Chapter 1-4 Reliability Basis of Eurocode 1

1-4 The Reliability Basis of Eurocode

M Holický, JV Retief & PE Dunaiski

1 INTRODUCTION

One of the primary advances made with the development and implementation of the Eurocode
Standards for structural design (CE 2002) is the extension of the general principles of structural
reliability as formulated in ISO 2394 General principles on the reliability of structures (ISO 1998)
into operational reliability based partial factor limit states design procedures. The principles and
procedures are formulated in general terms in EN 1990 Basis of structural design (CEN 2002,
Gulvanessian 2002) and then applied in EN 1991 Actions on structures and EN 1992 – EN 1999
which specifies the materials related resistance requirements of structural design.
By capturing the principles of the reliability basis of structural design into a head Standard EN
1990, Eurocode not only strengthened its operational development into reliability based design
procedures, but also provides the basis for unified treatment of actions on structures and materials
based resistance, including foundations and earthquake resistance as the series of parts involved in the
design of an individual structure, but also between the parallel comparative performance requirements
for structures constructed from different structural materials as diverse and ranging from steel to
geotechnical materials.
The fundamental reliability based methodology taken in Eurocode at least at the conceptual
level was the only way in which the wide range of approaches and practice followed by Member
States could be harmonised. Even where such harmonisation could not be achieved, such deviations
were constrained to the National Determined Parameters, sometimes even to the range of values
allowed for basic variables. Furthermore concerted efforts are made to improve the degree of
consensus in the selection of NDP options and values presently and in the future.
This Chapter presents a review and analysis of the reliability basis of Eurocode, with an
assessment made from the view of the revision of the present South African Loading Code SANS
10160-1989 (Amended) The general procedures and loadings to be adopted in the design of buildings
(SANS 1994) into the proposed SANS 10160 (Draft) Basis of structural design and actions for
buildings and industrial structures. Since Eurocode has been selected as reference to the formulation
of SANS 10160 (Draft) such an assessment of the reliability basis of Eurocode serves as background
to the draft South African Standard.
The basic approach taken was to maintain the general reference level of reliability of the
existing Standard SANS 10160-1989 (Amended), but to apply the Eurocode reliability framework to
improve the consistency of reliability across the scope of application of the Draft Standard. It is argued
that this approach will improve harmonisation with international practice, particularly with that of
Europe which is a major South African trade partner. Future unification with structural resistance
design procedures will also be facilitated.
Following a presentation on the background to the development of standards for structural
design, an overview of the reliability framework of Eurocode as formulated in EN 1990 is presented.
The use of this framework in formulating application rules for action combinations, the specification
of actions on structures, and general guidelines for materials based resistance is presented next. The
theory of structural reliability is then applied in parametric analysis to determine appropriate partial
factors. In conclusion the merit of the Eurocode reliability basis is assessed to derive guidelines for the
formulation of SANS 10160 (Draft).

2 BACKGROUND TO STANDARDS FOR STRUCTURAL DESIGN

Chapter 1-4 Reliability Basis of Eurocode 2

The improved economy of structures designed using the principles of limit states procedures is
underpinned by the theory of structural reliability in order to achieve the required levels of safety
across the range of structures and conditions to which they are subjected.
Since structural reliability provides the theoretical and conceptual basis for the unification of
the Eurocode structural design standards, an overview is provided in this chapter of the theory of
structural reliability and its use to derive design procedures. First a brief overview is provided of the
development of structural development standards into reliability based partial factor limit states
procedures.

2.1 General principles for reliability based Limit States Design procedures

The principles for the application of the theory of structural reliability to derive design
procedures are captured in the International Standard ISO 2394:1998 General principles on reliability
for structures. This Standard is accepted internationally as the basis for the formulation of National
Standards. These general principles are reformulated and developed into practical application rules in
Eurocode EN 1990 Basis for structural design, which are then implemented the design standards EN
1991 – EN 1999 for actions and structural resistance.

2.1.1 Design requirements

It is required that the structural reliability has to be guaranteed during the whole economically
reasonable working life. In particular, the construction works must be designed and built in such a way
that the loading liable to act during its construction and use does not cause:
- collapse of the whole or part of the work,
- major deformations to an inadmissible degree,
- damage to other parts of the works, equipment or installed devices,
- disproportionate damage in relation to the original cause.

The operational methods of structural design are further based on the concept of design
situations and relevant limit states in conjunction with the partial factor method. A distinction is made
between ultimate limit states and serviceability limit states. Limit states can be further differentiated
into design situations to provide for all conditions that can reasonably be expected to occur during the
life of the structure. In general, four types of the design situations are recognised as persistent
situations for normal use; transient situations for temporary conditions; accidental situations for
exceptional conditions; seismic situations which refer to earthquake events.

2.1.2 Limit states of structural performance requirements

The limit states denote particular circumstances beyond which the structural performance
requirements are no longer satisfied:

The ultimate limit states are those associated with various forms of structural failure or states close to
structural failure and may require consideration of:
- loss of equilibrium of the structure considered as a rigid body;
- excessive deformation or settlement, rupture, or the loss of stability.

The serviceability limit states are those associated with the criteria for the structure related to its use
or function and may require consideration of:
- deformation or deflection;
- vibrations which limit the structural use;
- detrimental cracking.

2.1.3 Uncertainties
Chapter 1-4 Reliability Basis of Eurocode 3

It is well recognised that construction works are complicated technical systems suffering from
a number of significant uncertainties in all stages of execution and use. Some uncertainties can never
be eliminated absolutely and must be taken into account when designing or verifying construction
works. The following types of uncertainties can usually be identified, presented in the approximate
order of decreasing knowledge and theoretical tools for analysis and implementation in design:
- natural randomness of actions, material properties and geometric data;
- statistical uncertainties due to a limited size of available data;
- model uncertainties caused by a simplification of actual conditions;
- vagueness due to inaccurate definitions of performance requirements;
- gross errors in design, execution and operation of the structure;
- lack of knowledge of the behaviour of new materials in real conditions.

The lack of available theoretical tools is obvious in the case of gross errors and lack of
knowledge, which are nevertheless often the decisive causes of structural failures. To limit gross errors
due to human activity a quality management system including the methods of statistical inspection and
control may be effectively applied. Structural reliability forms the basis of contemporary systems of
quality control and their operational techniques (ISO 1993; ISO 1997).

2.1.4 Structural reliability

The theory of structural reliability has been developed to describe and analyse uncertainties in
a rational way and to take them into account in the design and verification of structural performance.
In fact, the development of the whole theory was initiated by observed insufficiencies and structural
failures caused by various uncertainties.
In Eurocode EN 1990:2002 it is required that a structure shall be designed and executed in
such a way that it will, during its intended life with appropriate degrees of reliability and in an
economic way:
- remain fit for the use for which it is required; and
- sustain all actions and influences likely to occur during execution and use.

The specification for reliability includes four important elements:

- performance requirements – the definition of the structural failure,
- time period – the specification of the required service-life T,
- reliability level – the specification of the probability of failure pf,
- conditions of use – the restriction of input uncertainties.

2.2 Historical developments of design procedures

During its historical development the design methods to take into account recognised
uncertainties and to ensure structural reliability have been closely linked to the available empirical,
experimental as well as theoretical knowledge of mechanics and the theory of probability. Empirical
methods for structural design gradually evolved into rationally based methods: permissible stresses;
global safety factors; partial factor limit states methods; reliability based design. The basis for these
methods is briefly reviewed.

2.2.1 Permissible stress methods

The first worldwide design method for civil structures is the method of permissible stresses. It
is based on the condition:
σmax < σper, where σper = σcrit / k (1)
Chapter 1-4 Reliability Basis of Eurocode 4

where the coefficient k is assessed with regard to uncertainties in the determination of local load effect
σmax and of resistance σper and therefore may ensure with an appropriate level of security the reliability
of the structure.
The main insufficiency of this method is perhaps the local verification of reliability (in the
elastic range) and the impossibility to consider separately the uncertainties of basic quantities and the
uncertainties of computational models for the assessment of action effects and structural resistance. In
this method, the probability of failure is controlled by one quantity only, the coefficient k.

2.2.2 Global safety factor method

The second widespread method of structural design is the method of global safety factor. It is
based on the condition:
s = Xresist / Xact > s0 (2)
according to which the calculated safety factor s must be greater than its specified value s0 through the
aggregate quantities of structural resistance Xresist and action effect Xact. Although the structural
element or its section is considered, the probability of failure can again be controlled by one quantity
only, the global safety factor s.

2.2.3 Partial factor method

At present, the most advanced operational method of structural design is the partial factor
format based on the condition:
Ed (Fd, fd, ad, θd) < Rd (Fd, fd, ad, θd) (3)
where the action effect Ed and the structural resistance Rd are assessed according to the design values
of basic quantities describing the action Fd, material properties fd, dimensions ad and model
uncertainties θd. The design values of these quantities are determined (taking into account their
uncertainties as well as the uncertainties of computational models) using their characteristic values
(Fk, fk, ak, θk), partial safety factors γ, combination factors ψ and other measures of reliability. Thus a
whole system of various partial factors and other reliability elements may be used to control the
probability of structural failure.
Partial factor methods provide operational design procedures adopted in many national and
international standards. General concepts of the partial factor method are described in International
Standard ISO 2394-1998. The fundamental Eurocode EN 1990:2002 Basis of Structural Design
describes detail of accepted procedures based on a partial factor method in Section 6 Verification by
the partial factor method and Annex A1 (Normative) Application for buildings. Additional
background information and the theoretical basis of the method are described in Annex C
(Informative) Basis for partial factor design and reliability analysis.
The partial factor method was introduced with limit states design in which differentiated
requirements were set for safety (ultimate) and operational (serviceability) limits. From the view of
reliability based design the two limit states can be considered as basic reliability differentiation in
terms of consequences of the limits being exceeded.

2.2.4 Probabilistic design methods

Probabilistic design methods that are provided for in ISO 2394-1998 are based on the
condition that the probability of failure pf does not exceed during the service life of a structure T a
specified target value pt
pf ≤ pt (4)

The probability of failure pf is assessed using a computational structural model, defined

through basic quantities X [X1, X2, ... , Xn] for actions, mechanical properties and geometrical data. The
limit state of a structure is defined by the function for the safe state:
Chapter 1-4 Reliability Basis of Eurocode 5

g(X) ≥ 0 (5)

Design directly based on probabilistic methods is allowed in EN 1990 Clause 3.5(5) as an

alternative to the partial factor method. The JCSS Model Code provides consistent procedures for the
design of structures from a probabilistic point of view (JCSS 2001).

2.2.5 The Design Value Method

The design value method, which is also called a semi-probabilistic (level I) method is a very
important step from probabilistic design methods toward an operational partial factors method. The
design value method is directly linked to the basic principle of EN 1990, according to which it should
be verified that no limit state is exceeded when the design values of all basic variables are used in the
models of structural resistance R and action effect E in terms of the design values Ed and Rd, a
structure is considered as reliable when the following expression holds
Ed < Rd (6)

where Ed = E {Fd1, Fd2, ... ad1, ad2, ... θd1, θd2 , ...} (7)

Rd = R {Xd1, Xd2, ... ad1, ad2, ... θd1, θd2, ...} (8)

Here F is a general symbol for actions, X for material properties, a for geometrical properties, and θ
for model uncertainties.

