Evaluating the design value of the bearing capacity of steel elements designed using numerical models

Abstract

Introduction. Using numerical models to analyze the behavior of complex or new structural solutions becomes increasingly popular. New software can be used by a beginner to easily create numerical models of structures, and this is, on the one hand, an undeniable advantage, which, on the other hand, raises concerns about the accuracy and reliability of the results to be obtained. It is noteworthy that the regulatory engineering framework ignores this area. Moreover, research and design communities lack any uniform approaches to modeling, and, more importantly, to interpreting results and ensuring the structural reliability of solutions. Materials and methods. The article proposes a method for analyzing design values of the bearing capacity designated for target reliability levels, taking into account the changeability of basic variables and the modeling error. This method was developed using the Bayesian approach to quantile prediction provided that the number of validation results was limited. Results. The article presents the implementation of the proposed method of analyzing the design value of the bearing capacity using the results of FEM (finite element method) modeling of the bearing capacity of corrugated steel beams. The influence of the assumption about the standard deviation of the modeling error is analyzed. Conclusions. The work presents a method for determining the design values of the bearing capacity for the target levels of reliability, taking into account the changeability of basic variables and the modeling error. The factors, having a great impact on evaluating the design value of the bearing capacity and deserving further research, are substantiated. First, it is necessary to draw attention to the justification and regulation of target levels of structural reliability in regulations. Second, it is necessary to draw attention to studying statistical parameters of the modeling error and developing recommendations about the designation of apriori statistical data and maximum evaluations in respect of the standard deviation of the modeling error. Thirdly, attention must be drawn to development of criteria and formats for checking limit states in the course of design based on numerical models of the bearing capacity.