Probabilistic representation of duration of load effects in timber structures.

Staffan Svensson

Research output: Contribution to journalJournal articleResearchpeer-review

Abstract

Reliability analysis of structures for the purpose of code calibration or reliability verification of specific structures requires that the relevant failure modes are represented and analyzed. For structural timber, sustaining a life load, two failure cases for each failure mode have to be considered. These two cases are maximum load level exceeding load-carrying capacity and damage accumulation (caused by the load and its duration) leading to failure. The effect of both load intensity and load duration on the capacity of timber has been an area of large interest over the last decades; the effect of load duration is in general referred as duration of load (DOL) effect. Several research projects did address the DOL effect in the past and a number of different damage models, based on experimental evidence, have been formulated to describe the phenomenon. In the present paper, three different models for the representation of damage accumulation in timber are considered. The parameters of these models are calibrated by the use of test data whereas uncertainties associated with the model formulation and limited sample size are considered. The DOL effect is usually taken into account in code-based design of timber structures in terms of a modification factor View the MathML source which is multiplied on the short-term resistance of the timber material. The scenario of a beam subject to office space life loads is analyzed and the modification factor View the MathML source is calibrated by using the three different damage models.
Original languageEnglish
JournalEngineering Structures
Volume33
Issue number2
Pages (from-to)462-467
ISSN0141-0296
DOIs
Publication statusPublished - 2011

Fingerprint Dive into the research topics of 'Probabilistic representation of duration of load effects in timber structures.'. Together they form a unique fingerprint.

Cite this