The design values Ed and Rd may be found by:

P(E > Ed ) = Φ(+αEβ) (9)

P(R ≤ Rd ) = Φ( −αRβ) (10)

where β is the target reliability index, αE and αR, with |α| ≤ 1, are the values of the FORM weight
(sensitivity) factors; taken in EN 1990 as αE = − 0,7 and αR = 0,8.

2.3 Calibration

The assessment of various reliability measures (characteristic values, partial and combination
factors) in the new structural design standards is partially based on probabilistic considerations, but to
a great extent they are based on historical and empirical experience. Moreover, the choice of these
reliability measures is affected in EN 1990 by the intention to simplify the calculation in practical
design. Figure 1 shows possible calibration procedures indicated in EN 1990.
Chapter 1-4 Reliability Basis of Eurocode 6

Deterministic methods Probabilistic methods

Historical methods FORM Probabilistic

Empirical methods level II level III

Calibration (b)

Calibration(a) Semi probabilistic methods

Calibration Calibration (c)
Level I

Partial Factor Method

Figure 1 Alternative calibration procedures

The most frequently used calibration procedure (a) is based on past experience represented by
historical and empirical design methods. However, reliability principles (calibrations (b) and (c))
represent very important theoretical bases enabling further improvement and generalization of existing
design procedures. The theory of structural reliability makes it possible to extend the general
methodology for new structural systems and construction materials.

3 RELIABILITY FRAMEWORK AND SPECIFICATION IN EUROCODE

The International Standard ISO 2394 General principles on the reliability of structures (ISO
1998) is globally accepted to provide the reliability basis for structural design on which structural
design standards need to be based, as stated in its Introduction: This International Standard constitutes
a common basis for defining design rules relevant to the construction and use of the wide majority of
buildings and civil engineering works, whatever the nature or combination of the materials used. … It
has a conceptual character and it is of a fairly general nature. Whilst Eurocode EN 1990 Basis of
structural design is consistent with ISO 2394, its development of reliability based partial factor Limit
States design is taken further in order to provide operational design rules for application in EN 1991 –
EN 1999.

3.1 Comparison of ISO 2394-1998 and EN 1990-2002

A comparison of the main features of ISO 2394 and EN 1990 is presented in Table 1.
Although the two Standards are comparable in total length, the normative part of EN 1990 is more
substantial. The main differences remain in the treatment of design verification procedures where ISO
2394 considers probability-based and partial factor procedures on an equal footing (including also the
assessment of existing structures), EN 1990 develops the partial factor method into operational
procedures. An important common feature is the inclusion of the reliability basis partial factor design
in informative annexes, including design assisted by testing.

Table 1 Comparison of the scope of ISO 2394 and EN 1990

ISO 2394 EN 1990

# pages
INTRODUCTORY AND GENERAL SECTIONS
Chapter 1-4 Reliability Basis of Eurocode 7

1-3 Scope, Definitions, Symbols 21 37

1-2 General; Requirements
4 Requirements and concepts 3 Principles of limit states design
5 Principles of limit states design 4 Basic variables
6 Basic variables 5 Structural analysis and design assisted by
7 Models testing
DESIGN VERIFICATION PROCEDURES
8 Principles of probability-based design 2,5 - Provided for in Clause 3.5(5)
General; Systems reliability vs. element reliability;
Specified degrees of required reliability;
Calculation of failure probabilities; Implementation
9 Partial factors format 4 19 6 Verification by the partial factors method
Design conditions and values; Action Annex A1 (Normative) Applications for
characteristic, combination, frequent, quasi- buildings Design values (Actions, effects;
permanent values; Characteristic values (Materials Material; Geometry, Resistance); Ultimate LS
including soils; Geometry); Load cases and load (Static equilibrium & resistance; Combination of
combinations; Action effects and resistances; actions; Partial factors for actions &
Fatigue; Calibration combinations; Materials & products);
10 Assessment of existing structures 2,5 Serviceability LS (Verification; Criteria;
Relevant cases; Principles of assessment; Basic Combinations; Partial factors for materials)
variables; Investigation; Assessment of damage
INFORMATIVE ANNEX
A Quality management and quality assurance 40 30 B Management of structural reliability for
B Examples of permanent, variable and construction works
accidental actions
C Models for fatigue
D Design based on experimental models D Design assisted by testing
E Principles of reliability-based design C Basis for partial factor design and reliability
F Combination of actions, estimation of a-values analysis
G Example of a method of combination of
actions
TOTAL 81 86

3.2 Design working life

The design working life of a structure is intended to relate to design and maintenance as
summarised in Table 2. Included are also indicative values from EN 1990 which presents a slight
elaboration of values from ISO 2394.

