LIMIT STATES DESIGN AND STRUCTURAL RELIABILITY: THE CANADIAN CONTEXT

Abstract eng:
This paper discusses reliability-based design procedures in Canada. The National Building Code of Canada adopted, in the 1970’s, a policy of having all design Codes and guidelines follow a “Limit States Design” format, with partial load and resistance factors . The intention was that, as these guidelines were developed, the older and customary “allowable stress design procedures” were to be phased out. The limit states design provisions remained prescriptive, but introduced new design equations based on characteristic demands (linked to specific return periods) and capacities (based on a lower percentile). The characteristic demands and capacities were to be modified, respectively, by “load factors” and “resistance factors”. Combination load factors were also given for the case of several demands acting together. This format is essentially identical to the one currently adopted for design in many countries. The load and resistance factors, controlling the achieved safety or reliability, had to be calibrated so that the resulting design achieved a minimum target reliability. This calibration process has not been completed, neither in Canada nor in other countries. The paper discusses why this approach, although it uses reliability theory for the calibration, is not a full implementation of that theory. One of the stumbling blocks in the calibration has been the definition of a “minimum target reliability”. This paper discusses this issue with examples in the Canadian context. In addition, some design procedures (notably, for earthquake and geotechnical engineering) have lagged behind in the development of a limit states design formulation. The calibration exercise relies on being able to obtain a reasonable estimate of the reliability level achieved with a given design, and this depends on reliable models for both demands and capacities. In many instances, the development of such models forms one of the greater stumbling blocks, particularly when the prediction of demand or capacity involves complex nonlinear behavior. On the other hand, although forty years ago there was still much confusion about how to estimate reliability in general problems, today the field of reliability theory and associated software has developed to an extent that has outpaced the actual applications. The paper also discusses this issue in the Canadian context, along with the advantages or disadvantages that have resulted from the switch to “Limit States Design”. Finally, the paper discusses instances where reliability theory has been fully implemented, and the upcoming emphasis in “performance-based design”.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket