 16 34
E-mail: jean.de.lannoy@infoboard.be Fax (352) 499 888 444 Tel. (42-2) 24 22 94 33 Fax (972-2) 27 12 19
E-mail: mpk@pt.lu Fax (42-2) 24 22 94 33
Ubrakie europeen ne/Europeee Boekhandel URL: www.mpk.lu E-mail: nkposp@dec.nis.cz
URL: www.nis.cz RUSSIA
Rue de la Loi 244/
Wetstraat 244 NEDERLAND CCEC
B-1 040 Bruxelles/Brussel CYPRUS
60-letiya Oktyabrya Av. 9
Tl. (32-2) 295 26 39 SDU Servicecentrum Uitgevers 117312MOSCOW
Fax (32-2) 735 08 60 Cyprus Chamber Of Commerce & Industry
Tel.(095)135 52 27
Christoffel Planjnstrat 2 38, Grivas Digenis Ave Fax (095) 135 52 27
Postbus 20014 Mail orders:
DANMARK ' 2500 EA 's-Gravenhage PO Box 1455
Tei. (31-70)378 98 80 CY-1 509 Nicosia AUSTRALIA
J. H. Schultz Information A/S Fax (31-70) 378 97 83 Tel. (357-2) 44 95 00/46 23 12
E-mail: sdu@sdu.nl Fax 357-2) 361 044 Hunter Publications
Herstedvang 10-12 URL: www.sdu.nl. E-mail: cy1691_eic_cypros@vans.infonet.com PO Box 404
DK-2620 Albertslund 3167 Abbotsford, Victoria
Tlf . (45) 43 63 23 00 Tel. (61-3)941753 61
Fax (45) 43 63 19 69 OSTERREICH MAGYARORSZG
Fax (61 -3) 941 9 71 54
E-malI: schultz@schultz.dk
URL: www.schultz.dk Euro Info Service
Manz'sche Verlags- und Unlversltlts CANADA
buchhandlung Gmbh Europa Hz
Margitsziget
DEUTSCHLAND Siebenbrunnengasse 21 PO Box 475 Uniquement abonnements/
Postfach 1 H-1396 Budapest 62 Subscriptions only:
Bundesanzeiger Verlag A-1 050 Wien Tel. (36-1)11 16 061/11 16 216
Tel. (43-1)53 161334/340 Fax (36-1) 302 50 35 Renouf Publishing Co. Ltd
Breite StraUe 78-80 Fax (43-1) 53 161 339 E-mail: euroinfo@maiI.matav.hu 1 294 Algoma Road
Postfach 10 05 34 E-mail: ausliefeaing@manz.co.at URL: www.euroinfo.hu/index.htm K1 B 3W8 Ottawa, Ontario
D-50667 Kln URL: www.austria.EU.net:81/manz Tel. (1-613)74173 33
Tel. (49-221) 20 29-0 Fax (1-613) 741 54 39
Fax (49-221) 20 29 278 MALTA
E-mail: renouf@fox.nstn.ca
PORTUGAL URL: fox.NSTN.Ca/~renouf
Miller Distributors Ltd
GREECE/ Imprensa NaclonahCasa da Moeda, EP Malta International Airport
PO Box 25
G.C. Eleftheroudakls SA Rua Marqus de S da Bandeira, 1 6 A LQA 05 Malta EGYPT
P-1050 Lisboa Codex Tel.(356)66 44 88
International Bookstore Tel. (351-1) 353 03 99 The Middle East Observer
Panepistimiou 17. Fax (356) 67 67 99
Fax (351-1)353 02 94/384 01 32 41, Sherif Street
GR-105 64 Athens
Tel. (30-1)3314180/1/2/3 POLSKA Cairo
Distribuidora de Uvros Bertrand Ld.' Tel. (20-2) 39 39 732
Fax (30-1) 323 98 21
E-mail: elebooks@netor.gr Rua das Terras dos Vales, 4 A Ars Polona Fax (20-2) 39 39 732
Apartado 60037 Krakowskie Przedmiescie 7
P-2701 Amadora Codex JAPAN
Skr. pocztowa 1001
ESPAA Tel. (351 -1 ) 495 90 50/495 87 87 PL-00-950 Warszawa
Fax (351-1 496 02 55 Tel. (48-2)2612 01 PSWapan
Mundi Prensa Libros, SA Fax (48-2) 26 62 40 Asahi Sanbancho Plaza #206
Castell, 37 SUOMI/FINLAND 7-1 Sanbancho, Chiyoda-ku
E-28001 Madrid TURKIYE Tokyo 102
Tei. (34-1) 431 33 99/431 32 22 Tel. (81-3)3234 69 21
Akateeminen Kirjakauppa / Fax (81-3) 3234 69 15
Fax (34-1) 575 39 98 DQnya Infotel A.S.
E-mail: mundiprensa@tsai.es Akademiska Bokhandeln E-mail: psijapan@gol.com
Pohjoisesplanadi 39/ Istikll Cad. No: 469 URL: www.psi-japan.com
URL: www.tsai.es/mprensa TR-80050 Tnel-Istanbul
Norra esplanaden 39
PL/PB128 Tei. (90-212)2519196
Boletn Oficial del Estado (90-312 427 0210 SOUTH AFRICA
FIN-00101 Helsinki/Helsingfors
Trafalgar, 27-29 Tel. (358-9) 121 41 Fax 90-212 251 91 97
Safto
E-28071 Madrid Fax (358-9) 121 44 35