Table 2 Application and indicative values for design working life

Application of Design Working Life (DWL)

• Assumed period for which a structure or part of it is to be used for its intended purpose with anticipated maintenance
but without major repair being necessary;
• Design: Actions (e.g. imposed, wind, earthquake) and deterioration of material (e.g. fatigue, creep, durability);
• Maintenance: Strategies and procedures for maintenance and renovation.
Categor Notional DWL
Examples
y (years)
1 10 Temporary structures (e.g. scaffolding)
2 10-25 Replaceable structural parts, e.g. gantry girders, bearings
3 15-30 Agricultural and similar structures (e.g. buildings where people do not normally enter)
4 50 Building structures and other common structures (e.g. hospitals, schools etc.)
5 100 Monumental building structures, bridges, other civil engineering structures (e.g. churches)
Chapter 1-4 Reliability Basis of Eurocode 8

3.3 Scope of operational procedures presented in EN 1990

The background and an interpretation of EN 1990-2002 is presented by Gulvanessian (2002)

and Holický & Vrouwenvelder (2005). The EN 1990 operational procedures which derive from the
principles, which also closely agree with ISO 2394 as indicated above, can be summarised as follows:
• Reliability framework: The various reliability elements are developed into a reliability framework
consisting of {reliability differentiation; limit states; design situations; basic variables}
• Action combination schemes: Appropriate action combination schemes are developed for the
respective “cells” of the reliability framework
• Design situations: Provision is made and implemented for ranges of design situations {transient;
persistent; accidental; seismic (accidental); equilibrium; strength (structural; geotechnical)}; for
serviceability {irreversible; reversible; long-term}
• Design verification: Provision is made for {basic variables (characteristic values); partial factors;
combination factors (combination; frequent; quasi-permanent)}
• Actions: Standard range extended to include {transient (execution); geotechnical; accidental
(unidentified [integrity & robustness]; fire; impact; internal explosion); seismic}
• Materials: Standard range of structural materials {concrete; steel; composite; timber; masonry;
alloy aluminium} extended to include {geotechnical structures}
• Scope of structures: {generic; buildings; bridges; others}

3.4 EN 1990 reliability differentiation guidelines as compared to ISO 2394

EN 1990 establishes reliability classes RC1 – RC3 (also called consequence classes CC1 –
CC3), with β values shown in Table 3:

RC3 High consequence for the loss of human life; economic, social or environmental
consequences very great;

RC2 Reference class of medium consequences of most conventional structures;

RC1 Low consequences for the loss of human life; small or negligible economic, social or
environmental consequences.

The EN 1990 specification to achieve reliability required for structures … by design … and
by appropriate execution and quality management measures (Clause 2.2(1)P) with the choice of the
levels of reliability for a particular structure should take account of the relevant factors (Clause
2.2(3)) also provides that different levels of reliability may be adopted inter alia for structural
resistance; for serviceability (Clause 2.2(2)). Guidelines for the application of reliability management
in terms of reliability differentiation, provision for design supervision and inspection during execution,
including adjustment of partial factors for resistance are presented in Annex B (Informative).

Table 3 Target reliability levels (β) according to EN 1990

Reliability EN 1990 Minimum values for β

Class Ultimate LS Fatigue Serviceability LS
Reference
1 year 50 years 1 year 50 years 1 year 50 years
period
RC3 5,2 4,3
RC2 4,7 3,8 1,5 to 3,8 2,9 1,5
RC1 4,2 3,3

3.5 EN 1990 Nationally Determined Parameters

Chapter 1-4 Reliability Basis of Eurocode 9

The fundamental alternative expressions for action combinations for transient and persistent
design situations for the Ultimate Limit State are presented in symbolic terms in the normative
Standard as either Expression (6.10) or Expressions (6.10a) and (6.10b). The Nationally Determined
Parameter (NDP) choices are all collected presented in Annex A1 (Normative) Application for
buildings. NDP choices include the expressions to be selected for the various design situations, for
which no recommendations are made; and the partial action and combination factor values for which
values are recommended.

3.6 Design assisted by testing

The informative EN 1990 Annex D elaborates design assisted by testing, providing procedures
that may be followed when testing is used in the design process and in the verification of limit states.
The procedures provide the basis for complying to the principle of Clause 5.2(2)P Design assisted by
testing shall achieve the level of reliability required for the relevant design situation. Provision is
made for the determination of either
- representative or characteristic values of actions, resistance or material properties; to
be used together with appropriate partial factors;
- design values determined directly.

The procedures provide an important basis for activities related to design such as testing of materials
or structural products, wind tunnel testing and soil tests.

4 DESIGN VERIFICATION FOR BUILDINGS

The design verification procedures are presented in EN 1990 Section 6 Verification by the
partial factor method in symbolic terms for all the various limit states and design situations and
applied to buildings in Annex A1 (Normative) Application for buildings. The alternative combination
schemes and the recommended partial factors for buildings are discussed here.

4.1 Action combination values

Representative values for actions consist of:

Characteristic values (Fk)

Combination value of a variable action (ψ0Qk): Chosen so that the probability that the action
effect values caused by the combination will be exceeded is approximately the same as when
a single action is considered

Frequent value of a variable action (ψ1Qk): Determined so that the total time, within a
chosen period of time, during which it is exceeded, is only a small given part of the chosen
period of time; or the frequency of its excess is limited to a given small value

Quasi-permanent value of a variable action (ψ2Qk): Determined so that the total time,
within a chosen period of time during which it is exceeded, is of the magnitude of half the
chosen period

The recommended variable action combination values for buildings are listed in Table 4.

Table 4 Recommended combination factors for actions applied to buildings

Chapter 1-4 Reliability Basis of Eurocode 10

Variable Action ψ0 ψ1 ψ2
Imposed loads on buildings
- Category A/B: domestic residential, office areas 0,7 0,5 0,3
- Category C/D: congregation, shopping areas 0,7 0,7 0,6
- Category E: storage areas 1,0 0,9 0,8
- Category F/G: traffic area (vehicle < 30kN / >30kN) 0,7 0,7/0,5 0,6/0,3
- Category H: roofs 0 0 0
Wind loads on buildings 0,6 0,2 0
Temperature loads in buildings 0,6 0,5 0

4.2 Action combination schemes for various design situations

4.2.1 Design situations

The various combination schemes for actions are given in symbolic terms in Table 5 for the
different design situations of the ultimate and serviceability limit states.

