Finite element analysis for predicting the short-term and long-term behaviour of timber-concrete composite structures
- Publication Type:
- Issue Date:
NO FULL TEXT AVAILABLE. Access is restricted indefinitely. ----- This research focuses on developing finite element methods for predicting the short-term and long-term behaviour of timber and timber-concrete composite (TCC) beams. Using experimental investigations is sometimes expensive and time consuming for research programs such as long-term behaviour of TCC beams and therefore, developing efficient analytical models is of great importance in this regard. Further, the existing FE models can be typically classified as continuum-based and discrete frame elements which have different domains of applicability. Offering the versatility required for detailed analysis and modelling of the local effects is the main advantage of continuum-based FE models; however, they are not as efficient as frame FE models from a computational point of view. The frame models are a good compromise between accuracy and efficiency and this study concentrates on the development of a computationally efficient, discrete element for predicting short-term and long-term response of TCC members with reasonable accuracy. Accordingly a 1D frame FE model cast in the framework of the force-based formulation in conjunction with lumped connections is developed in this study for non-linear analysis of TCC beams under service and ultimate loads. Timber is an anisotropic material from a computational point of view, with different behaviours along the grain and perpendicular to grain directions. Various constitutive laws in 2D and 3D frameworks have been developed in literature; however, most of them are too complex with many input parameters which should be calibrated for different types of timber. Accordingly two novel strategies for modelling timber as an orthotropic material in the context of 2D continuum-based FE model are proposed namely: Hashin damage criteria; and a layered approach with Hypoelastic constitutive law. Further, a simple strategy for modelling TCC connections is employed in which the connector is modelled with non-linear spring and its shear-slip response can be obtained from existing analytical models and/or push-out tests. The proposed approach provides a sensible basis for comparing the results of 1D frame and 2D continuum-based FE models with available experimental results and it is concluded that the developed 1D frame FE model has adequate accuracy for modelling of TCC beams under short-term loads. Furthermore, a finite difference scheme with an implicit approach is developed to analyse diffusion within the timber section due to variation of relative humidity in the ambient condition. The time-dependent behaviour of TCC components (i.e. timber, concrete and connection) in the framework of a total secant approach is readily derived. The derived formulation in conjunction with the FD method is implemented in the developed 1D frame FE model to capture the long-term behaviour of TCC beams. The accuracy of the developed 1D model is validated by different experimentally measured data taken from literature. The main feature of this frame model is its simplicity and superior efficiency compared with continuum-based FE models for capturing the short-term and long-term behaviour of timber and TCC beams, particularly for design oriented parametric studies and sensitivity analyses. Finally, a comprehensive parametric study on different aspects of TCC beams is conducted to investigate short-term and long-term behaviour with different geometries, material properties, types of connections and ambient conditions. Various conclusions are drawn and some ranges for long-term input parameters of LVL timber beams are suggested.
Please use this identifier to cite or link to this item: