Simulación del comportamiento mecánico de la madera en uniones estructurales y su aplicación mediante modelos tridimensionales de elementos finitos

  1. Iraola Sáenz, Borja Iñaki
Dirigée par:
  1. José Manuel Cabrero Ballarín Directeur
  2. Beatriz Gil Rodríguez Directrice

Université de défendre: Universidad de Navarra

Fecha de defensa: 22 juillet 2016

Jury:
  1. Antonio Martín Meizoso President
  2. Mikel Landa Esparza Secrétaire
  3. Gustavo Adolfo Vargas Silva Rapporteur
  4. Nuno Dourado Rapporteur
  5. Santiago Sánchez Beitia Rapporteur
Département:
  1. (ETSA) Construcción, Instalaciones y Estructuras

Type: Thèses

Teseo: 121763 DIALNET lock_openDadun editor

Résumé

Wood is an orthotropic material, with different elasticity in its three primary directions. Its outstanding load bearing capacity and mass relation is used in order to obtain efficient structural elements. However, the efficiency of these structures is limited, amongst other reasons, by the notable safety margins that these structures require for the difficulties in the study of such a complex ma­ terial. This research reviews the morphological and mechanical properties of wood with the aim of understanding its complexity. The usually applied simplifications to model it by menas of the Finite Element method are also presented in this work. Aiming at a reduction in the applied simplifications, and at obtaining a research tool that incorporates the actual characteristics of wood, an algorithm is developed that reproduces the mechanical behavior of wood accounting for different behavior in each of the three directions, and tension and compression. Moreover, the algorithm enables the implementation of different failure criteria for each tensional state to improve failure detection. The algorithm is designed to detect the failure mode that triggers the co­ llapse and allows to exhaustively control the damage evolution in wood. The algorithm is written in Fortran and has been successfully implemented in the Finite Elements Analysis software ABAQUS® by the use of a USDFLD user subroutine, which gives great versatility and simplicity to the model in order to adapt it to the different properties of wood species. The algorithm has been va­ lidated by a complete series of tests. It has also been validated by its comparison with laboratory tests and tests found in the literat ure. Connections are the critica! points in wood structures. Most of the optimi­ zation improvements for these structures are based on their study. Thus, the algorithm is applied to joint models in which additional techniques are required to improve the behavior simulation. Among these techniques, the use of cohesive elements in order to simulate the failure made by the combination of tensión and shear forces, stands out. The use of a pressure-overclosure law that defi­ nes the contact stiffness between wood and the metal elements of the joints is highlighted too. This law is based on a pressure and penetration criterion that allows to improve the elastic stiffness of joints, a topic in which sorne researchers have already pointed out a tendency to be overestimated when simulating with Finite Elements. Finally, after validating the joint models that incorporate the algorithm and the techniques presented above, a preliminary application of the models is performed, aiming to optimize the use of the material. Studies on the group effect and load distribution between connectors are initiated . The effect of the variation of the joint geomet ry and the substitution of steel by more elastic materials are studied too. The obtained results demonstrate the capacity of the model to reproduce complex situations and show that the proposed optimization hypothesis really improve the efficiency of the joint , at least for sorne cases. Nonetheless, the results obtained from these preliminary studies are limited, but it is expected to increase the knowledge on the studied joints and to propose design rules continuing the initiated studies. These preliminary studies show that although these kind of tests are very complex, there is a great potential for the developed tools.