Table 5 Action combination schemes for design situations

Design Situation Ultimate Limit State

∑ γ G , jGk , j "+" γ P P "+" γ Q ,1Qk ,1 "+" ∑ γ Q,iψ 0,iQk ,i [(6.10)]

j ≥1 i >1

Persistent, Alternative scheme:

Transient ∑ γ G, jGk , j "+" γ P P "+" γ Q ,1Ψ0,1Qk ,1 "+" ∑ γ Q,iψ 0,iQk ,i [(6.10a)]

j ≥1 i >1

∑ ξγ G, jGk , j "+" γ P P "+" γ Q ,1Qk ,1 "+" ∑ γ Q,iψ 0,iQk ,i [(6.10b)]

j ≥1 i >1

∑ Gk , j "+" P "+" Ad "+" (ψ 1,1 orψ 2,1 )Qk ,1 "+" ∑ψ 2,iQk ,i

Accidental j ≥1 i >1

∑ Gk , j "+" P "+" AE , d "+" ∑ψ 2,iQk ,i

j ≥1 i >1
Seismic
An importance factor γI is applied to the seismic action AE,d.

Serviceability Limit State

Characteristic, ∑ Gk , j "+" P "+" Qk ,1 "+" ∑ψ 0,iQk ,i

(Irreversible) j ≥1 i >1

Frequent ∑ Gk , j "+" P "+" Ψ1,1Qk ,1 "+" ∑ψ 2,iQk ,i

(Reversible) j ≥1 i >1

Quasi-permanent ∑ Gk , j "+" P "+" ∑ψ 2,iQk ,i

(Long-term) j ≥1 i ≥1

4.2.2 Recommended partial factors for ultimate limit states buildings

Chapter 1-4 Reliability Basis of Eurocode 11

Recommended partial factors for the various limit states and design situations are summarised
in Table 6; with ξ = 0,85. Application of the two alternative combination schemes {Expression (6.10)
or Expression (6.10 a and b)} is presented (Brozzetti (2000); a third alternative with γQ,a = 0 is
revealed; this option is quite similar to the present SANS 10160-1989 (Amended) scheme (Holický &
Retief 2005). Set A provides for static equilibrium, whilst Set B and Set C are defined in order to
provide for geotechnical design, as discussed below. For the ultimate limit state in the accidental and
seismic design situations the partial factors are specified as 1,0.

Table 6 Recommended partial factors for ultimate limit state combination sets

Action
Permanent Variable
Set Ultimate Limit State
Unfavourable Unfavourable
/Favourable /Favourable
Set A EQU: Static equilibrium for buildings
- Equilibrium only 1,10 / 0,90 1,50 / 0
- with Resistance (instead 1,35 / 1,15 1,50 / 0
of A & B)
Set B STR/GEO: Alternative structural/geotechnical general
equations
Alt 1 - Equation (6.10) 1,35 / 1,00 1,50 / 0
Alt 2 - Equation (6.10a) 1,35 / 1,00 1,50 ψ0 (= 1,05)
- Equation (6.10b) ξ1,35 / ξ1,00 /0
1,50 / 0
Alt 3 - Equation (6.10a) modified 1,35 / 1,00 0
- Equation (6.10b) ξ1,35 / ξ1,00 1,50 / 0
Set C STR/GEO: Geotechnical actions and ground resistance
involved
- General equation 1,00 1,30 / 0

4.2.3 The serviceability limit state

For serviceability limit states the partial factors for actions should be taken as 1,0 except if
differently specified in EN 1991 to EN 1999. Serviceability limit states in buildings should take into
account criteria related, for example, to floor stiffness, differential floor levels, storey sway or/and
building sway and roof stiffness. The serviceability criteria should be specified for each project and
agreed with the client. The serviceability criteria may be defined in the National annex.
The serviceability criteria for deformations and vibrations shall be defined depending on the
intended use in relation to the serviceability requirements and independently of the materials used.

4.2.4 Alternative Design Approaches for geotechnical design

Three alternative design approaches (Approach 1, 2 & 3) for the design of structural members
(footings, piles, basement walls, etc.) (STR) involving geotechnical actions and the resistance of the
ground (GEO) are introduced with various combinations of the use of Set B and Set C together with
alternative sets of material and resistance factors specified in EN 1997-1 Geotechnical design –
General rules.
Chapter 1-4 Reliability Basis of Eurocode 12

• Approach 1 alternatively applies Set B and Set C for structural and geotechnical actions; usually
resulting in structural resistance governed by the Set B loading case and sizing of foundations
governed by the Set C loading case.
• Approach 2 applies Set B to both structural and geotechnical actions.
• Approach 3 applies Set B to structural actions and Set C to geotechnical actions in a single load
case.

Combination with the alternative sets of material factors (M1 & M2) and resistance factors
(R1 – R4) according to EN 1997-1 are summarized in Table 7, where Set B & C are presented by A1
& A2 respectively.

Table 7 Sets of partial factors for alternative ULS design approaches

Design Approach (DA) Combination Scheme

General [Piles & Anchors]
Design Combination 1 A1 “+” M1 “+” R1
Approach 1 A2 “+” M2 [M1*] “+” R1 [R4]
Combination 2
Satisfy both: (* M2 for negative skin friction & lateral pile load)
Design Approach 2 A1 “+” M1 “+” R2
Design Approach 3 {A1(STR) or A2(GEO)} “+” M2 “+” R3

An extract from the partial action (A), material (M) and resistance (R) factors recommended in
EN 1997-1 Annex A (Normative) Partial and correlation factors for ultimate limit states and
recommended values is provided in Table 8, which also provide some clarification of the intent,
implications and differences for the alternative design approaches. Selection of the Design Approach
is an NDP for EN 1997-1.

