Assessing damage and collapse capacity of reinforced concrete structures using the gradient inelastic beam element formulation

Abstract In the recent years, there have been growing efforts to simulate the response of reinforced concrete (RC) framed structures in the post-peak range and until collapse using distributed-plasticity beam-column elements, as opposed to concentrated-plasticity elements. Concentrated-plasticity elements require component-level testing data for their calibration, which can be scarce, whereas distributed-plasticity elements only require material testing data, which are abundant, thereby making them much more attractive. However, in the presence of softening constitutive relations, distributed-plasticity elements suffer from strain localization, which causes loss of response objectivity and convergence failures of the member-level solution algorithms. To address these deficiencies, the authors recently formulated the gradient inelastic (GI) beam theory and developed the GI force-based (FB) element formulation. The capabilities of this new element formulation are herein compared with those of modeling strategies employing existing displacement-based (DB) and FB distributed-plasticity frame elements commonly used to model softening systems. First, comparisons are performed against RC column test data that include not only force vs. displacement curves, but also curvature distributions. Subsequently, predictions of damage and collapse of RC framed structures obtained by the modeling strategies using different distributed-plasticity elements are investigated. Two RC moment frames of two and four stories and a two-span RC bridge with a single-column pier are simulated by the aforementioned modeling approaches, under static pushover loading and earthquake ground motions, including Incremental Dynamic Analyses (IDAs). Comparisons with experimental data demonstrate the capability of the GI FB element to more accurately and realistically capture the curvature distributions over the length of RC members, while comparisons of the IDA results show that models employing the GI element formulation predict lower collapse capacities by about 15% and greater damage.