Modelling cyclic degradation of bridge R. C. columns subjected to concurrently seismic events

Lower magnitude events typically are concomitant with seismic events, thus cumulative damage in structures has become an increasing concern in the structural engineering community. Currently, structural non linear models are capable of considering cyclic degradation, however there are several difficulties to perform the calibration of those parameters. In this research, the application of Approximate Bayesian Computation algorithms is explored to perform the inference of cyclic cumulative degradation parameters of a model constructed in OpenSees software, using laboratory tests results as a dataset. The data has been taken from a scaled reinforced concrete column subjected to a dynamic load, performed in PUCP laboratory. Two models are presented in this study: (1) a blind prediction using nominal model parameters for concrete columns and (2) a simplified model that accounts for possible non-measured lateral slip; both models do not include degradation to latter demonstrate the importance of this matter. Results show that the simplified model, calibrated using an Approximate Bayesian Computation (ABC) technique, better predicts the test deformation records; also, the strengths of the method for the inference of those parameters are displayed, as they give information to allow a probabilistic seismic assessment of a reinforced concrete column that includes the modelling uncertainties. Later on, this model will be used to evaluate the effects of cumulative damage on structural design criteria.