Reliability Assessment of Multi—Member Structures: Fid-Bau Portal

Reliability Assessment of Multi—Member Structures

Rackwitz, R.

Abstract Traditionally structural safety considerations are based on checking safety at the component level. Only implicitly, adjustments of the required safety margins with respect to the system reliability aspect have been made in some cases. For example, in stability problems in soil mechanics the representative values of the soil properties are taken close to their means in order to take account of the averaging of the property fluctuations along the yield surfaces the sections of which are to be viewed as a parallel system. However, that almost all structures involve a system reliability aspect must be considered as common knowledge. It is recognized that some systems behave like series systems where system collapse is implied by the failure of a component in any of its failure modes. Such systems are correctly categorized as especially critical systems. It is also recognized that many civil engineering structures are redundant structures in the sense that given the (first) failure(s) of some component(s), element(s), joint(s), cross section(s), etc., the structural system not necessarily fails, too. Without such inherent redundancy many of damaged structures would actually have failed. Thus, redundancy has been identified as a desirable feature of a structure although it usually does not mean that the user will leave a structural component unrepaired once it failed. Vague notions like robustness have been introduced as requirements for structural design meaning essentially that after the failure of one component in an extreme loading event the structure still maintains its primary function though with somewhat reduced capability. And it is not surprising that already in the very early deterministic works on the subject the ductility of structural components has been identified as one of the dominating factors when assessing system reliability.