A Bayesian regression framework for concrete creep prediction improvement: application to Eurocode 2 model

Concrete is the most widely used material in the construction industry due to its strength, workability and durability. But under a sustained load, concrete is prone to creep causing excessive long-term deflection of structural members, cracks in tensile members, redistribution of stresses over time in composite structures, and loss of prestressing force in prestressed concrete elements. Therefore, structural engineers must accurately predict the concrete creep over the long-term. The concrete creep coefficient is an important entry in many calculations and analyses of reinforced concrete structures. Currently, many models such as the Eurocode 2 model have predicted the concrete creep coefficient. Based on a large database for creep tests, this study aims to improve the prediction of the Eurocode 2 creep coefficient model at long-term by implementing correction coefficients into the model. Since Bayesian-type inferences are suitable tools for revising and updating design codes, the correction coefficients are calculated based on Bayesian linear regression. Statistical indicators demonstrate the accuracy and effectiveness of the proposed improvement and modification.