Vous êtes sur la page 1sur 13

Journal of Constructional Steel Research 62 (2006) 1047–1059

www.elsevier.com/locate/jcsr

The European standard family and its basis


Gerhard Sedlacek ∗ , Christian Müller
RWTH Aachen, Institute of Steel Construction, Mies-van-der-Rohe Strasse 1, D-52074 Aachen, Germany

Abstract

The Eurocodes as European unified design rules for structures are part of the European Standard Family comprising also product standards,
testing standards, standards for execution, European Technical Approvals and European Technical Approval Guidelines.
A key feature of all these standards is consistency that has been obtained by consistent definitions of material and product properties and by
basing any calculative way of defining structural properties on test evaluations.
As a consequence all rules in Eurocode 3 are justified by test evaluations with a standardised method that introduced full transparency into the
harmonisation works and allowed new innovative design approaches.
Some examples for determining characteristic values of actions and combination factors for actions as well as for determining characteristic
values and design values of resistances, in particular for the rules for choice of material to avoid brittle fracture, the harmonisation of various types
of stability checks and the new interpretation of the plate buckling rules highlight the benefits of the standardised evaluation method.
c 2006 Elsevier Ltd. All rights reserved.

Keywords: Eurocodes; Steel; Unified design rules; Examples

1. Introduction

The Eurocodes have been developed since 1979 in several


steps. The start for steel was the development of Eurocode
3 under a contract between the Commission and the ECCS.
The Commission later mandated CEN to continue the work
to prepare ENV-Eurocodes that after an inquiry have been
transferred to the final EN versions.
As 2005 is the year of the completion of technical works, see
Fig. 1, it is an appropriate time for remembrance of the basis of
the work agreed across different kinds of material and ways of Fig. 1. 2005—year of completion of 10 Eurocodes (58 parts).
construction in interdisciplinary groups, where Prof. Patrick J.
Dowling, first chairman of the groups for preparing Eurocode requires International Code Families in order to avoid
3, played a key role. inconsistencies due to the use of various national codes.
So far there are two sources of International Code Families:
one in the USA, the other in Europe, each consisting of a design
2. Globalisation and international standard families code in connection with product standards and testing codes,
see Fig. 2.
The globalisation of the construction market comprising The European code family prepared by CEN so far includes
construction products, engineering and construction services 10 Eurocodes with 58 parts with design rules and many
hundreds of EN standards for products and testing. It also
contains so far ∼170 European Technical Approvals and
∗ Corresponding author. Tel.: +49 241 80 5177; fax: +49 241 888 8140. European Approval Guidelines worked out by the European
E-mail address: stb@stb.rwth-aachen.de (G. Sedlacek). Organisation for Technical Approvals (EOTA).

c 2006 Elsevier Ltd. All rights reserved.


0143-974X/$ - see front matter
doi:10.1016/j.jcsr.2006.06.027
1048 G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059

Fig. 2. Overview of international code families.

Fig. 3. Dissemination of international standard families.

The European Standard Family is technically coordinated


and constitutes the most advanced standard system in the world.
Up to 2010 it will be implemented in countries of the EU
and EFTA. It may also be chosen by other countries that wish
to participate in the European market and European technical
developments Fig. 3.
3. Basis of the European standard family
The Eurocodes and the European product and testing
standards as well as ETAs and ETAGs are tools to fulfil
the “Essential requirements” of the European “Construction
Product Directive (CPD)” with sufficient reliability, in
particular the requirements “Mechanical resistance and Fig. 4. Essential requirements and tools for fulfilment.
stability” and “Resistance to fire” Fig. 4.
The conditions for the application and use of the Eurocodes The crucial condition in Guidance Paper L for the
and the product and testing standards are laid down in Guidance architecture of the design rules in Eurocode 3 and all
Paper L [1] agreed by the Commission and Member States. other Eurocodes is that the manufacturer may determine the
G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059 1049

Fig. 5. Testing of prefabricated components.

Fig. 6. Reliability basis.

