Beruflich Dokumente
Kultur Dokumente
1991-1:1996
Eurocode 1:
Basis of design and
actions on structures —
Part 1: Basis of design —
ICS 91.040
DD ENV 1991-1:1996
The preparation of this National Application Document for use in the UK with
ENV 1991-1:1996 was entrusted by Technical Committee B/525, Building and
civil engineering structures, to Subcommittee B/525/1, Actions (loadings) and
basis of design, upon which the following bodies were represented:
Contents
Page
Committees responsible Inside front cover
National foreword ii
Text of National Application Document v
Foreword 2
Text of ENV 1991-1 7
© BSI 04-2000 i
DD ENV 1991-1:1996
National foreword
ii © BSI 04-2000
DD ENV 1991-1:1996
This NAD also provides clarification to certain clauses which were considered
ambiguous; these clarifications are not intended to change the original intention
of the drafters of ENV 1991-1.
The main reasons for publishing the UK NAD to Eurocode 1: Part 1 are as follows.
a) For essentially informative purposes, to enable structural designers in the
UK to familiarize themselves with the contents of ENV 1991-1, which, as
referred to above, has no existing equivalent British Standard.
b) Following from a), to enable UK comments on ENV Eurocode 1: Part 1 to be
obtained during its ENV period, so that these can be considered during the
conversion to an EN.
c) To provide information in cases where the “Basis of design” sections of
DD ENV 1992 to DD ENV 1997 require amplification, for example under
unusual circumstances not adequately covered by those Eurocodes.
d) To provide definitions for certain terms and symbols used in
DD ENV 1997-1.
e) To enable comparative designs to be performed which compare the approach
of ENV 1991-1 with those of DD ENV Eurocodes predating ENV 1991-1.
Compliance with DD ENV 1991-1:1996 does not of itself confer immunity
from legal obligations.
For consideration of the conversion of ENV 1991-1 into a full European Standard,
it is important to get as much feedback as possible from practising engineers.
Such feedback is therefore strongly encouraged, and users of this document are
invited to comment on its technical content, ease of use and any ambiguities or
anomalies. These comments will be taken into account when preparing the UK
national response to CEN on the question of whether the ENV can be converted
into an EN.
Comments should be made in writing to the Secretary of Subcommittee B/525/1,
BSI, 389 Chiswick High Road, London W4 4AL, quoting this document, the
reference to the relevant clause and, if possible, a proposed revision.
Summary of pages
This document comprises a front cover, an inside front cover, pages i to x,
the ENV title page, pages 2 to 52 and a back cover.
This standard has been updated (see copyright date) and may have had
amendments incorporated. This will be indicated in the amendment table on the
inside front cover.
National Application
Document
© BSI 04-2000 v
DD ENV 1991-1:1996
Contents of
National Application Document
Page
Introduction vii
1 Scope vii
2 Informative references vii
3 Partial load factors, combination factors and other values vii
4 Loading codes vii
5 Reference standards vii
6 Additional recommendations vii
Table 1 — Table and equation substitutions vii
Table 2.1 — Notional classification of design working life viii
List of references ix
vi © BSI 04-2000
DD ENV 1991-1:1996
6.3 Clause 9. Verification by the partial factor d) 9.4.3 The second sentence of (2)P should be
method replaced by the following:
a) 9.3.2 A new note should be added to (2) as NOTE Examples of where this may apply are as follows.
follows: i) When considering Case A of Table 2 for the static
equilibrium of balanced cantilevers.
NOTE Information on appropriate values of *Sd for the ii) When considering Case B of Table 2 for the bending
analysis of bridges is given in BS 5400-1.”. strength needed within a span of a multispan beam which
b) (3)b) The phrase: “the partial factor is applied” has adjacent span lengths that differ greatly.”
should be replaced by the phrase: “the partial 6.4 Annex A to Annex D (informative)
factor *F is applied.”. Annex A to Annex D need further development work
c) 9.4.2 In (3)a) equations (9.10a) and (9.10b) before they can be considered adequately validated
should not be used, pending calibration work. for design purposes.
In (6) a note should be added at the end of the 6.5 Table 2.1
paragraph:
Table 2.1 should be replaced by the one listed below.
“NOTE For example, in the design of a section subject to
both bending moment and axial force due to a single action,
the axial force may be reduced by 20 % if it is a favourable
action effect.”.
Table 2.1 — Notional classification of design working life
Notional design
Class Examples
working life (years)
Informative reference
© BSI 04-2000 ix
x blank
EUROPEAN PRESTANDARD ENV 1991-1
PRÉNORME EUROPÉENNE
September 1994
EUROPÄISCHE VORNORM
ICS 91.040.00
Descriptors: Buildings, civil engineering, structures, building codes, design, safety, reliability, mechanical strength, verification
English version
CEN
European Committee for Standardization
Comité Européen de Normalisation
Europäisches Komitee für Normung
Central Secretariat: rue de Stassart 36, B-1050 Brussels
2 © BSI 04-2000
ENV 1991-1:1994
© BSI 04-2000 3
ENV 1991-1:1994
4 © BSI 04-2000
ENV 1991-1:1994
Page
Annex A (informative) Partial factor design 37
Annex B (informative) Fatigue 43
Annex C (informative) Serviceability
limit state: verification of structures
susceptible to vibrations 44
Annex D (informative) Design assisted
by testing 46
Figure A.1 — Overview of reliability methods 38
Figure A.2 — Design point definition
according to first order reliability
methods (FORM) 40
Table 2.1 — Design working life classification 16
Table 9.1 — Design values of actions for
use in the combination of actions 31
Table 9.2 — Partial factors: ultimate limit
states for buildings 33
Table 9.3 — ? factors for buildings 34
Table 9.4 — Design values of actions for
use in the combination of actions 35
Table A.1 — Relation between " and P1 38
Table A.2 — Indicative values for the
target reliability index ". 39
Table A.3 — Design values for various
distribution functions 40
Table A.4 — Expression for ?o 42
Table D.1 — Values of kn for the 5 %
characteristic value 49
Table D.2 — Values of kn for the ULS
design value, if X is dominating
(P{X < Xd} = 0,1 %) 50
Table D.3 — Values of kn for the ULS
design value, if X is non-dominating
(P{X < Xd} = 10 %) 50
© BSI 04-2000 5
6 blank
ENV 1991-1:1994
Section 1. General
1.1 Scope
(1) This Part 1 of ENV 1991 establishes the principles and requirements for safety and serviceability of
structures, describes the basis for design and verification and gives guidelines for related aspects of
structural reliability.
(2)P Part 1 of ENV 1991 provides the basis and general principles for the structural design of buildings and
civil engineering works including geotechnical aspects and shall be used in conjunction with the other parts
of ENV 1991 and ENVs 1992 to 1999. Part 1 relates to all circumstances in which a structure is required
to give adequate performance, including fire and seismic events.
(3) Part 1 of ENV 1991 may also be used as a basis for the design of structures not covered in
ENVs 1992 to 1999 and where other materials or other actions outside the scope of ENV 1991 are involved.
(4)P Part 1 of ENV 1991 is also applicable to structural design for the execution stage and structural design
for temporary structures, provided that appropriate adjustments outside the scope of ENV 1991 are made.
(5) Part 1 of ENV 1991 also gives some simplified methods of verification which are applicable to buildings
and other common construction works.
(6) Design procedures and data relevant to the design of bridges and other construction works which are
not completely covered in this Part may be obtained from other Parts of Eurocode 1 and other relevant
Eurocodes.
(7) Part 1 of ENV 1991 is not directly intended for the structural appraisal of existing construction in
developing the design of repairs and alterations or assessing changes of use but may be so used where
applicable.
(8) Part 1 of ENV 1991 does not completely cover the design of special construction works which require
unusual reliability considerations, such as nuclear structures, for which specific design procedures should
be used.
(9) Part 1 of ENV 1991 does not completely cover the design of structures where deformations modify direct
actions.
© BSI 04-2000 7
ENV 1991-1:1994
1.3 Assumptions
The following assumptions apply:
— The choice of the structural system and the design of a structure is made by appropriately qualified
and experienced personnel.
— Execution is carried out by personnel having the appropriate skill and experience.
