Abstract
One of the promising directions of development of structural design standards is the transition to full probabilistic methods of structural design for a given target reliability level. For estimation of the reliability index, effective accounting and modeling of aleatory and epistemic uncertainties of data is necessary. The article describes a probabilistic approach to the design of steel truss elements taking into account incomplete statistical data. This approach allows giving an estimate of the failure probability for a steel truss under uncertainty data. The failure probability presented in the form of an interval that will narrow as additional statistical data about loads, material properties, geometric imperfections, etc. grows. To solve this problem, the article uses two numerical approaches to calculate the failure probability: discretization of p-boxes into Dempster-Shafer structures and the interval Monte Carlo (IMC) method. The use of the presented approach in practice also allows performing a comparative evaluation of different technical and economic solutions of steel trusses based on the reliability factor.