Table 8 Partial action, material and resistance factors recommended in EN 1997-1 Annex A

Action (STR), (GEO) A1 A2

Permanent Unfavourable / Favourable 1,35/1,0 1,0/1,0
Variable Unfavourable / Favourable 1,5/0 1,3/0
Material M1 M2
Angle of shear resistance (tan φ); effective cohesion 1,0 1,25
Shear strength: Undrained, unconfined 1,0 1,4
Density 1,0 1,0
Resistance R1 R2 R3 R4
Pile (driven/bored/CF Base 1,0/1,25/1,1 1,1 1,0 1,3/1,3/1,45
Auger) Shaft (compression) 1,0 1,1 1,0 1,3
Total/combined (compression) 1,0/1,15/1,1 1,1 1,0 1,3/1,3/1,4
Spread foundation Bearing 1,0 1,4 1,0
Retaining Sliding 1,0 1,1 1,0
Equilibrium (EQU), Uplift (UPL) A
M = M2
Hydraulic heave (HYD) EQU UPL HYD
Permanent Unfavourable / Favourable 1,1/0,9 1,0/0,9 1,35/0,9
Variable Unfavourable / Favourable 1,5/0 1,5(/0) 1,5(/0)
Angle of shear resistance (tan φ), effective cohesion 1,25
Undrained shear strength, tensile pile & anchorage 1,4
resistance
Chapter 1-4 Reliability Basis of Eurocode 13

5 THEORETICAL RELIABILITY VERIFICATION IN EUROCODE

In accordance with the partial factor methods accepted in EN 1990 to EN 1999 the design
values of the basic variables, Xd and Fd, are expressed in terms of their representative values Xrep and
Frep, which may be the characteristic Xk and Fk or nominal values Xnom and Fnom. The representative
values Xrep and Frep should be divided and/or multiplied, respectively, by the appropriate partial factors
to obtain the design values Xd and Fd expressed as:
Xd = Xk/γM (11)

Fd = γF Fk (12)
where γM and γF denotes the materials and action partial factors, both in most cases > 1.

5.1 General partial factors

Both partial factors should include model uncertainties, which may significantly affect the
reliability of a structure. Design values for model uncertainties may be incorporated into the design
expressions through the partial factors γEd and γRd, applied as follows:
Ed = γ Ed E {γ gj Gkj ; γ P P; γ q1Qk 1; γ qiψ 0i Qki ; ad ...} (13)

Rd = R{ηX k / γ m ; ad ...}/ γ Rd (14)

Here η denotes a conversion factor appropriate to the material property.

Reductions in the design values of variable actions are applied as ψ0 , ψ1 or ψ2 to

simultaneously occurring accompanying variable actions. The following simplifications may be made
to expressions (13) and (14), when required.
• On the loading side (for a single action or where linearity of action effects exists) :
Ed = E {γF,iF rep,i, ad} (15)
• On the resistance simplifications of expression (14) may be applied in the relevant material-
oriented documents provided the level of reliability is not reduced.

For non-linear resistance and actions models, and multi-variable action or resistance models
the above relations become more complex.

5.2 Partial factors for material properties

The partial factor for resistance γm is defined in equation (14) by fractiles Xk and Xd. Taking
into account the general expression for fractiles of random variables the factor γm may be written as
X k µ X + u0.05 σ X 1 + u0.05 wX
γm = = = , p = Φ (−0.8β ) (16)
Xd µ X + u pσ X 1 + u p wX
where wX denotes coefficients of variation of X, u0,05 or up denotes 5%- p-fractile of the standardised
random variable having the same probability distribution as the resistance X. In case of a lognormal
distribution having the lower bound at zero, equation (16) may be written as
Chapter 1-4 Reliability Basis of Eurocode 14

1
(
exp u0.05 ln(1 + wX2 ) )
X
γm = k =
1+ wX2
≅
exp u0.05 × wX ( ) p = Φ (−0.8β )
( ) ( )
, (17)
Xd 1
exp u p ln(1 + wX2 ) exp u p × wX
1 + wX2
where u denotes the normal standardised variable.

The increase of γm with increasing the reliability index β is considerably greater for the normal
distribution than for the lognormal distribution, particularly for coefficients of variation wX greater
than 0,10. Figure 2 shows the dependence of γm on wX for normal (N), lognormal (LN, x0 = 0) and
lognormal distribution with the skewness (LN, α = 0,5) assuming β = 3,8.