— Adequate supervision and quality control is provided during execution of the work, i.e. in design
offices, factories, plants, and on site.
— The construction materials and products are used as specified in this Eurocode or in
ENVs 1992 to 1999 or in the relevant supporting material or product specifications.
— The structure will be adequately maintained.
— The structure will be used in accordance with the design assumptions.
— Design procedures are valid only when the requirements for the materials, execution and
workmanship given in ENVs 1992 to 1996 and 1999 are also complied with.
1.5 Definitions
For the purposes of this prestandard, the following definitions apply.
NOTE Most definitions are reproduced from ISO 8930:1987.
1.5.1 Common terms used in the Structural Eurocodes (ENVs 1991 to 1999)
1.5.1.1
construction works
everything that is constructed or results from construction operations
NOTE This definition accords with ISO 6707-1. The term covers both building and civil engineering works. It refers to the complete
construction worlds comprising structural, non-structural and geotechnical elements.
1.5.1.2
type of building or civil engineering works
type of construction works designating its intended purpose, e.g. dwelling house, retaining wall, industrial
building, road bridge
1.5.1.3
type of construction
indication of principal structural material, e.g. reinforced concrete construction, steel construction, timber
construction, masonry construction, composite steel and concrete construction
1.5.1.4
method of construction
manner in which the execution will be carried out, e.g. cast in place, prefabricated, cantilevered
8 © BSI 04-2000
ENV 1991-1:1994
1.5.1.5
construction material
material used in construction work, e.g. concrete, steel, timber, masonry
1.5.1.6
structure
organized combination of connected parts designed to provide some measure of rigidity
NOTE ISO 6707-1 gives the same definition but adds “or a construction works having such an arrangement”. In the Structural
Eurocodes this addition is not used in order to facilitate unambiguous translation.
1.5.1.7
form of structure
the arrangement of structural elements, such as beam, column, arch, foundation piles
NOTE Forms of structure are, for example, frames, suspension bridges.
1.5.1.8
structural system
the load-bearing elements of a building or civil engineering works and the way in which these elements
function together
1.5.1.9
structural model
the idealization of the structural system used for the purposes of analysis and design
1.5.1.10
execution
the activity of creating a building or civil engineering works
NOTE The term covers work on site; it may also signify the fabrication of components off site and their subsequent erection on site.
1.5.2 Special terms relating to design in general
1.5.2.1
design criteria
the quantitative formulations which describe for each limit state the conditions to be fulfilled
1.5.2.2
design situations
those sets of physical conditions representing a certain time interval for which the design will demonstrate
that relevant limit states are not exceeded
1.5.2.3
transient design situation
design situation which is relevant during a period much shorter than the design working life of the
structure and which has a high probability of occurrence
NOTE It refers to temporary conditions of the structure, of use, or exposure, e.g. during construction or repair.
1.5.2.4
persistent design situation
design situation which is relevant during a period of the same order as the design working life of the
structure
NOTE Generally it refers to conditions of normal use.
1.5.2.5
accidental design situation
design situation involving exceptional conditions of the structure or its exposure, e.g. fire, explosion, impact
or local failure
1.5.2.6
design working life
the assumed period for which a structure is to be used for its intended purpose with anticipated
maintenance but without substantial repair being necessary
© BSI 04-2000 9
ENV 1991-1:1994
1.5.2.7
hazard
exceptionally unusual and severe event, e.g. an abnormal action or environmental influence, insufficient
strength or resistance, or excessive deviation from intended dimensions
1.5.2.8
load arrangement
identification of the position, magnitude and direction of a free action
1.5.2.9
load case
compatible load arrangements, sets of deformations and imperfections considered simultaneously with
fixed variable actions and permanent actions for a particular verification
1.5.2.10
limit states
states beyond which the structure no longer satisfies the design performance requirements
1.5.2.11
ultimate limit states
states associated with collapse, or with other similar forms of structural failure
NOTE They generally correspond to the maximum load-carrying resistance of a structure or structural part.
1.5.2.12 serviceability limit states
States which correspond to conditions beyond which specified service requirements for a structure or
structural element are no longer met.
1.5.2.12.1
irreversible serviceability limit states
limit states which will remain permanently exceeded when the responsible actions are removed
1.5.2.12.2
reversible serviceability limit states
limit states which will not be exceeded when the responsible actions are removed
1.5.2.13
resistance
mechanical property of a component, a cross-section, or a number of a structure, e.g. bending resistance,
buckling resistance
1.5.2.14
maintenance
the total set of activities performed during the working life of the structure to preserve its function
1.5.2.15
strength
mechanical property of a material, usually given in units of stress
1.5.2.16
reliability
reliability covers safety, serviceability and durability of a structure
1.5.3 Terms relating to actions
1.5.3.1
action
a) Force (load) applied to the structure (direct action)
b) An imposed or constrained deformation or an imposed acceleration caused for example, by
temperature changes, moisture variation, uneven settlement or earthquakes (indirect action).
10 © BSI 04-2000
ENV 1991-1:1994
1.5.3.2
action effect
the effect of actions on structural members, e.g. internal force, moment, stress, strain
1.5.3.3
permanent action (G)
action which is likely to act throughout a given design situation and for which the variation in magnitude
with time is negligible in relation to the mean value, or for which the variation is always in the same
direction (monotonic) until the action attains a certain limit value
1.5.3.4
variable action (Q)
action which is unlikely to act throughout a given design situation or for which the variation in magnitude
with time is neither negligible in relation to the mean value nor monotonic
1.5.3.5
accidental action (A)
action, usually of short duration, which is unlikely to occur with a significant magnitude over the period of
time under consideration during the design working life
NOTE An accidental action can be expected in many cases to cause severe consequences unless special measures are taken.
1.5.3.6
seismic action (AE)
action which arises due to earthquake ground motions
1.5.3.7
fixed action
action which has a fixed distribution over the structure such that the magnitude and direction of the action
are determined unambiguously for the whole structure if this magnitude and direction are determined at
one point on the structure
1.5.3.8
free action
action which may have any spatial distribution over the structure within given limits
1.5.3.9
single action
action that can be assumed to be statistically independent in time and space of any other action acting on
the structure
1.5.3.10
static action
action which does not cause significant acceleration of the structure or structural members
1.5.3.11
dynamic action
action which causes significant acceleration of the structure or structural members
1.5.3.12
quasi-static action
dynamic action that can be described by static models in which the dynamic effects are included
1.5.3.13
representative value of an action
value used for the verification of a limit state
© BSI 04-2000 11
ENV 1991-1:1994
1.5.3.14
characteristic value of an action
the principal representative value of an action. In so far as this characteristic value can be fixed on
statistical bases, it is chosen so as to correspond to a prescribed probability of not being exceeded on the
unfavourable side during a “reference period” taking into account the design working life of the structure
and the duration of the design situation
1.5.3.15
reference period
See 1.5.3.14.
1.5.3.16
combination values
values associated with the use of combinations of actions (see 1.5.3.20) to take account of a reduced
probability of simultaneous occurrence of the most unfavourable values of several independent actions
1.5.3.17
frequent value of a variable action
the value determined such that:
— the total time, within a chosen period of time, during which it is exceeded for a specified part, or
— the frequency with which it is exceeded,
is limited to a given value.
1.5.3.18
quasi-permanent value of a variable action
the value determined such that the total time, within a chosen period of time, during which it is exceeded
is a considerable part of the chosen period of time
1.5.3.19
design value of an action Fd
the value obtained by multiplying the representative value by the partial safety factor *F
1.5.3.20
combination of actions
set of design values used for the verification of the structural reliability for a limit state under the
simultaneous influence of different actions
1.5.4 Terms relating to material properties
1.5.4.1
characteristic value Xk
the value of a material property having a prescribed probability of not being attained in a hypothetical
unlimited test series. This value generally corresponds to a specified fractile of the assumed statistical
distribution of the particular property of the material. A nominal value is used as the characteristic value
in some circumstances
1.5.4.2
design value of a material property Xd
value obtained by dividing the characteristic value by a partial factor *M or, in special circumstances, by
direct determination
1.5.5 Terms relating to geometrical data
1.5.5.1
characteristic value of a geometrical property ak
the value usually corresponding to the dimensions specified in the design. Where relevant, values of
geometrical quantities may correspond to some prescribed fractile of the statistical distribution
12 © BSI 04-2000
ENV 1991-1:1994
1.5.5.2
design value of a geometrical property ad
generally a nominal value. Where relevant, values of geometrical quantities may correspond to some
prescribed fractile of the statistical distribution
1.6 Symbols
For the purposes of this prestandard, the following symbols apply.