properties of prefabricated components to be declared for CE 3. Eurocodes have a double role; beside their role as a tool for
marking either by tests or by calculations and that for the determining Rk they shall also be suitable for the design
calculative determination of properties the Eurocodes are the of structures. That design needs design values Rd that
only design codes referred to, see Fig. 5. shall be determined using the declared characteristic values
By this condition a link between experimental results from Rk . Hence the design values Rd needed for the design of
tests with prefabricated components and the design rules in structures shall be
the Eurocodes is established, that is specified by the reliability Rk
requirements in EN 1990 – Eurocode: Basis of Structural Rd =
γM
Design [2] – in the following way:
1. The product property to be declared, that may be determined where γ M is a global factor related to the resistance Rk . It is
directly from testing, shall represent a certain fractile of therefore not possible to use separate partial factors γ Mi to
the statistical distribution of the experimental results. It parameters X i in the formula for R(X i ), e.g. partial factors
is denoted as characteristic value Rk (in general the 5%- for stiffness, slenderness or the strengths of constitutive
fractile) and this value declared with CE marking will be materials.
acknowledged throughout Europe without any impact from 4. The choice of the global partial factors γ M is the respon-
national safety levels. The method to determine Rk from sibility of Member States (Nationally Determined Parame-
tests is therefore a unified European rule in EN 1990 — ters); however the Eurocodes provide recommendations for
Annex D. the numerical values for these NDPs that result from the
2. Eurocodes shall, as an alternative to experimental testing, same test evaluations that are used to verify Rk . If national
provide by their design rules calculation-based methods for choices are different to these recommendations, these differ-
determining numerical values of Rk , that are in competition ences should be justifiable.
with those from direct experimental tests. Therefore the
characteristic values Rk in the Eurocodes must be calibrated The basic reliability targets for design values for ULS
to test results such that the manufacturer prefers them to any recommended in EN 1990 are based on a semi-probabilistic
experimental determination. approach, see Fig. 6, with the reliability index β = 3.80 for a
1050 G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059

Fig. 7. Guidance Paper L and model for traffic actions.

Fig. 9. Snow load in Munich–Riem.

Fig. 8. Definition of characteristic values of actions and action effects.

reference period of 50 years and weighting factors α E = −0.7


for action effects and α R = 0.8 for resistances.
This schematic diagram applies to the simple case where just
one action effect is relevant. More elaborate models have been
developed for combinations of actions, as will be highlighted
below.
Before more details on the test evaluation, in particular for
Eurocode 3, are presented, first the consequences of EN 1990
for the determination of characteristic values and design values
of actions are highlighted.
Fig. 10. Wind load in Munich–Riem.
4. Actions and action effects

The procedure for determining the load models in the An example for the use of these definitions is the preparation
Eurocodes [3] can be best described with the load model for of the loading specifications for the Allianz–Arena in Munich
traffic loads on road bridges, see Fig. 7. for the football world championship in 2006. Fig. 9 shows
From traffic measurements (axle loads and distances the statistical distribution of the annual extremes of the
between axles) in the Paris–Lyon highway at Auxerre (that have snowloads at the location of the stadium and the subsequent
been agreed to be adopted as representing “European traffic”) characteristic value for snow on the ground defined by a return
traffic effects E(Q) on typical bridges were calculated with period of 50 years or the 0.98-fractile of the annual extreme
dynamic simulation models. value distribution. Fig. 10 illustrates the determination of the
From statistical evaluations, functions of the characteristic characteristic peak velocity pressure according to EN 1991, Part
values E k (Q) were determined that were used to calibrate a 1-4 for wind, and Fig. 11 gives the characteristic values of air
fictitious engineering load model Q k composed of a suitable temperature related to a reference temperature of +10 ◦ C.
loading pattern and the magnitudes of its components. In In Fig. 12 all characteristic values and design values
conclusion action models are all action-effect-oriented. determined from measurements of magnitudes of actions are
For various actions the definitions of characteristic values assembled.
are given in Fig. 8. Fig. 8 also shows the definition of For the determination of a combination factor the
combination factors from characteristic values of combined consideration of single actions is not sufficient. Fig. 13
action effects. shows how the action effects from snow and wind may be
G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059 1051

Fig. 11. Air temperature in Munich–Riem.

Fig. 12. Evaluated climatic actions.

Fig. 13. Combination rule of climatic actions.

Fig. 14. Characteristic values of effects of combined actions.


1052 G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059

Fig. 15. Combination factor ψ0 .

determined for various locations at a structure, Fig. 14 gives


the characteristic values of the action effects for these locations
depending on the weighing parameter and Fig. 15 eventually
shows the maximum value of the combination factor applicable
to the concept of a leading and an accompanying action. Fig. 16
shows the Allianz–Arena after completion.

5. Calibration of steel structures design rules to tests

The central role of the test evaluation for the development of


sustainable design rules with sufficient stability and continuity
for steel structures [4] (see Fig. 17) is demonstrated in Fig. 18:
1. Prefabricated steel components to be tested experimentally
shall have properties representative for a larger population
and comply with the requirements of the Product Standards Fig. 16. Allianz–Arena Munich.

Fig. 17. Standard system for steel structures.


G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059 1053

Fig. 18. Determination of characteristic values Rk and γ M values from tests.

Fig. 19. Use of test evaluation method for various regulatory routes.

