FATIGUE RELIABILITY OF STRUCTURES BASED ON PROBABILITY AND POSSIBILITY MEASURES

Abstract

The randomness of fatigue life prediction has been commonly recognized, but recently researchers have begun to recognize the fuzziness of fatigue failure descriptions resulting from the lack of distinct definitions and precise information. In this paper, both probability- and possibility-based reliabilities of components subjected to fatigue loading are investigated. In the probability-based reliability, the material properties and the amplitude of a loading history are assumed to be random. In the possibility-based method, the fuzzy linear regression method is employed to account for the fuzziness. This work employs S-functions to define fuzzy limit state functions. Failure probabilities are calculated in terms of probability distributions of fuzzy damage and fuzzy limit state functions. Both probability- and possibility-based reliabilities of a portal frame are examined. Accordingly, the reliability of systems in the presence of fuzziness is calculated and compared to the results obtained from the probability-based method.