1,20

γm
N

1,10

LN, x0= 0

LN, α = 0,5

1,00
0,00 wX 0,10

Figure 2 Partial factor γm versus the coefficient of variation wX for normal

and lognormal (x0 = 0; skewness α = 0,5) distributions assuming β = 3,8

5.3 Partial factors for Permanent Action

Consider a permanent load G (self-weight) having a normal distribution. It is assumed that the
characteristic value Gk is defined as the mean µG; with σG denoting the standard deviation and wG the
coefficient of variation; the sensitivity factor αG = − 0,7. The design value Gd is then:
Gd = µG − αG × β × σG = µG + 0,7× β × σG = µG(1 + 0,7× β × wG) (18)

The partial factor γG of G is given as:

γg = Gd / Gk (19)

Taking into account of the fact that Gk = µG it follows from equations (5.7) and (5.8) that
γg = (1 +0,7× β × wG) (20)

Figure 3 shows the variation of the partial factor γg with the reliability index β for selected
values of the coefficient of variation wG = 0,05; 0,10; 0,15 and 0,20.
Chapter 1-4 Reliability Basis of Eurocode 15

2,5
γg
2
wG = 0,20
0,15
1,5 0,10
0,05

0,5
0 1 2 β 3 4 5

Figure 3 Variation of γg with β for a normal distribution of G

Consider a permanent load with a coefficient of variation wG = 0,10; then γg = 1,28; if the model
uncertainty γEd = 1,05 is included, then:
γG = γg γEd = 1,28×1,05 = 1,33 ≈ 1,35
Note that the value γG = 1,35 is recommended in EN 1990, corresponding to wG = 0,1.

5.4 Partial factors for Variable Action

A similar procedure can be used for estimation of the partial factors γQ for variable loads Q.
Assuming the Gumbel distribution with mean µG and coefficient of variation wQ; the characteristic
value is usually defined as 0,98 fractile of annual extremes (or extremes related to a certain basic
reference period) and is given as
Qk = µQ (1 − wQ (0,45 + 0,78 ln(−ln(0,98)))) (21)

The design value Qd related to the design working life or other reference period is given as
Qd = µQ (1 − wQ (0,45 - αT 0,78 ln(N)+ 0,78 ln(−ln(Φ-1(−αEβ)))) (22)
where N denotes the ratio of the working design life, for example 50 years (or other reference period)
to the basic reference period for the variable action (e.g. occupancy period for imposed loads or annual
maximum for wind load), and αT (usually taken = 1) is the time sensitivity factor, given by the ratio of
w’Q/wQ , where w’Q refers to the time variable part of wQ.
Figure 4 shows the variation of γq = Qd / Qk with the coefficients of variation wQ for selected
values of β assuming a Gumbel distribution of Q, and the period ratio N = 10 (the reference period 10
times greater than the basic reference period). For a variable load which has the coefficient of variation
wQ = 0,3 the partial factor γq = 1,48. If the model uncertainty γEd = 1,05 is accepted, then it follows
that:
Chapter 1-4 Reliability Basis of Eurocode 16

γQ = γq γEd = 1,48×1,05 = 1,54 ≈ 1,5

It appears that the value γQ = 1,5, recommended in EN 1990, is a reasonable approximation.

γq β = 4,3
1.8

1.6
β = 3,8

β = 3,3
1.4

1.2

1
0 0.2 0.4 0.6 0.8 1
wQ

Figure 4 Variation of γq with the coefficients of variation wQ for selected values of β

assuming a Gumbel distribution for Q, with N =10 and αT = 1.

5.5 Combination factor ψ0

The combination rules provided in EN 1990 are based on the concept of one leading and
various accompanying variable actions. Consider two variable actions, a leading action Q1 having the
basic reference period T1 and an accompanying action Q2 having the basic reference period T2. The
characteristic value Q2k corresponds to the probability Φ(αEβ) that the extremes corresponding to the
basic reference periods (e.g. annual extremes) exceed Q2k, thus
Q2k = µ Q [1−wQ (0,45 +0,78 ln(−ln(Φ-1(−αEβ)))] (23)

The representative (reduced characteristic) value Q2,rep corresponds to the increased

probability Φ(0,4αEβ) (additional factor 0,4 reduces β) that the extreme corresponding to the greater of
the basic periods T1 and T2 exceeds Q2,rep, thus
Q2,rep = µ Q [1−wQ (0,45 +αT 0,78 ln(N)+0,78 ln(−ln(Φ-1(−0,4αEβ)))] (24)

In equation (24) N denotes the integer taken as a rounded value of the smaller from the ratios
T/T1 and T/T2. The reduction (combination) factor ψ0 is then given as
ψ0 = Q2,rep/ Q2k (25)
where Q2k and Q2,rep are given by (23) and (24). Figure 5 shows the variation of ψ0 with the
coefficient of variation related to the reference period T, for β=3,8, for N = 7 and selected time
sensitivity factors αT.
Chapter 1-4 Reliability Basis of Eurocode 17

ψ0
0.8

αT= 0,5
0.6
αT= 0,6
0.4
αT= 0,7
0.2
αT= 1,0

0 . .
0 0.2 0.4 0.6 0.8
wQ

Figure 5 Variation of ψ0 with wQ related to the reference period T; for β=3,8; N = 7 and
selected time sensitivity factors αT

Consider a combination of an imposed load Q1 having the basic reference period T1 = 7 years
and an accompanying action (wind) Q2 having the basic reference period T2 = 1 year. Consider a time
sensitivity factor αT = 0,5. Considering the reference period T = 50 years then N = 50/7≈7. It follows
from Figure 5 that for β=3,8, wQ=0,35 the factor ψ0 ≈ 0,46 (for αT = 1, ψ0 is less than 0,4).

