A method for structural reliability analysis of trusses with interval uncertainty of random variables

Abstract

The article describes a method for the reliability analysis of hinge-rod systems (trusses) with an interval estimation of the variability limits of random variables. In practical tasks of structural reliability analysis, it is often not possible to obtain complete statistical data that is necessary for a probabilistic and statistical approach in accordance with GOST 27751-2014. The use of alternative theories of data analysis makes it possible to obtain a define assessment of reliability with incomplete statistical data without introducing unconfirmed hypotheses about the form and shape of probability distribution functions of random variables. In this paper, it is proposed to use data on the limits of variability of random variables (an interval analysis) to assess the reliability of hinge-rod systems on the truss example. A comparative analysis of the truss reliability is given by two approaches (the classical probabilistic-statistical and the proposed interval approach). By the interval approach, the reliability of the truss is lower due to the fact that the interval estimation takes into account other possible distributions within a given interval that differ from the normal probability distribution. In the case when it is impossible to determine the type of probability distribution, but the limits of variability for random variables are known, the use of the presented method is a more cautious for the safety of truss operation.