NOTE The notation used is based on ISO 3898:1987
Latin upper case letters
A Accidental action
Ad Design value of an accidental action
AEd Design value of seismic action
AEk Characteristic seismic action
Ak Characteristic value of an accidental action
Cd Nominal value, or a function of certain design properties of materials
E Effect of an action
Ed Design value of effects of actions
Ed,dst Design effect of destabilizing actions
Ed,stb Design effect of stabilizing actions
F Action
Fd Design value of an action
Fk Characteristic value of an action
Frep Representative value of an action
G Permanent action
Gd Design value of a permanent action
Gd,inf Lower design value of a permanent action
Gid Characteristic value of permanent action j
Gd,sup Upper design value of a permanent action
Gind Indirect permanent action
Gk Characteristic value of a permanent action
Gk,inf Lower characteristic value of a permanent action
Gk,sup Upper characteristic value of a permanent action
P Prestressing action
Pd Design value of a prestressing action
Pk Characteristic value of a prestressing action
Q Variable action
Qd Design value of a variable action
Qind Indirect variable action
Qk Characteristic value of a single variable action
Qk1 Characteristic value of the dominant variable action
Qid Characteristic value of the non-dominant variable action i
R Resistance
Rd Design value of the resistance
Rk Characteristic resistance
X Material property
Xd Design value of a material property
Xk Characteristic value of a material property
© BSI 04-2000 13
ENV 1991-1:1994
14 © BSI 04-2000
ENV 1991-1:1994
Section 2. Requirements
© BSI 04-2000 15
ENV 1991-1:1994
2.5 Durability
(1) It is an assumption in design that the durability of a structure or part of it in its environment is such
that it remains fit for use during the design working life given appropriate maintenance.
(2) The structure should be designed in such a way that deterioration should not impair the durability and
performance of the structure having due regard to the anticipated level of maintenance.
(3)P The following interrelated factors shall be considered to ensure an adequately durable structure:
— the intended and possible future use of the structure;
— the required performance criteria;
— the expected environmental influences;
— the composition, properties and performance of the materials;
— the choice of the structural system;
— the shape of members and the structural detailing;
16 © BSI 04-2000
ENV 1991-1:1994
© BSI 04-2000 17
ENV 1991-1:1994
3.1 General
(1)P Limit states are states beyond which the structure no longer satisfies the design performance
requirements.
(2) In general, a distinction is made between ultimate limit states and serviceability limit states.
NOTE Verification of one of the two limit states may be omitted if sufficient information is available to prove that the requirements
of one limit state are met by the other.
(3) Limit states may relate to persistent, transient or accidental design situations.
18 © BSI 04-2000
ENV 1991-1:1994
(2) Design values are generally obtained by using the characteristic or representative values (as defined in
sections 4 to 6 and specified in ENVs 1991-1999) in combination with partial and other factors as defined
in section 9 and ENV 1991 to 1999.
(3) In exceptional cases, it may be appropriate to determine design values directly. The values should be
chosen cautiously and should correspond to at least the same degree of reliability for the various limit
states as implied in the partial factors in this code (see also section 8).
NOTE 1 Partial factor design is discussed in Annex A.
NOTE 2 Principles and application rules for verification are given in section 9.
© BSI 04-2000 19
ENV 1991-1:1994
20 © BSI 04-2000
ENV 1991-1:1994
(6) The self-weight of the structure can, in most cases, be represented by a single characteristic value and
be calculated on the basis of the nominal dimensions and mean unit masses. The values are given in
ENV 1991-2.
(7)P For variable actions the characteristic value (Qk) corresponds to either:
— an upper value with an intended probability of not being exceeded or a lower value with an intended
probability of not falling below, during some reference period;
— a nominal value which may be specified in cases where a statistical distribution is not known.
Values are given in ENVs 1991-2 and 1991-3.
(8) The following may be assumed for the time-varying part for most cases of characteristic values of
variable actions:
— the intended probability is [0,98];
— the reference period is [one] year.
However in some cases the character of the action makes another reference period more appropriate. In
addition, design values for other variables within the action model may have to be chosen, which may
influence the probability of being exceeded for the resulting total action.
(9) Actions caused by water should normally be based on water levels and include a geometrical parameter
to allow for fluctuation of water level. Tides, currents and waves should be taken into account where
relevant.
(10) For accidental actions the representative value is generally a characteristic value Ak corresponding to
a specified value.
(11) Values of Ak for explosion and for some impacts are given in ENV 1991-2-7.
(12) For accidental actions arising from fire, information is given in ENV 1991-2-2.
(13) Values of AEd for seismic actions are given in ENV 1998-1.
(14) For accidental actions on bridges arising from the traffic, characteristic values to be used as design
values are given in ENV 1991-3.
(15) For multi-component actions [see 4.1(7)] the characteristic action is represented by groups of values,
to be considered alternatively in design calculations.
© BSI 04-2000 21
ENV 1991-1:1994
(7)P These representative values and the characteristic value are used to define the design values of the
actions and the combinations of actions as explained in section 9. The combination values are used for the
verification of ultimate limit states and irreversible serviceability limit states. The frequent values and
quasi-permanent values are used for the verification of ultimate limit states involving accidental actions
and for the verification of reversible serviceability limit states. The quasi-permanent values are also used
for the calculation of long term effects of serviceability limit states. More detailed rules concerning the use
of representative values are given, for example, in ENVs 1992 to 1999.
(8) For some structures or some actions other representative values or other types of description of actions
may be required, e.g. the fatigue load and the number of cycles when fatigue is considered.
NOTE Further information concerning the specification and combination of actions is given in Annex A and other parts of
ENV 1991.
22 © BSI 04-2000
ENV 1991-1:1994
(1)P Properties of materials (including soil and rock) or products are represented by characteristic values
which correspond to the value of the property having a prescribed probability of not being attained in a
hypothetical unlimited test series. They generally correspond for a particular property to a specified fractile
of the assumed statistical distribution of the property of the material in the structure.
(2) Unless otherwise stated in ENVs 1992 to 1999, the characteristic values should be defined as the 5 %
fractile for strength parameters and as the mean value for stiffness parameters.
NOTE For operational rules, see Annex D; for fatigue, information is given in Annex B
(3)P Material property values shall normally be determined from standardized tests performed under
specified conditions. A conversion factor shall be applied where it is necessary to convert the test results
into values which can be assumed to represent the behaviour of the material in the structure or the ground
(see also ENVs 1992 to 1999).
(4) A material strength may have two characteristic values, an upper and a lower. In most cases only the
lower value will need to be considered. In some cases, different values may be adopted depending on the
type of problem considered. Where an upper estimate of strength is required (e.g. for the tensile strength
of concrete for the calculation of the effects of indirect actions) a nominal upper value of the strength should
normally be taken into account.
(5) Where there is a lack of information on the statistical distribution of the property a nominal value may
be used; where the limit state equation is not significantly sensitive to its variability a mean value may be
considered as the characteristic value.
(6) Values of material properties are given in ENVs 1992 to 1999.
© BSI 04-2000 23
ENV 1991-1:1994
(1)P Geometrical data are represented by their characteristic values, or in the case of imperfections directly
by their design values.
(2) The characteristic values usually correspond to dimensions specified in the design.
(3) Where relevant, values of geometrical quantities may correspond to some prescribed fractile of the
statistical distribution.
(4)P Tolerances for connected parts which are made from different materials shall be mutually compatible.
Imperfections which have to be taken into account in the design of structural members are given in
ENVs 1992 to 1999.
24 © BSI 04-2000
ENV 1991-1:1994
© BSI 04-2000 25
ENV 1991-1:1994
8.1 General
(1)P Where calculation rules or material properties given in ENVs 1991 to 1999 are not sufficient or where
economy may result from tests on prototypes, part of the design procedure may be performed on the basis
of tests.
