Statistical procedures to quantify the influence of quality control in component properties: A methodology for structural safety assessments

Abstract In structural safety assessments, the selection of suitable parameters for stochastic models of component properties — such as concrete compressive strength or geometric dimensions — is a prime requirement. During their production, several quality control strategies are adopted to verify the specified value of diverse properties of structural members. Bayesian statistics form a suitable basis for updating the statistical distribution parameters when input from quality control is available. Previous work in the field of concrete compressive strength has demonstrated that such strategies introduce a filtering effect on the stochastic description of concrete strength distribution, which may have a positive influence on the resulting safety level. However, a generic approach applicable to a wider scope of properties remains to be established. This paper offers a methodology to update the distributional parameters based on input from quality control and further integration in safety assessments. By considering the filtering effect of quality control on the distributions of material properties, the proposed methodology aims to assist practitioners to unlock existing safety margins in structural design and, thus, propose more material and resource efficient solutions.