Euro-SiBRAM’2002 Prague, June 24 to 26, 2002, Czech Republic

Session 4 – Comment

Reference Value vs. codes and design approaches

Prof. Ing. Břetislav Teplý, CSc.

BUT Brno, Žižkova 17, 66237 Brno, Czech Republic

teply.b@fce.vutbr.cz

Abstract

The determination of reference value in the context of limit states and present and future codes of praxis is discussed.

Key words: codes, design, limit states, reference value, reliability index.

Introduction

The goal of structural design is to ensure appropriate safety, serviceability, durability and economy of  structures – this requires a certain optimisation. The requirements concerning safety are usually expressed in terms of the accepted minimum reliability index or the accepted maximum failure probability with respect to some adverse events – a limit state.

Within the concept of limit states (LS), the reference value (RV) is a necessary measure utilized in the probabilistic assessment of structures and it is implicitly given in codes. Unfortunately, the exact and unique definition is lacking and some questions may arise. The RV might be a deterministic value or a random variable - (see [1]). While LS characterizes the possible failure  in a quantitative way, reference quantity characterizes the resistance of the structure in a measurable way!

Note that the probabilistic approach was mentioned in codes as early as in 1988 in Czechoslovak national standard [2]. At present, the Euro code documents are being introduced and the situation prevails as in EC 1  [3], paragraph 3.5 reads: “As an alternative, a design directly based on probabilistic methods may be used” (showing some details in ANNEX C). Naturally, the Probabilistic Model Code [4] worked out by the Joint Committee for Structural Safety (JCSS) deals with such an approach too. Let us focus our attention on the Euro code approach: to determine an appropriate reference value one has to decide about or to select relevant

(i)                  design situation and limit state

(ii)                reliability class

(iii)               definition of LS – and, accordingly the numerical value of RV !

(iv)              reliability level (target value of reliability index or failure probability)

(v)                reference time period.

It can be seen that the proper determination of RV is not an easy task. The following table shows this situation on an example of the Ultimate Limit State.

Note: PeBD is the abbreviation of the Performance-Based Design approach, which appeared recently as a new trend and a part of the project PeBBu – see [5]. It should bring the innovation in the building process as the prescriptive criteria in current regulations (e.g. Euro codes) proved to be a barrier for the cost optimization and trade.

Now, to the question marks shown in the table:

-         it is certainly not reasonable to use the same value of target reliability index for so different LS definitions (and relevant RV´s) as are e.g. the onset of yielding and the mechanism formation. The consequences would differ considerably and so the same should hold for the reliability level;

-         in future codes (PeBD) the definition of LS (and of relevant RV) will probably be more open to the investor’s wishes

-         the consideration of durability or dynamic loading might bring some more uncertainty regarding the RV determination.

Acknowledgement

This paper was prepared as a part of the activities in project 103/00/0603 backed by GA CR.

References

[1] Jaroslav Menčík: Reference Values (Safety, Serviceability and Durability). Proceedings of Colloquium SiBRAM´2002, Vol 1, pp. 38-39.

[2] ČSN 730031 Reliability of Structures. Basis of Design, Czechoslovak code, 1988 (in Czech).

[3] Euro code – Basis of structural design, prEN 1990:2001.

[4] www.jcss.ethz.ch

[5] www.cibworld.nl