NOTE Some of the clauses in this section may also be helpful in cases where the performance of an existing structure is to be
investigated.
(2)P Tests shall be set up and evaluated in such a way that the structure has the same level of reliability
with respect to all possible limit states and design situations as achieved by design based on calculation
procedures specified in ENVs 1991 to 1999, including this Part of ENV 1991.
(3) Sampling of test specimens and conditions during testing should be representative.
(4) Where ENVs 1991 to 1999 include implicit safety provisions related to comparable situations, these
provisions shall be taken into account in assessing the test results and may give rise to corrections. An
example is the effect of tensile strength in the bending resistance of concrete beams, which is normally
neglected during design.
26 © BSI 04-2000
ENV 1991-1:1994
iv) the way that safety factors as partial factors or additive elements are applied to get design values
(see 9.3).
The partial factor used in method a) should be chosen in such a way that there is sufficient similarity
between the tests under consideration and the usual application field of the partial factor used in numerical
verifications. (see also 3.4).
(3) When for special cases method b) is used, the determination of the design values should be carried out
by considering:
— the relevant limit states;
— the required level of reliability;
— the statistical and model uncertainties;
— the compatibility with the assumptions for the action side;
— the classification of design working life of the relevant structure according to Section 2;
— prior knowledge from similar cases or calculations.
(4) Further information may be found in ENVs 1992 to 1999.
NOTE see also Annex A and Annex D.
© BSI 04-2000 27
ENV 1991-1:1994
9.1 General
(1)P In ENVs 1992 to 1999 the reliability according to the limit state concept is achieved by application of
the partial factor method. In the partial factor method, it is verified that, in all relevant design situations,
the limit states are not exceeded when design values for actions, material properties and geometrical data
are used in the design models.
(2)P In particular, it shall be verified that:
a) the effects of design actions do not exceed the design resistance of the structure at the ultimate limit
state; and
b) the effects of design actions do not exceed the performance criteria for the serviceability limit state.
Other verifications may also need to be considered for particular structures e.g. fatigue. Details are
presented in the relevant parts of ENV 1991 and in ENVs 1992 to 1999.
NOTE see also Annex A and Annex B.
(3)P The selected design situations shall be considered and critical load cases identified. For each critical
load case, the design values of the effects of actions in combination shall be determined.
(4) A load case identifies compatible load arrangements, sets of deformations and imperfections which
should be considered simultaneously for a particular verification.
(5) Rules for the combination of independent actions in design situations are given in this section. Actions
which cannot occur simultaneously, for example, due to physical reasons, should not be considered together
in combination.
(6) A load arrangement identifies the position, magnitude and direction of a free action. Rules for different
arrangements within a single action are given in ENVs 1991-2, 1991-3 and 1991-4.
(7) Possible deviations from the assumed directions or positions of actions should be considered.
(8) The design values used for different limit states may be different and are specified in this section.
where:
*F is the partial factor for the action considered taking account of:
— the possibility of unfavourable deviations of the actions;
— the possibility of inaccurate modelling of the actions;
— uncertainties in the assessment of effects of actions.
Frep is the representative value of the action.
28 © BSI 04-2000
ENV 1991-1:1994
(2) Depending on the type of verification and combination procedures, design values for particular actions
are expressed as follows:
Gd = *GGk or Gk
Qd = *QQk, *Q?0Qk, ?1Qk, ?2Qk or Qk (9.2)
Ad = *A Ak or Ad
Pd = *P Pk or Pk
AEd = AEd
(3)P Where distinction has to be made between favourable and unfavourable effects of permanent actions,
two different partial factors shall be used.
(4) For seismic actions the design value may depend on the structural behaviour characteristics
(see ENV 1998).
9.3.2 Design values of the effects of actions
(1) The effects of actions (E) are responses (for example internal forces and moments, stresses, strains and
displacements) of the structure to the actions. For a specific load case the design value of the effect of
actions (Ed) is determined from the design values of the actions, geometrical data and material properties
when relevant:
Ed = E(Fd1, Fd2, ... ad1, ad2, ... Xd1, Xd2, ...) (9.3)
where:
Fd1, ..., ad1, ... and Xd1, ... are chosen according to 9.3.1, 9.3.3 and 9.3.4 respectively.
(2) In some cases, in particular for non-linear analysis, the effect of the uncertainties in the models used in
the calculations should be considered explicitly. This may lead to the application of a coefficient of model
uncertainty, *Sd applied either to the actions or to the action effects, whichever is the more conservative.
The factor *Sd may refer to uncertainties in the action model and/or the action effect model.
(3) For non-linear analysis, i.e. when the effect is not proportional to the action, the following simplified
rules may be considered in the case of a single predominant action.
a) When the effect increases more than the action, the partial factor is applied to the representative
value of the action.
b) When the effect increases less than the action, the partial factor is applied to the action effect of the
representative value of the action.
In other cases more refined methods are necessary which are defined in the relevant Eurocodes (e.g. for
prestressed structures).
9.3.3 Design values of material properties
(1)P The design value Xd of a material or product property is generally defined as:
Xd = )Xk / *M or Xd = Xk / *M (9.4)
where:
*M is the partial factor for the material or product property, given in ENVs 1992 to 1999, which covers:
— unfavourable deviations from the characteristic values;
— inaccuracies in the conversion factors; and
— uncertainties in the geometric properties and the resistance model.
) is the conversion factor taking into account the effect of the duration of the load, volume and scale
effects, effects of moisture and temperature and so on.
In some cases the conversion is implicitly taken into account by the characteristic value itself, as indicated
by the definition of ), or by *M.
9.3.4 Design values of geometrical data
(1)P Design values of geometrical data are generally represented by the nominal values:
ad = anom (9.5)
© BSI 04-2000 29
ENV 1991-1:1994
where %a takes account of the possibility of unfavourable deviations from the characteristic values.
%a is only introduced where the influence of deviations is critical, e.g. imperfections in buckling analysis.
Values of %a are given in ENVs 1992 to 1999.
9.3.5 Design resistance
(1)P Design values for the material properties, geometrical data and effects of actions, when relevant, shall
be used to determine the design resistance Rd from:
Rd = R(ad1, ad2, ... Xd1, Xd2, ...) (9.7)
where:
— *R is a partial factor for the resistance;
— *m is a material factor;
— *rd covers uncertainties in the resistance model and in the geometrical properties.
NOTE For further information, see Annex A
(3) The design resistance may also be obtained directly from the characteristic value of a product
resistance, without explicit determination of design values for individual basic variables, from:
Rd = Rk/*R (9.7d)
This is applicable for steel members, piles, etc. and is often used in connection with design by testing.
where:
Ed ,dst is the design value of the effect of destabilizing actions;
Ed ,stb is the design value of the effect of stabilizing actions.
where:
Ed is the design value of the effect of actions such as internal force, moment or a vector representing
several internal forces or moments;
Rd is the corresponding design resistance, associating all structural properties with the respective
design values.
30 © BSI 04-2000
ENV 1991-1:1994
In some cases it may be necessary to replace expression (9.9) by an interaction formula. The required load
cases are identified as described in 9.1.
9.4.2 Combination of actions
(1)P For each critical load case, the design values of the effects of actions (Ed) should be determined by
combining the values of actions which occur simultaneously, as follows:
a) Persistent and transient situations: Design values of the dominant variable actions and the
combination design values of other actions.
b) Accidental situations: Design values of permanent actions together with the frequent value of the
dominant variable action and the quasi-permanent values of other variable actions and the design value
of one accidental action.
c) Seismic situations: Characteristic values of the permanent actions together with the quasipermanent
values of the other variable actions and the design value of the seismic actions.
(2) When the dominant action is not obvious, each variable action should be considered in turn as the
dominant action.
(3) The above combination process is represented in Table 9.1.