Fig. 20. Procedure to obtain reliable values Rk .


1054 G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059

Fig. 21. Choice of material.

Fig. 22. Choice of material to EN 1993-1-10.

for materials and semi-finished products and with the


execution rules in EN 1090—Part 2. These representative
properties make them suitable to determine representative
resistance values Rexp,i .
2. For interpreting the test results an engineering model R(X )
is applied that allows us to determine by calculation the
resistance properties Rcalc,i of the test components using the
measured parameters X i .  
Rexp
3. The plotting of the ratios Rcalc versus X i demonstrates
i
the quality of the engineering model (the ratios should be
independent of X i variations, i.e. horizontal lines).
4. By direct comparison of Rexp,i and Rcalc,i the mean value
correction Rm and the scatter distribution sδ can be found.
These values allow estimation of the initial value Rk = R5% Fig. 23. Choice of material to EN 1993-1-10.
G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059 1055

Fig. 24. Common design rules.

Fig. 25. Test evaluation for buckling curves and γ M values.


1056 G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059

Fig. 26. Mechanical background of column and lateral torsional buckling.

and the design value Rd , for which the recommended value the transfer of research results to practical applications, see
of the reliability index β = 3.8 and the weighting factor Fig. 19.
α R = 0.8 for the resistance are used. This method has been used to justify the characteristic values
5. From Rk and Rd the γ M -value for the particular problem can of strength and also the numerical values of the partial factors
be obtained, or as suggested in Eurocode 3, a suitable class γ M recommended in Eurocode 3—EN 1993 [4]. Fig. 20 gives a
of γ M is chosen from the following possibilities: survey on the various recommended γ M values associated with
γ M0 = 1.00 where large deflections due to yielding ( f y ) the different ductile failure modes distinguished in Eurocode
define the ULS 3—EN 1993. It also shows that test evaluations were performed
γ M1 = 1.10 where component failure due to instability to determine the design strength functions Rd depending on
occurs (λ) the yield strength f y or the tensile strength f u according to
γ M2 = 1.25 where failure is caused by disintegration of the relevant failure mode in the first instance, whereas the
material ( f u ). characteristic values Rk were obtained from Rd by multiplying
6. The initial characteristic value is then corrected to comply with the recommended γ M -values.
with the partial factor γ Mi chosen as The method has also been used to adjust the toughness oriented
Rk = γ Mi · Rd . safety checks for brittle failure in the low temperature domain
to target reliability in order to prepare the rules for the choice
7. Finally, the statistical parameters obtained from this test
of material to avoid brittle fracture in EN 1993—Part 10. This
evaluation allow the determination of quality requirements
choice is the prerequisite to base the design on ductile failure
for the product standards and execution standards to comply
modes only.
with the γ M values. All test evaluations have demonstrated that the model
With this evaluation method, which is more detailed in EN uncertainty sδ of any engineering model R is the main
1990 and Eurocode 3, a transparent unified European basis for controlling parameter for γ M , so that the format Rd = γRMk
the equal treatment of research results, unique verifications, used in Eurocode 3 is also justified from the statistical point of
technical approvals and design codes is available that facilitates view.
G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059 1057

at the most unfavourable location of the structure. The safety


check is performed using stress intensity factors K and reads:
K appl,d ≤ K mat,d
where K appl,d is the toughness requirement and K mat is the
toughness resistance.
Fig. 22 shows the table for permissible product thicknesses
calculated for different steel grades, temperatures and stress
levels with this model. Fig. 23 shows a typical application for
the cast steel nodes for the grandstand roof of the Olympic
stadium in Berlin.

