Partial safety factor for resistance model uncertainties in 2D non-linear finite element analysis of reinforced concrete structures

This work evaluates the partial safety factor related to the resistance model uncertainties in non-linear finite element analyses (NLFEAs) for reinforced concrete structures. Various experimental tests concerning different typologies of structures with different behaviours and failure modes, i.e., walls, deep beams, panels, are simulated by means of appropriate two-dimensional finite elements (FE) structural models (i.e., plane stress configuration). Several FE structural models are defined for each experimental test to investigate the model uncertainty influence on the 2D NLFEAs of reinforced concrete structures in terms of global resistance, considering different modelling hypotheses to describe the mechanical behaviour of reinforced concrete members (i.e., epistemic uncertainties). Subsequently, the numerical results are compared to the experimental outcomes. Then, a consistent treatment of the resistance model uncertainties is proposed following a Bayesian approach. Specifically, the prior distributions of the resistance model uncertainties for the different modelling hypotheses are evaluated and then each one is updated on the basis of the data obtained from the other models to evaluate the posterior distributions. After that, the mean value and the coefficient of variation characterizing the resistance model uncertainties are identified. Finally, in agreement with the safety formats for NLFEAs of reinforced concrete structures, the partial safety factor related to the resistance model uncertainties is evaluated.