Table 9.1 — Design values of actions for use in the combination of actions
Single variable actions Qd Accidental actions or
Design situation Permanent actions Gd
Dominant Others seismic actions Ad
(9.10)
(9.10a)
(9.10b)
[K] is a reduction factor for *Gj within the range 0.85 and 1. From the expressions (9.10a) and (9.10b) the more favourable may
be applied instead of expression (9.10) under conditions defined by the relevant National Application Document.
b) Combinations for accidental design situations
(9.11)
(9.12)
where:
“+” implies “to be combined with”;
implies “the combined effect of”;
© BSI 04-2000 31
ENV 1991-1:1994
32 © BSI 04-2000
ENV 1991-1:1994
© BSI 04-2000 33
ENV 1991-1:1994
9.4.4 ? factors
(1) ? factors for buildings are given in Table 9.3. For other applications see relevant parts of ENV 1991.
Table 9.3 — ? factors for buildings
Action ?0 ?1 ?2
(9.13)
(9.14)
In this case the effect of actions should also be verified for the dominant variable actions using
expression (9.13).
(2) The *G values are given in Table 9.2.
9.4.6 Partial safety factors for materials
Partial safety factors for properties of materials and products are given in ENVs 1992 to 1999.
9.5 Serviceability limit states
9.5.1 Verifications of serviceability
(1)P It shall be verified that:
Ed k Cd (9.15)
where:
Cd is a nominal value or a function of certain design properties of materials related to the design
effects of actions considered; and
Ed is the design value of the action effect (e.g. displacement, acceleration), determined on the basis
of one of the combinations defined in 9.5.2.
34 © BSI 04-2000
ENV 1991-1:1994
(2) Three combinations of actions for serviceability limit states are defined symbolically by the following
expressions:
a) Characteristic (rare) combination
(9.16)
b) Frequent combination
(9.17)
c) Quasi-permanent combination
(9.18)
(9.19)
© BSI 04-2000 35
ENV 1991-1:1994
(9.20)
In this case the effect of actions should also be verified for the dominant variable action using
expression (9.19).
(2) Where simplified prescriptive rules are given for serviceability limit states, detailed calculations using
combinations of actions are not required.
9.5.6 Partial factors for materials
Partial factors for the properties of materials and products are given in ENVs 1992 to 1999.
36 © BSI 04-2000
ENV 1991-1:1994
Annex A (informative)
Partial factor design
A.1 General
(1) This annex gives information and theoretical background concerning the partial factor method as
described in section 9. This annex is also an introduction to Annex D. The information in these annexes
may be used if the verification rules of ENVs 1991-1999 are not considered adequate for the case
considered.
(2) In the partial factor method it is verified that all relevant limit states are not exceeded, given design
values for actions, resistances and geometrical data. The design values are the products or quotients of the
characteristic values and the appropriate partial factors and ? values, as indicated in 9.3 to 9.5. In
general, partial factors are intended to take account of:
— unfavourable deviations from the representative values;
— inaccuracies in the action models and structural models;
— inaccuracies in the conversion factors.
(3) The value of the partial factors should depend on the degree of uncertainty in the actions, resistances,
geometrical quantities and models, and on the type of construction works and the type of limit state.
(4) In principle there are two ways to determine numerical values for partial factors:
a) on the basis of calibration to a long and successful history of building tradition; for most of the factors
proposed in the currently available Eurocodes this is the leading principle;
b) on the basis of the statistical evaluation of experimental data and field observations; this should be
done within the framework of a probabilistic reliability theory.
(5) In practice, the two methods described in A.1(4) can also be used in combination. In particular, a mere
statistical (probabilistic) approach usually fails from a lack of sufficient data. Some reference to traditional
design methods should always be made. Where there has been a long and successful building tradition, it
is of great value to obtain a rational understanding of that success. Understanding may justify the
reduction of some factors for specified conditions, which in their turn may lead to economy. From this point
of view, the statistical methods should be considered as giving added value to the more traditional
approach.
A.2 Overview of reliability methods
(1) Figure A.1 presents an overview of the various methods for reliability verification and the interactions
between them. The probabilistic verification procedures can be subdivided into two main classes: exact
methods and first order reliability methods (FORM), sometimes referred to as level III and level II methods
respectively. In both methods the measures of reliability are failure probabilities P1 for the failure modes
under consideration and for some appropriate reference period. These values are calculated and compared
with some preset target value P0. If the failure probability is larger than the target, the structure is
considered to be unsafe.
© BSI 04-2000 37
ENV 1991-1:1994
(2) In the level II procedures one generally works with an alternative measure of safety, the so-called
reliability index ", which is related to P1 by:
P1 = 9(-") (A.1)
(3) According to Figure A.1, the safety elements of the partial factor method (level I) can be obtained in
three ways:
a) from calibration to historical and empirical design methods;
b) from calibration to probabilistic methods;
c) as a simplification of FORM via the (calibrated) design value method as described in A.3.
The present generation of Eurocodes has been primarily based on method a), with amendments based
on c) or equivalent methods, mainly in the field of design assisted by testing.
(4) Indicative target values for " in various design situations are given in Table A.2. Values are given for
the design working life (see Table 2.1 of ENV 1999-1) and for one year. Values for one year might be
relevant for transient design situations and temporary structures where human safety is of great
importance.
(5) The values in Table A.2 are intended as “appropriate for most cases”. For reasons related to the type
and consequences of failure and economy of building, it may be appropriate to use higher or lower values
(see 2.2). A class difference in reliability level is usually associated with differences in " values in the order
of 0,5 to 1,0. A difference of reliability level may be desired for a total building, some specific components
or some specific hazards.
NOTE 1 A given reliability level may lead to different partial factors for various material properties and loads, depending on their
variability and influence, see A.3 and A.4. This should not be confused with reliability differentiation.
NOTE 2 Choosing a different target reliability index is not the only possible measure for reliability differentiation; other measures
are related to the accuracy of calculation, the degree of quality assurance and the stringency of detailing rules.
38 © BSI 04-2000
ENV 1991-1:1994
(6) The values in Table A.2 should be considered as reasonable minimum requirements, following from
calibration calculations to design codes in various countries. In these calibrations lognormal or Weibull
distributions for resistance parameters and model uncertainties were usually used. Normal distributions
were usually taken for self weight and extreme value distributions for variable loads. It should be noted,
however, that these calibrations showed a wide scatter, depending on the code at hand, the type of
structural component, and the quantification of the various uncertainties.
(7) The value of 3.8 for the ultimate limit state target reliability index is in particular, accepted for many
applications mainly relating to resistance. This, however, does not mean that standard design according to
the Eurocodes automatically would lead to "-values equal or close to this target. In fact, up to now, the
present generation of Eurocodes have not been wholly evaluated in this way. Such an evaluation is not very
straightforward as serviceability, durability, round off effects or multimodal distribution effects may
disturb the picture in many cases. Additionally design rules in codes may also include implicit safety
differentiations depending on the type of failure, especially ductile or brittle behaviour.
(8) Finally, it should be stressed that a "-value and the corresponding failure probability are formal or
notional numbers, intended primarily as a tool for developing consistent design rules, rather than giving a
description of the structural failure frequency.
Table A.2 — Indicative values for the target reliability index "
Target reliability index
Limit state Target reliability index (one year)
(design working life)
where
E is the action effect;
R is the resistance;
F is action;
f is material property;
a is geometrical property;
F model uncertainty.
Note that expression (A.2) is partly symbolic and that sometimes a more general formulation is necessary.
(2) The set of design values for the design point corresponds to the point on the failure surface having the
highest probability of occurrence (see Figure A.2). In this way the design value method is related to the
probabilistic level II method [see A.2(1)].
(3) The design value of action effects Ed and the resistances Rd are defined such that the probability of
having a more unfavourable value is as follows:
P (E > Ed) = 9 (+ µE") = 9 (– 0,7") (A.3a)
P (R < Rd) = 9 (– µR") = 9 (– 0,8") (A.3b)
© BSI 04-2000 39
ENV 1991-1:1994
where:
! is the FORM weight factor (– 1 k ! k + 1);
" is the target value for the reliability index (see Table A.2).
For a load ! is negative; for a resistance parameter ! is positive.
(4) The essence of the method is the setting of the !E and !R values to – 0,7 and + 0,8 respectively. The
validity range for these values is limited for the case " = 3,8 (accepting a maximum deviation of 0.5) to the
ratios:
0,16 < BE/BR < 7,6
Outside this range it is recommended to use ! = ± 1.0 for the variable having the largest value of B.
