The generalized local model: a methodology for probabilistic assessment of fracture under different failure criteria
- Autores: Miguel Muñiz Calvente
- Directores de la Tesis: Alfonso Carlos Fernández Canteli (dir. tes.)
- Lectura: En la Universidad de Oviedo ( España ) en 2017
- Idioma: español
- Tribunal Calificador de la Tesis: Eugenio Giner Maravilla (presid.), María Cristina Rodríguez González (secret.), Sabrina Vantadori (voc.)
- Materias:
- Texto completo no disponible (Saber más ...)
- Resumen
When designing structural and mechanical components, the criteria of structural integrity must be accomplished. Such criteria consist in assuring that the component will perform its designed function of supporting loads or any kind of actions causing stress and strains in the material without catastrophic failure. This implies from the point of view of fracture mechanics safe prediction of static, dynamic and fatigue failure what requires previous knowledge of the loads but also the mechanical properties of the component material such as density, hardness, strength and toughness but also of dimensions (shape, size and thickness) which should be suitable to withstand the actions operating on it.
Moreover, this concept requires performing an experimental program and its statistical analysis, as well as the interpretation of the results obtained, in order to achieve the mechanical characterization of the material. The latter consists in the probabilistic definition of the critical values of the reference parameter, denoted generalized parameter (GP) that will be used as a property related to the material to predict failure of the components made of such material.
In this doctoral thesis, an approach is presented aiming at providing an objective definition of the material characterization in order to guarantee transferability from the experimental results obtained irrespective of the test chosen and the shape and size of the specimen to the component design. This methodology, which is denoted Generalized Local Model (GLM) because it represents a general solution without restrictions of former particular local models, provides the statistical information related to the distribution of the reference parameter considered, i.e. the primary failure cumulative distribution (PFCDF), which is identified as a characteristic property of the material under study. Moreover, this allows a possible assessment and eventual selection of the generalized parameter (GP) to be made as a result of the evaluation of its suitability.
After the GLM is introduced ad justified, some of its extensions are presented in order to overcome some usual situations arising during the evaluation of experimental results. First, the GLM is applied to the joint evaluation of experimental data from different programs, using different specimen or even test types, as they were a unique sample. In this way, higher reliability in the derivation of the primary failure cumulative distribution function (PFCDF) is achieved. Second, a methodology is developed and applied to solve the assessment of the so-called confounded data case (concurrent flow populations) as an extension of the GLM. This allows the PFCDF to be obtained independently of the failure type involved. Third, the extension of the GLM to the fatigue lifetime analysis is justified as a result of the conjunction of the GLM and the fatigue Weibull model developed by Castillo and Canteli.
The application of the models proposed in this doctoral thesis is illustrated using practical examples of such as: a) individual and joint assessment of brittle fracture of glass plates; b) individual and joint assessment of experimental data from different test types on a glass composite material; c) analysis of the influence of the statistical scale effect (size and thickness) in cleavage fracture; d) assessment of fatigue results on notched specimens from standard test data to predict lifetime of structural bolted joints, and e)probabilistic assessment of fatigue results in the high cycle (HCF) and very high cycle fatigue (VHCF) domains.
As a final consideration, it is worth to mention that the GLM is a fundamental tool for obtaining the PFCDF this being applicable without loss of generality, regardless of the possible complexity of the selected failure criterion, the geometry of the specimen and the type of test or load considered in the experimental program, as well as the characteristics of the material. For the valid application of the GLM, only suitability of the proposed failure criterion is required, this being related to the selection of the appropriate GP, provided applicability of the weakest link principle and the existence of statistical independence between the results of the different specimen and test types used.
