Modeling of random static loads on a structural cover with limited statistical data

Abstract

Relevance. Loads on structures are complex stochastic elements that include several types of uncertainties simultaneously. The article describes a probabilistic approach to the load modeling on structural covers taking into account limited statistical data, when the parameters of distribution functions are presented in an interval form. The aim of the work is development of an approach to modeling the probabilistic distribution of random load on the structural surface in conditions of limited (incomplete) statistical information about the design load. Methods. The probability distribution of a particular type of loading is represented as p-boxes (probability boxes). A numerical example shows an algorithm for determining a p-box consisting of a sum of p-boxes that characterize different loads with different boundary distribution functions. Results. Based on the proposed approach, it is possible to define the intervals of normative and design loads with a given confidence level, to estimate the failure probability of structural elements, to assess the risk of an accident and also to make selection for structural element cross-section at the target level of reliability.