(5) When the load or resistance model contains several basic variables (other loads, conversion factors,
more materials) expressions (A.3a) and (A.3b) only hold for the dominating variables. For non-dominating
variables:
P {E > Ed} = 9 (– 0,4 × 0,7 × ") = 9 (– 0,28 ") (A.4a)
P {R < Rd} = 9 (– 0,4 × 0,8 × ") = 9 (– 0,32 ") (A.4b)
For " = 3,8 these values correspond approximately to the 0,90 and 0,10 fractiles respectively.
(6) Table A.3 gives expressions for calculating the design values for given ! and ".
Table A.3 — Design values for various distribution functions
Distribution Design values Remarks
Figure A.2 — Design point definition according to first order reliability methods (FORM)
40 © BSI 04-2000
ENV 1991-1:1994
— values calibrated to reach the aimed reliability, e.g. coefficients and model factors.
Additionally there is a set of partial safety factors and load combination factors.
(2) The design values for actions F, material properties X and geometrical properties follow from:
Fd = *fFk or Fd = *f?Fk (? can be ?0, ?1 or ?2) (A.5)
Xd = Xk/*m (A.6)
ad = anom ± %a (A.7)
In this model:
*f takes account of:
— the possibility of unfavourable deviations of the action values from the representative values.
*m takes account of:
— the possibility of unfavourable deviations of the material properties from the characteristic values;
— the systematic part of the conversion factor [if relevant, see also 8.3(1)];
— uncertainties of the conversion factor.
%a takes account of:
— the possibility of unfavourable deviations of the geometrical data from the characteristic (specified)
values governed by the tolerance specifications;
— the importance of variations;
— the cumulative effect of a simultaneous occurrence of several geometrical deviations.
*Rd takes account of:
— the uncertainties of the resistance model if these uncertainties are not covered by the model itself.
*Sd takes account of the uncertainties:
— in the action model;
— in the action effect model.
? takes account of reductions in design values for loads, in particular:
— the combination value ?0*FFk is determined in such a way that the probability of combined action
effect values being exceeded is approximately the same as when a single variable action only is
present. Within the context of a design value approach (A.3), operational formulae are presented in
Table A.4 for the case of two fluctuating loads.
— The frequent value of a variable action ?1Fk corresponds to the value which is exceeded either 5 %
of the time or 300 times per year; the highest value should be chosen.
— The quasi-permanent value ?2Fk corresponds to the time average or to the value with a probability
of being exceeded of 50 %.
(4) The procedure described by expressions (A.8) and (A.9) is theoretically perfect but cumbersome from a
practical point of view. Therefore the following simplifications are made.
a) On the loading side (for a single loading):
Ed = E {*FFk, anom} (A.10)
© BSI 04-2000 41
ENV 1991-1:1994
Providing R is proportional to the strength X, the model uncertainty F and the geometrical property a,
i.e. R ! FaX. the following simple relations apply:
*M = *m*Rd/{1 + %a/anom} (ENVs 1992 and 1995) (A.15)
*R = *m*Rd/{1 + %a/anom} (ENV 1993) (A.16)
*rd = *Rd/{1 + %a/anom} (ENV 1994) (A.17)
For non-linear models, or in the case of multi-variable load or resistance models, commonly encountered in
Eurocodes, these relations become more complicated.
Table A.4 — Expression for ?0
Distribution
?o = Fnon dom/Fdom
General
Normal
(approximation)
Gumbel
Fs() is the probability distribution function of the extreme value of the non dominating load in the
design period T;
9() is the standard normal distribution function;
N is the T/T1;
T is the design period;
T1 is the period of an independent load variation of the slower varying load;
" is the reliability index;
V is the coefficient of variation for the non dominating load.
NOTE For intermittent loads, the parameter T1 is equal to the duration of the load and Fs ( ) represents the unconditional
distribution function of the load intensity; so Fs( ) is not the conditional distribution function given that the load is active.
42 © BSI 04-2000
ENV 1991-1:1994
A.5 Closure
It is clear from A.1 to A.4 that the same level of formal reliability can be obtained in many different ways.
Some partial factors may be put equal to 1,0 and the required safety margin may be included in another
factor. High characteristic values and low safety factors may be taken or vice versa. The various safety
elements form a set of “communicating vessels”. For every individual design situation, however, there is
the possibility of calibrating the specific coefficients in order to obtain the required reliability level.
In the currently available set of Eurocodes, the characteristic values for loads and strength parameters and
the geometrical properties are generally taken in accordance with A.2 to A.4. ENV 1991-1 gives values for
the partial load factors and the material-related design codes give values for the partial resistance factors.
This is mainly done in a global way, partly based on probabilistic considerations, partly on a historical or
empirical motivation. Furthermore, the choice of the representative values and the corresponding values
for partial factors was done taking into account the needs for and aspects of an easy and economic
application of the verification procedure in practical design. This has led to the following requests.
— For common structures the design values for actions or action effects should be independent of the
design values of the resistance.
— There should be only a small set of *F values.
— Only one constant *M value should be given for each material property.
— Further simplifications concerning the safety and serviceability verification as well as in structural
analysis should be possible, i.e. avoiding the need to consider too many load arrangements, load cases,
and load combinations in the relevant design situations.
Annex B (informative)
Fatigue
B.1 The fatigue phenomenon
(1) Fatigue is a local material deterioration caused by repeated variations of stresses or strains.
(2) Low cycle fatigue and high cycle fatigue may be distinguished.
low cycle fatigue is associated with non-linear material and geometric behaviour, e.g. alternating plastic
strains in plastic zones. Criteria to exclude low cycle fatigue are given in ENVs 1992 to 1999.
high cycle fatigue is mainly governed by elastic behaviour. Therefore the analysis model can be elastic.
(3) Criteria for determining whether fatigue assessment is needed are given in ENVs 1992 to 1999.
B.2 Fatigue resistance
(1) Except for cases where the fatigue strength of members is determined in specific tests with a load-time
history close to the actual loading to which they are subjected, the fatigue behaviour of structural members
is generally studied for code purposes by simplified tests. In these tests the members are subjected to
constant amplitude load variations, until excessive deformations or fractures due to cracking occur.
(2) The fatigue strength of a given detail is then defined by a %BR – NR relationship, which approximately
represents the 95 % fractile of survival; where %BR is the stress range and NR the number of cycles up to
failure.
This relationship may be modelled by a standardized linear, bilinear or trilinear line in double logarithmic
scale.
(3) For a range of details, a system of such equidistant %BR – NR curves may be established to allow for
classification.
B.3 Determination of fatigue action effects compatible with the fatigue resistance
(1) Fatigue actions are specified in the other Parts of ENV 1991.
(2) When stress-time histories representative of the fatigue action on a given detail are available, any
stress-time history may be evaluated using the reservoir counting method or rainflow counting method.
These methods enable stress ranges and the numbers of cycles to be determined, together with the
associated mean stresses when these are relevant.
(3) The stress ranges and the number of cycles may be ordered in stress-range frequency distributions or
stress-range spectra.
(4) The stress-range frequency distributions or stress range spectra may be transformed to
fatigue-damage-equivalent constant amplitude stress-range spectra using Miner’s rule.
© BSI 04-2000 43
ENV 1991-1:1994
Annex C (informative)
Serviceability limit state: verification of structures susceptible to vibrations
C.1 General
C.1.1 Objective
(1) This annex gives guidance for serviceability limit state verifications of structures susceptible to
vibrations.
(2) It deals with the treatment of the action side, the determination of the structural response and the
limits to be considered for the structural response to ensure that vibrations are not disturbing or harmful.
(3) Dynamic effects relating to ultimate limit states or fatigue are treated in the other Parts of ENV 1991
and therefore are not considered in this annex.
C.1.2 Sources of vibrations
(1) Vibrations may be induced by the following:
a) people, e.g. on:
— pedestrian bridges;
— floors where people walk;
— floors for sport or dance activities;
— floors with fixed seating and spectator galleries.
b) machines, e.g. on:
— machine foundations and supports;
— bell towers;
— ground with transmitted vibrations.
c) wind, e.g. on:
— buildings;
— towers;
44 © BSI 04-2000
ENV 1991-1:1994
© BSI 04-2000 45
ENV 1991-1:1994
(6) Damping ratios should be evaluated by using suitable experimental procedures, approved theoretical
methods and values derived from collections of reliable measurements of homogenous structural classes.