7. Harmonisation of stability rules

A field of traditionally complex design rules is the field


of stability verifications, namely for flexural buckling, lateral
Fig. 27. Comparison of buckling curves and LTB curves. torsional buckling, plate buckling and shell buckling.
To demonstrate the efficiency of the test evaluation method
6. Choice of material to avoid brittle fracture a simplified unique approach for the verification of these
stability phenomena is used that takes the availability of FEM
The choice of material to avoid brittle fracture is based on a
programmes into account.
fracture mechanics model as illustrated in Fig. 21. For a given
Fig. 24 gives the unified approach for the verification of
detail the accidental existence of an initial crack with the size
flexural buckling, lateral torsional buckling, plate buckling
a0 is assumed, that normally should have been detected and
and shell buckling by using the “Global system slenderness”
repaired during welding inspections.
λ concept together with appropriate reduction curves χ (λ).
Under service conditions, initial cracks may grow due to
This concept is not limited to specific geometrical boundary
fatigue until they are detected by service inspections.
conditions or loading conditions [5].
It is assumed that in the time interval between two
The reduction curves χ (λ), namely column buckling curves,
inspections a fatigue load equivalent to the fatigue damage
lateral torsional buckling curves, plate buckling curves and shell
∆σi3 · n i buckling curves, are defined by technical classes with imperfec-
P
1
D= = tion parameters a0 , a, b, c, d. These parameters were calibrated
4 ∆σc3 · 2 · 106
to test results according to EN 1990—Annex D in such a way
is applied to the structure, so that due to fatigue the crack with that the required characteristic values Rk and the recommended
size a0 at the beginning of the interval develops to its design values for the partial factors γ M were obtained, see Fig. 25.
value ad at the end of the interval. The engineering models used for the buckling curves are
At the end of the interval an accidental design situation is based on mechanical models for structural elements with
applied for the structure with the minimum temperature TEd as imperfections, see Fig. 26. Fig. 27 shows the shape of the
leading action together with the “frequent load” effect σ Ed = buckling curves depending on the product used and the limit
σG + ψ1 · σ Q as accompanying action and the crack size ad state considered.

Fig. 28. Application of global slenderness concept for a bridge supporting frame.
1058 G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059

on stress limitations and the method using effective cross-


sections could be consistently interpreted in the course of test
evaluations.
Whereas the performance of the member before plate
buckling of its components in compression can be easily
described by the stress limitation method based on gross cross-
sections and linear elastic behaviour the formation of effective
cross-sectional properties presumes that after a certain amount
of non-linear deformation a strain ε up to the yield strain ε y can
be reached, see Fig. 29.
Fig. 30 shows the equivalence of the stress limit σlimit for the
gross area and of the limit by yielding for the effective area for
a single plate element.
It also shows how after the first attainment of the stress limit
σlimit for the weakest plate element the distribution of stress
limits σlimit over the full cross-section may be obtained, that
is fully equivalent to the distribution of effective areas related
to f y over the cross-section
Fig. 29. Modelling of plate buckling.
Hence both the stress limit concept and the effective cross-
section concept lead to the same results for resistance if in the
The possibilities that are offered by this new harmonised stress limit concept the distribution of different stress limits is
method are demonstrated by the stability check of a complete integrated as for hybrid sections or composite sections.
frame as given in Fig. 28 [6].
9. Conclusions
8. New interpretation of plate buckling rules

The two methods offered in EN 1993-1-5 Eurocode 3— The Eurocodes are part of the European Standard Family and
Part 1-5 for plate buckling verification, i.e. the method based will be completed by the end of 2005.

Fig. 30. Modelling of plate buckling.


G. Sedlacek, C. Müller / Journal of Constructional Steel Research 62 (2006) 1047–1059 1059

The Eurocodes have a double role. On the one hand they Prof. Patrick J. Dowling as former chairman of the
give rules to determine the characteristic values of product subcommittee of CEN/TC 250 for Eurocode 3 played a key rule
properties for CE marking by calculation instead of testing; in introducing this strategy for harmonising technical rules in
on the other hand they are technical reference documents for Europe.
design works in connection with National Annexes.
This double role requires that all design rules are based on References
test evaluations using an appropriate test evaluation method.
Such a test evaluation method initially developed for Eurocode [1] European Commission: Enterprise Directorate-General. Single Market:
3 (former Annex Z of ENV 1993) is now standardised in Regulatory Environment, Standardisation and New Approach. Construc-
tion. ENTR/G5: Guidance paper L (concerning the Construction Products
Annex D of EN 1990 – Eurocode Basis of Structural Design
Directive — 89/106/EEC) Application and use of eurocodes. Brussels. 27
– applicable to all kinds of materials and ways of construction. November 2003.
Various examples are given to show the benefits of the [2] European Committee for Standardization CEN: EN 1990–Eurocode–Basis
evaluation method both for the determination of characteristic of structural design. Brussels.
values of actions and for determining characteristic and design [3] European Committee for Standardization CEN: EN 1991–Actions on
values of resistances. structures. Brussels.
The evaluation method has led to a transparent system that [4] European Committee for Standardization CEN: EN 1993–Eurocode
3–Design of steel structures. Brussels.
enabled us to introduce new innovative approaches for design, [5] Müller C. Zum Nachweis ebener Tragwerke aus Stahl gegen seitliches
e.g. for the choice of material to avoid brittle fracture and Ausweichen, Dissertation. Heft 47, 2003.
harmonised general rules for stability checks including more [6] Sedlacek G, Müller C. Eurocodes et International Advantages des
consistent approaches for plate buckling. Eurocodes, Colloque Européen sur les Eurocodes. Paris 12/2004.

Vous aimerez peut-être aussi