5.6 Combination factor ψ1 and ψ2

Factors ψ1 and ψ2 are defined in such a way that the corresponding representative Qrep = ψ1Qk or ψ2Qk
are exceeded with the probability 0,01 or 0,50 respectively. Assuming Gumbel distribution the
reduction factors may be determined as
η 1( 2 )
1 − w{0 , 45 + 0,78 ln[ − ln( Φ (1 − ))]}
q (26)
ψ 1( 2 ) =
1 − w{0, 45 + 0,78 ln[ − ln( Φ ( 0,98 ))]}
where η1(2) is the required probability (0,01 or 0,50 respectively) and q is the probability of variable
action Q being non-zero
Consider the following informative examples of frequent and quasi-permanent values of
different variable actions Q. Recall that the probability ρ = 1−η/q, where η is a fraction of the reference
period (0,01 or 0,5) during which Q1 or Q2 are exceeded, q denotes the probability of Q being non
zero.
• Imposed load: It is assumed that the coefficient of variation wQ = 1,1 (approximates a point in time
distribution), short term imposed load is supposed to be on about 18 days a year, then ρ1 = 1− η1/q1
=1−0,01/0,05 ~ 0,8 and it follows from Figure 5 that ψ1 ~ 0,5, long term load is almost always on,
then ρ2 =1− η2/q2 = 1−0,5/1 ~ 0,5 and it follows from Figure 5 that ψ2 ~ 0,2,
• Wind load: The assumed coefficient of variation is wQ = 0,5, the load is on about 10×8 hours a year
ρ1=1− η1/q1 =1−0,01/0,009 < 0 (not applicable), ψ1 ~ 0,0, for quasi-permanent value ρ2 = 1−
η2/q2=1−0,5/0,009 < 0, (not applicable) and ψ2 ~ 0,0,
Chapter 1-4 Reliability Basis of Eurocode 18

Based on these results factors ψ0, ψ1 and ψ2 recommended in EN 1990 (see §4.1) seem to be
rather conservative and should be revised taking into account local conditions.

6 CONCLUSIONS

The following conclusions are made on the basis of structural design applied in Eurocode
through the head Standard EN 1990-2002 in terms of its relevance to the revision of SANS 10160 in
particular and South African structural design standards in general:
• International harmonisation: The consistency between EN 1990 and ISO 2394 provides a sound
basis for harmonisation by referencing to the Eurocode Standard, not only to European Member
States, but also to international practice based on the International Standard.
• Unified design: The extensive development of reliability based partial factor Limit States Design
in EN 1990 as the basis for unification with the treatment of actions on structures and structural
resistance through EN 1991 – EN 1999 (Sedlacek et al 1996) provides the platform for the
consistent treatment of actions in the present revision of SANS 10160 and the application of
unified procedures with the future revision of materials-based design standards.
• Reliability framework: The extended reliability framework set up in EN 1990 provides the basis
for an extended scope of design situations and actions that can be provided for, whilst maintaining
consistent levels of reliability based on rational analysis and calibration; specifically providing
operational procedures for transient (execution of structures); persistent (equilibrium, strength,
geotechnical) and accidental (robustness, seismic) situations.
• Parameter selection: Although the selection of Nationally Determined Parameters are constrained
and managed, the options are sufficiently wide to accommodate present South African practice
whilst maintaining effective consistency with Eurocode in a revised SANS 10160. It also provides
the basis for improvement of the consistency of reliability (SAKO 1999; BRE 2003; Holický &
Retief 2005).

It is therefore concluded that Eurocode EN 1990-2002 Basis of structural design represents a

significant advancement in presenting rational structural design procedures, that could serve as useful
reference in the revision of the present South African Standard SANS 10160-1989 (Amended).

7 REFERENCES

BRE (2003) UK BRE Review of SAKO Report. Gulvanessian H – personal communication.

Brozzetti J, Sedlacek G (2000). Background to the combination rules (6.10) or (6.10a) and
(6.10b) in prEN 1990 Basis of design. Document CEN/TC 250 N419. CEN/TC 250 Secretariat.
CE (2002). European Commission Guidance Paper L Application and use of Eurocodes.
Document CONSTRUCT 01/483 Rev.1. Brussels.
CEN (2002). European Committee for Standardisation EN 1990:2002 Eurocode – Basis of
structural design. Brussels.
Gulvanessian H, Calgaro J-A & Holický M (2002). Designer’s guide to EN 1990 Eurocode:
Basis of structural design. Thomas Telford. 192 pp.
Holický M & Retief JV (2005). Reliability Assessment of Alternative Load Combination
Schemes for Structural Design. Journal SAICE, Vol. 25, No 1, pp 15 – 20, 2005.
Holický M & Vrouwenvelder T (2005). Basic concepts of structural reliability. Chapter I in
Implementation of Eurocodes Handbook 2 Basis of structural reliability and risk engineering.
Leonardo Da Vinci Pilot Project.
ISO (1993). International Standards Organisation. ISO 3534 Statistics- Vocabulary and
Symbols: Part 1: Probability and general statistical terms, Part 2: Statistical quality control.
ISO (1997). International Standards Organisation. ISO 12491 Statistical methods for
quality control of building materials and components. 1997.
Chapter 1-4 Reliability Basis of Eurocode 19

ISO (1998). International Standards Organisation. ISO 2394:1998 General Principles

on Reliability for Structures. International Standard. [Also issued as SANS 2394:2004]
JCSS (2001). Joint Committee on Structural Safety. Probabilistic model code.
http://www.jcss.ethz.ch/.
SABS (1994). SA Bureau of Standards, Pretoria. Code of Practice: The general procedures
and loadings to be adopted in the design of buildings. SABS 0160:1989 (Amended 1990, 1991, 1994).
SAKO (1999). Joint Committee of NKB and INSTA-B. Basis of Design of Structures.
Proposal for Modification of Partial Safety Factors in Eurocodes. NKB Committee and Work Report
1999:01 E, pp. 55.
Sedlacek G, Brozzetti J, Hanswille G, Litzner HU, Casinius TDG (1996). Relationship
between Eurocode 1 and the “material” oriented Eurocodes. IABSE Colloquium Background and
application of Eurocode 1. Delft.