C.4 Evaluation of the structural responses
C.4.1 General
(1) The evaluation of structural responses depends on the limits that are specified for them.
(2) Limits may be expressed in terms of:
a) r.m.s (root-mean-square) values determined for a certain exposure time
(C.1)
where:
aeff the effective or r.m.s value or the response e.g. the effective acceleration
T is the exposure time;
ai is the response value (e.g. acceleration) for each time step %ti;
t is time;
%ti is the time step.
b) extreme values during a certain exposure time T for narrow banded stochastic responses only
(C.2)
where:
n is the natural frequency of the structure
amax is the expected maximum value of the response, e.g. the maximum acceleration
(3) The structural responses aeff or amax should be compared with the specified limits.
C.4.2 Limiting values for vibrations
C.4.2.1 Human comfort
(1) Where conditions for human comfort are specified, these conditions should be given in terms of an
acceptance criteria according to ISO 2631.
(2) The acceptance criteria should include the relevant acceleration (aeff) – frequency (fs) line for the selected
exposure time and direction of vibration.
C.4.2.2 Functioning of machines
(1) Limits for the movements of the machines should be specified in terms of maximum deflections and
frequency (maximum deflection-frequency lines).
C.4.2.3 Other limits
(1) Limits not covered by acceleration-frequency lines or deflection-frequency lines may be:
— the attainment of a maximum stress (e.g. to avoid permanent deformations);
— the attainment of a maximum stress range (e.g. to avoid a limited fatigue life or accumulative
deflections);
— the attainment of a maximum deformation (e.g. to avoid bumping and for continuous operations).
These limits should be given in the design specifications.
Annex D (informative)
Design assisted by testing
D.1 Scope and objectives
(1) This annex is intended to give guidance for the planning and evaluation of experiments to be carried
out in connection with structural design as indicated in Section 8, when the number of tests is sufficient
for a statistical interpretation of their results.
46 © BSI 04-2000
ENV 1991-1:1994
© BSI 04-2000 47
ENV 1991-1:1994
Where structural properties are conditioned by one or several effects of actions which are not varied
systematically, then these effects should be specified at least by their design values. Where they are
independent of the other parameters of the loading path, design values related to estimated values of
lead combination may be adopted.
e) Testing arrangements
Special attention should be given to measures to ensure sufficient strength and stiffness of the loading
and supporting rigs, and clearance for deflections, etc.
f) Measurements
A list should be made of all relevant properties of each individual test specimen to be determined prior
to the execution of the tests. Similarly a list should be made of observation points and methods for
observation and recording, e.g. time histories of displacements, velocities, accelerations, strains, forces
and pressures, required frequency and accuracy of measurements and measuring devices. Depending
on the type of test it could be recommended to have some measurements available during the test.
g) Evaluation and reporting of the test
For specific guidance, see ENVs 1992 to 1999.
D.3 Evaluation of test results
D.3.1 General
(1) All test results should be evaluated critically. The general behaviour and failure modes should be
compared with the expected ones. When large deviations from the expectation occurs, an explanation
should be sought, involving additional tests if necessary.
(2) Where relevant, the evaluation of test results should be on the basis of statistical methods. In principle
the tests should lead to a statistical distribution for the preselected unknown variables, including the
statistical uncertainties. Based on this distribution, design values, characteristic values and partial safety
factors to be used in partial coefficient design may be derived. If possible, only the characteristic value may
be derived while the partial factor is taken from normal design procedure
(3) If the response (or the strength) of the material depends on the load duration or history, the volume or
scale, the environmental conditions, or other non-structural effects, then the calculation model should take
these items into account by use of appropriate factors (conversion) and scaling rules. Further guidance may
be found in ENVs 1991 to 1999. In particular where codes include implicit safety provisions related to
comparable situations, these provisions should also be applied when testing and may give rise to additional
safety elements in the formulae. An example is the effect of tensile strength in concrete test specimens,
which in many cases is neglected during design.
(4) The result of a test evaluation is valid for the specifications and load characteristics considered.
Extrapolation to cover other design parameters and loadings requires additional information, e.g. from
previous tests or theoretical considerations.
D.3.2 Statistical evaluation of resistance/material tests
D.3.2.1 General
(1) This clause is intended to give the operational formulae for deriving design values from the test types
a) and b) for resistance and material testing [see D.1(3)], where the characteristic value is determined from
a standardized or established distribution of the material properties. Use will be made of Bayesian
procedures with vague prior distributions.
NOTE This leads to almost the same result as classical statistics with confidence levels equal to 0.75.
(2) In 8.3 two different methods are distinguished. In method a) a characteristic value is derived first and
then divided by the relevant partial factor. In method b) a direct determination of the design value is made.
These methods are discussed in D.3.2.2 and D.3.2.3 respectively.
(3) The tables and formulae in D.3.2.2 and D.3.2.3 are based on:
— the normal distribution;
— a complete lack of prior knowledge for the mean;
— a complete lack of prior knowledge for the coefficient of variation in the case “Vx unknown” or, on the
other hand, full knowledge for the coefficient of variation in the case “Vx known”.
48 © BSI 04-2000
ENV 1991-1:1994
In practice there may be prior knowledge that the distribution type is of a more favourable nature (for
instance the lognormal distribution) and there might be partial prior knowledge, both about the mean and
about the standard deviation. This prior knowledge may be based on previous experience with similar cases
and will in general lead to more favourable design values. Further guidance is, however, outside the scope
of this annex.
D.3.2.2 Method a) Assessment via the characteristic value
Assume that a sample of n numerical test results is available. The design value of a variable X is obtained
from:
(D.1)
where:
*M is the partial factor for the design;
)d is the design value of the conversion factor;
Xk(n) is the characteristic value including statistical uncertainty;
The assessment of the conversion factor is strongly dependent on the type of test and the type of material.
No further guidance is given here.
The partial factor should be selected from the field of application under consideration in the test.
The value of kn follows from Table D.1. Table D.1 is based on the 5 % characteristic value and on the normal
distribution. Two cases are considered as follows.
i) The coefficient of variation Vx is known from pre-knowledge; pre-knowledge might be found from the
evaluation of previous tests in comparable situations. What is comparable is determined by engineering
judgement. In that case the row “Vx known” should be used.
The coefficient of variation Vx is not known from pre-knowledge, but must be estimated from the sample:
(D.2)
Vx = sx/mx (D.3)
In this case the row “Vx unknown” should be used.
Table D.1 — Values of kn for the 5 % characteristic value
n 1 2 3 4 5 6 8 10 20 30 Z
Vx known 2,31 2,01 1,89 1,83 1,80 1,77 1,74 1,72 1,68 1,67 1,64
Vx unknown — — 3,37 2,63 2,33 2,18 2,00 1,92 1,76 1,73 1,64
(D.4)
The meaning of all variables is the same in D.3.2.2, however, )d should now cover all uncertainties not
covered by the tests. The value of kn should now follow from Table D.2 or Table D.3.
If X is the dominating variable in the resistance model, kn may follow from Table D.2. The table is based on
the assumption that the design value corresponds to ¶ = 3,8 and ! = 0,8 (see Annex A) and that X is
normally distributed. This gives a value with about 0,1 % probability of observing a lower value.
© BSI 04-2000 49
ENV 1991-1:1994
If both a design value and a characteristic value are determined, a partial factor can be found from
*M = Xk/Xd.
If X is a non-dominating variable, then ! = 0,4 × 0,8 (see Annex A) and Table D.3 should be used. The
probability of observing a lower value is about 10 %.
Table D.2 — Values of kn for the ULS design value, if X is dominating (P{X < Xd} = 0,1 %)
n 1 2 3 4 5 6 8 10 20 30 Z
Vx known 4,36 3,77 3,56 3,44 3,37 3,33 3,27 3,23 3,16 3,13 3,08
Vx unknown — — — 11,40 7,85 6,36 5,07 4,51 3,64 3,44 3,08
Table D.3 — Values of kn for the ULS design value, if X is non-dominating (P{X < Xd} = 10 %)
n 1 2 3 4 5 6 8 10 20 30 Z
Vx known 1,81 1,57 1,48 1,43 1,40 1,38 1,36 1,34 1,31 1,30 1,28
Vx unknown — 3,77 2,18 1,83 1,68 1,56 1,51 1,45 1,36 1,33 1,28
where:
X is the vector of random variables;
W is the set of measurable deterministic variables;
Rt is the theoretical model;
R is the measurable result of the experiment;
D is the unknown coefficient, to be determined by the experiment.