Tel. (34-1 ) 538 22 95 (Libros)/ E-mail: akatiIaus@stockmann.mailnet.fi BALGARUA 5th Floor Export House,
384 1715 (Suscripciones) URL: booknetcultnet.fi/aka/index.htm CNR Maude & West Streets
Fax (34-1) 538 23 49 (Libros)/ Europress-Euromedla Ltd PO Box 782 706
384 17 14 (Suscripciones) 2146Sandton
59, Bid Vitosha Tel. (27-1 1)883 37 37
URL: www.boe.es SVERIGE BG-1000 Sofia Fax (27-1 1)883 65 69
Tel. (359-2) 80 46 41
Mundi Prensa Barcelona BTJAB Fax (359-2) 80 45 41
Conseil de Cent, 391 Traktorvgen 1 1 UNITED STATES OF AMERICA
E-08009 Barcelona PO Box 200 HRVATSKA
Tei. (34-3) 488 34 92 Bernan Associates
S-22100 Lund
Fax 34-3 487 76 59 Tel. (46-46) 18 00 00 Mediatrade Ltd 461 1 -F Assembly Drive
Fax (46-46) 18 01 25 MD20706 Lanham
Pavia Hatzal Tel. (301) 459 2255 (toll free telephone)
E-mail: bUo@mail.b|.se HR-1 0000 Zagreb
FRANCE URL: www.btj.se/media'eu Fax (800) 865 3450 (toll free fax)
Tel. (385-1)43 03 92 E-mail: query@beman.com
Journal officiel Fax (385-1) 44 40 59 URL: www.beman.com
UNITED KINGDOM
Service des publications des CE
26, rue Desaix
ROMANIA MEXICO
F-75727 Paris Cedex 15 The Stationery Office Ltd
(Agency Section) Euromedia Mundl-Prensa Mexico, SA de CV
Tl. (33-1)40 58 77 01/31
Fax (33-1) 40 58 77 00 51 , Nine Elms Lane Str. G-ral Berthelot Nr 41 Rio Panuco, 141
London SW8 5DR RO-70749 Bucuresti Delegacin Cuauhtemoc
Tel. (44-171) 873 9090 Tel. (40-1) 210 44 01/614 06 64 ME-06500 Mxico DF
RELAND Fax (44-171) 873 8463 Fax (40-1) 210 44 0/312 96 46 Tel. (52-5) 553 56 58/60
URL: www.the-stationery-office.co.uk Fax (52-5) 514 67 99
Government Supplies Agency SLOVAKIA E-mail: 104164.23compuserve.com
Publications Section ICELAND Slovenske Technicka Knlznlca
4-5 Harcourt Road RPUBLIQUE DE CORE
Dublin 2 Nmestie slobody 1 9
Tel. (353-1) 661 31 11 Bokabud Larusar Blondel SLO-81223 Bratislava 1 Kyowa Book Company
Fax (353-1) 475 27 60 Skiavrdustig, 2 Tel. (42-7) 53 18 364 1 F1 . Phyung Hwa Bldg
IS-101 Reykjavik Fax (42-7) 53 18 364 41 1-2 Hap Jeong Dong, Mapo Ku
Tel.(354)5515 650 E-mail: europ@tbb1 .sltk.stuba.sk 121-220 Seoul
ITALIA Fax (354) 55 25 560 Tl. (82-2) 322 6780/1
SLOVENIA Fax (82-2) 322 6782
Licosa SpA E-mail: kyowa2@ktnet.co.kr.
NORGE Gospodarskl Vestntk
Via Duca di Calabria, 1/1
Casella postale 552 Zalozniska skupina d.d. ANDERE LANDER/OTHER COUNTRIES/
NIC Info A/S Dunajska cesta 5
1-50125 Firenze AUTRES PAYS
Tel. (39-55)64 5415 0stenjoveien18 Sl-1 000 Ljubljana .
Fax (39-55) 64 12 57 Boks 651 2 Etterstad Tel. (386)61 133 03 54 Bitte wenden Sie sich an ein Bro Ihrer
E-mail: licosa@ftbcc.it N-0606 Oslo Fax (386) 61 133 91 28 Wahl / Please contact the sales office of
URL: t382.cIea.Wrtua!_Ubrary/biWioWetrina/ Tel. (47-22) 97 45 00 E-mail: belicd@gvestnik.s your choice / Veuillez vous adresser au
Ncosatf1.htm Fax (47-22) 97 45 45 URL: www.gvesmik.si bureau de vente de votre choix

12/96*
 NOTICE TO THE READER

All scientific and technical reports published by the European Commission are announced in the
periodical 'euro abstracts', published every two months, and in the R&TD publications database of
CORDIS, the Community Research and>Development Information Service. o
o
For subscription (1 year: ECU 65) please write to the address below.

OD

CO

m
2

Price (excluding VAT) in Luxembourg: ECU 30

ISBN =l_-fl_-D__7-5

* OFFICE FOR OFFICIAL PUBLICATIONS

*__&_.
OP *.
OF THE EUROPEAN COMMUNITIES
=,= -* 9 789282"801673 >
* * * L-2985 Luxembourg