Seismic Behaviour of R. C. Walls

Research focus

The Peer review has evaluated this group as Average

The participants in this research group have a strong background on modelling of RC structures in the non linear range. At the end of 2002 the participants entered into an International Research Project, putting together their knowledge in the field and thus gaining a strong advantage from this joined effort. The research work stemmed from the consideration that in the last decades the seismic codes in many countries have experienced a rapid evolution, introducing both non linear procedures of analysis and new design concepts. As a first consequence, reliable and robust modelling strategies in the non linear range must be devised which can be adopted by the design engineers; their performance must be investigated in connection to that of the non linear analysis procedures in whichthey are inserted. Secondly, it is necessary to assess the performance of existing structures, designed according to seismic provisions that have been judged insufficient. The two problems have been faced with reference to a widely spread class of structures, the RC shear walls, adopting as a term of comparison the results of a series of experimental shaking table tests on the specimen, named Camus I, of a 5 story lightly reinforced shear wall, 1/3 scale. Three different modelling scales have been investigated, with decreasing levels of accuracy and complexity. At the highest level, the micro-scale of the finite elements (FE), an ad-hoc developed computer code (EF2002) has been upgraded on account of the most recent works in the literature, to better reproduce the shear dominated behaviour of structural elements. The code tracks the non linear behaviour of the materials, both concrete and steel, at a local level. Modelling shear behaviour relies on the choice of adequate failure criteria for concrete in-plane stress and on the kinematics of the FE formulation. For this reason the FE model (a) uses a carefully formulated 2D strength envelope model for the tension–compression stress states; (b) adopts higher-order elements for the concrete component in order to avoid shear locking and to capture non-uniform strain distributions; and (c) includes concrete-reinforcement interface elements, since the bond-related mechanisms play an important role in shear failure. At an intermediate level, the meso-scale of a fibre element based on Timoshenko beam kinematics has been adopted. The fibre model has been developed specifically to introduce the shear effects at a global level of discretization. To this purpose the truss and arch mechanisms of shear transfer are taken into consideration by two mechanical formulations, coupled with the bending response. This represents an innovative choice, since fibre elements usually consider either the bending behaviour only or uncoupled bending and shear behaviour. The modelling of the constitutive behaviour of materials, in terms of stress-strain relations, is specially conceived to reproduce the global structural behaviour through non linear dynamic analyses. At the lowest level the macro scale of a spread plasticity model has been adopted, where the material behaviour is described in terms of sectional properties (moment-curvatures). In the model, which primarily describes a flexural behaviour, the degradation effects in terms of stiffness due to shear have been introduced through the modification of the constitutive relations governing the behaviour of the plastic hinges. The approaches at the meso and macro scale are appealing for the possibility of modelling a whole structure with few elements; in addition the results in terms of internal forces and curvatures, familiar to designers, are provided by computationally light analyses. The micro and meso scales of structural modelling have been adopted inside the framework of the displacement based assessment and of the pushover analysis; the meso and macro scales have been used for the non linear dynamic analyses. The effectiveness of the proposed modelling approaches within the quoted non linear analysis procedures is evaluated in reproducing the experimental behaviour of the CAMUS I shear wall. The wall design aimed to yield the reinforcement with the formation of horizon137 tal cracks at each floor, according to the multi-fuse principle; in the tests the structure did not fully behave according to this design philosophy, showing shear cracking and failure, anchorage pull-out and the breaking of some bars. The wall is representative of structures where the position of critical zones and failure mode cannot be assumed a priori. These issues highlight the need for models capable of predicting, together with bending, the shear-related phenomena and failure. The multiscale modelling approach adopted provided a comprehensive representation of the structural behaviour, capturing with different details the global and local behaviour of RC elements in plane state of stress.

Dipartimento di afferenza

Dipartimento di Ingegneria Strutturale (DIS)

Docenti afferenti

Associate Professors
Maria Gabriella Mulas
Assistant Professors
Dario Coronelli
Luca Martinelli