The concrete-steel composite slabs show a complex structural characterization due to the different behaviours at the two materials. The materials are exposed to different deformations, large deflections and complex stresses with still a limited understanding of their micromechanics. Hence, current building codes rely on expensive and tedious laboratory tests that characterize the composite slab failure and the ultimate resistance. The Finite Element (FE) numerical simulations were introduced more than 25 years ago in composite slab studies as a mechanism to validate new design methods and also as an alternative to reduce laboratory tests requirements. However, the simulations historically observed a significant number of simplifications such as reduced scale models or simplified geometries.
This dissertation introduces initially a novel modeling and simulation methodology that enables new insights in the steel deck and concrete slab response for bending. Distinct full-scale finite element models were generated for four commercial steel deck profiles to simulate the laboratory tests. An intense and systematic optimization process was carried out as the computational costs and the simulation files size associated with the initial FE models were significant. The three-dimensional composite models detailed embossment depth and slope, steel thickness, or tilting angle, among several others. Common limitations and simplifications related to steel-concrete contact, adhesion, and cohesion factors in previous research efforts were addressed. Newton-Raphson was the simulation method and enabled the consideration of geometrical and materials nonlinearities. The proposed methodology was validated by comparison of the results from the bending simulations with the actual maximum loads, midspan deflection and end slip values obtained from laboratory bending tests. Based on the robustness of the bending simulations, parametric and boundary conditions analyses were performed through pull-out simulations.
Micromechanics phenomena that could not be observed during laboratory tests were investigated at the full-scale bending simulations. First, the neutral axes and vertical disconnection representations for the steel deck and concrete slab were characterized and subsequently they proved the existence of partial connection between the materials. Second, a new normal vertical tension parameter sshear was introduced to describe the vertical stresses at the steel deck and the concrete slab. Third, the longitudinal shear strength Zu was computed for different midspan deflections, loads and friction coefficients. The longitudinal shear failure is the most common failure phenomenon among open rib steel deck profiles and thus multiple studies were performed. The observation of a constant Zu value at the shear span of the bending test was novel and indicated that the Eurocode 4 Partial Connection Method was not capable to describe the complex longitudinal shear strength behaviour observed from the simulations.
The dissertation concludes with the introduction of a new characterization parameter tu,mechanical to assess the composite slab design efficiency. The parameter is defined as the longitudinal shear strength tu computed from the simulations for a null friction coefficient. The new parameter proved to accurately characterize the performance of the different composite slabs studied in this dissertation when compared with the maximum loads from the laboratory tests.
The combiation of the novel modeling and simulation methodology with the tu,mechanical computation enabled a new design process for steel deck profiles. The process developed an iterative computer-focused approach with the goal to reduce the reliance in the costly and tedious laboratory tests.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados