Modelos de comportamiento último y de plastificación de secciones de hormigón ajustados con ensayos cumpliendo requerimientos constructivos y sísmicos

Zusammenfassung

The prediction and simulation of the seismic behaviour of reinforced concrete structural elements is considered an issue of great importance, due to the need to accurately know the effects caused by seismic loads when acting upon a reinforced concrete structure, from a social and economic point of view. Thus, obtaining certain concepts involved in the yielding and ultimate behaviour of the cross section (yielding chord rotation, yielding curvature, yielding moment and ultimate chord rotation) is essential in the development of a realistic moment-curvature diagram, as a previous step in order to describe the hysteretic behaviour of the plastic hinges generated in the structure by seismic loads. Since the seismic simulation of reinforced concrete structures is very time-consuming, numerical models need to be as effective as possible from the point of view of computing time, whilst maintaining an acceptable level of accuracy. Following this principle, the main objective of this study is to develop simple models which are able to describe the yielding and ultimate behaviour of a reinforced concrete cross section. Another essential requirement for a computational model to be able to carry out realistic numerical simulations properly is a correct calibration with appropriate data. Thus, a selection of tests has been made from a database of more than 1000 tests, complying with several constructional and seismic requirements imposed by EC2, EC8 and ACI-318 for some parameters (dimensions of cross section, mechanical properties of materials, etc.) of reinforced concrete sections in order to obtain only structural elements used in habitual reinforced concrete buildings. When numerical models become complex (high number of parameters, non-linear behaviour, etc.), the use of optimization tools is clearly justified in order to develop more reliable and realistic models by calibrating them with experimental results. The use of these techniques is becoming very common, not only in structural engineering but also in many other fields of engineering, due to their robustness to converge towards satisfactory results in an acceptable computing time. In this context, Genetic Algorithms are used in the present work to improve some expressions previously obtained by other authors, through calibrating these expressions using the aforementioned selection of tests.