(4) Assume a series of n experiments (i = 1 ... n) is carried out, where:
— the values of W have been taken equal to wi;
— the values of X have been measured as xi;
— the values of R have been measured as ri.
(5) It is recommended that the observed experimental results ri are plotted against the calculated values
Rt (xi wi) according to the model and versus each of the observed basic variables. This plotting procedure is
intended to check whether the calculation models adequately account for the respective variables.
(6) If more than one failure mode is observed in the test results, it is recommended that the tests in a
number of series are repeated. In every series all modes but one should be excluded.
(7) From the test results the following set of observations for the unknown coefficient D may by derived:
di = ri / Rt {xi, wi} (D.6)
50 © BSI 04-2000
ENV 1991-1:1994
(8) It will be assumed that D has a normal distribution. It should be noted that the normal distribution may
be replaced by a lognormal distribution, provided that this can be justified from experience with similar
tests in the past.
(9) The further statistical evaluation of D is the same as in D.3.2. For cases where the deterministic
specifications W are varied and/or the random basic variables X are measured indirectly or not at all,
specialist literature should be consulted.
D.3.4 Design value for quality control tests
(1) Control tests, as defined as test type d in D.1(3), are intended to check the quality of the delivered
products or the consistency of the production characteristics.
(2) It will be assumed that the product under consideration is produced in batches. A batch is tentatively
defined as a set of units, produced by one producer, in a relatively short period, with no obvious changes in
production circumstances.
(3) For discrete products the definition of a unit is generally self-evident. For continuously produced
materials, a unit may be defined as one test specimen, e.g. a concrete test cube.
In practice, batches correspond to, for example:
— a single production of concrete from the same materials and plant;
— structural steel from one melt processed according to the same conditions;
— foundation piles for a specific site.
(4) Quality control may be performed on every unit (total control) or on samples (batch control). Typically,
testing all units requires a non-destructive testing technique. In general a non-destructive testing
technique is not able to predict the strength with the same precision as a destructive testing technique.
Therefore some kind of measurement error has to be included. In theory there is always a measurement
error present, but this can often be ignored.
(5) If sampling is used, a random sample is usually taken. In a random sample each unit of the batch has
the same probability of being sampled.
(6) If quality control is performed on the basis of pre-defined selection rules, the control may lead to three
possible outcomes:
— the batch or unit is rejected: d < 0;
— the batch or unit is critical: d = 0;
— the batch or unit is fully acceptable: d > 0.
Where d is a function of the test result of a single unit or of the combined test result of the units in a sample.
A common formulation for an acceptance rule is given by:
m x > X c + 2 n sx (D.7)
where:
mx is the sample mean;
sx is the sample standard deviation;
Xc is a fixed value, for instance the required characteristic value;
2n is a number, normally depending on n.
From which d = mx – 2n sx – Xc
The number of tests n and the parameters 2n and Xc should be determined in such a way that an economical
and efficient test is obtained.
(7) In practice two requirements are often defined which should be met simultaneously. In those cases the
batch is accepted only if for example d1 > 0 and d2 > 0. The second requirement is often related to the lowest
observation, and could be of the type:
xmin > Xc (D.8)
(8) The design value corresponding to given quality control criteria should be calculated on the basis of:
— the operations characteristic of the control rules; this is the probability of some given batch being
accepted;
© BSI 04-2000 51
ENV 1991-1:1994
— the production characteristic; this is the information about the batch-to-batch variation in the
uncontrolled supply.
General formulae are out of the scope of this annex.
(9) Consider by way of example the case that x has a normal distribution, a known standard deviation, that
there is no prior knowledge about the mean and that a single criterion [as in expression (D.7)] is present.
The design or characteristic value based on the critical batch (having d = 0) is then given by:
Xk or Xd = Xc + (2n – kn)Bx (D.9)
The value of kn follows from Table D.1, Table D.2 and Table D.3, where “Vx known” must be assumed. It
should be noted that in most quality control tests there is substantial information on the mean, which leads
to more favourable values. This is as also stated in D.3.4(8) and is outside the scope of this annex.
(10) Finally for total or unit-by-unit testing, it is reasonable to expect some substantial error, as this is
normally carried out by a non-destructive testing procedure. It is assumed here that an error e is present
with mean zero and standard deviation Be. It is further assumed that the mean and standard deviation of
x, either of the batch or the total supply, are known:
(D.10)
The result for this case is also conservatively based on the “critical unit” and not on the “arbitrary accepted
unit”. The value of kn follows from Table D.1, Table D.2, and Table D.3, where “Vx known” may be assumed.
D.3.5 Proof Loading
(1) Proof loading is a test on the actual structure, i.e. test type f in D.1(3). Special care should be taken that
the structure is not unnecessarily damaged during the test. This requires a continuous monitoring of the
load and the response.
(2) A distinction is made between:
— an acceptance test and
— a strength test.
(3) The acceptance test is intended to confirm that the overall structural performance complies with design
intentions. The load is raised to values between the characteristic value and the design value for the
ultimate limit state. Requirements may be set for the deformations, the degree of non-linearity and the
residual deformations after removal of the test loading.
(4) The strength test is intended to show that the structure or the structural element has at least the
strength that is assumed in the design. If an assessment for the test element only is required, it is sufficient
to raise the load to the design load for the ultimate limit state. Obviously, as already stated in D.3.5(1) care
should be taken not to damage the structure unnecessarily.
(5) If the strength test is intended to prove that other but similar elements also have the required strength,
a higher load is required. A minimum requirement in this respect would be to correct the design load for
the presence of better material properties in the tested element, compared to the design values. This means
that the material properties of the tested element have to be measured.
(6) If the relationship between the resistance and the material property is linear, the design strength Rd
corresponding to a successful test with test load Ft is:
Rd = Ft Xd / Xt (D.11)
Xt is the strength of material in the test.
From the requirement Rd U Fd, the minimum test load can be calculated.
(7) If it is not possible to measure the material properties, the design value for the resistance can
conservatively be found from:
Rd = Ft (1 – kn VR) (D.12)
Here VR is the known coefficient of variation for the resistance of the element population under
consideration and kn follows from Table D.2. The case with Vx unknown is outside the scope of this annex.
(8) It is also possible to use a combination of expressions (D.11) and (D.12), e.g. if only a part of the relevant
random variables can be measured. If V is not know by pre-knowledge, a more sophisticated analysis is
required. This is outside the scope of this annex.
52 © BSI 04-2000
blank
DD ENV
1991-1:1996
BSI — British Standards Institution
BSI is the independent national body responsible for preparing
British Standards. It presents the UK view on standards in Europe and at the
international level. It is incorporated by Royal Charter.
Revisions
It is the constant aim of BSI to improve the quality of our products and services.
We would be grateful if anyone finding an inaccuracy or ambiguity while using
this British Standard would inform the Secretary of the technical committee
responsible, the identity of which can be found on the inside front cover.
Tel: 020 8996 9000. Fax: 020 8996 7400.
BSI offers members an individual updating service called PLUS which ensures
that subscribers automatically receive the latest editions of standards.
Buying standards
Orders for all BSI, international and foreign standards publications should be
addressed to Customer Services. Tel: 020 8996 9001. Fax: 020 8996 7001.
Information on standards
Copyright
Copyright subsists in all BSI publications. BSI also holds the copyright, in the
UK, of the publications of the international standardization bodies. Except as
permitted under the Copyright, Designs and Patents Act 1988 no extract may be
reproduced, stored in a retrieval system or transmitted in any form or by any
means – electronic, photocopying, recording or otherwise – without prior written
permission from BSI.
This does not preclude the free use, in the course of implementing the standard,
of necessary details such as symbols, and size, type or grade designations. If these
details are to be used for any other purpose than implementation then the prior
written permission of BSI must be obtained.