The implementation of finite element methods (FEM) for fatigue analysis of complex structures in industry are becoming an increasingly effective and accepted practice. In the case of large plate-like structures, such as Load Haul Dumper (LHD) equipment, constructional frames and supports in plants and heavy vehicle trailers to name but a few, modeling can take place by implementation of either two dimensional shell elements or three dimensional solid elements. It is, however, not clear which shell element modeling procedure is the most realistic. Solid elements are accepted to give the closest resemblance since the element itself is the closest to reality in terms of geometry and also due to the fact that it is a three dimensional element. Due to economical and practical considerations, however, shell elements are used in industry - especially in large, plate-like structures. Another primary source of uncertainty lies with the definition of nominal stress in complex structures and the correct determination and extraction thereof from finite element obtained stress distributions. This situation occurs as a consequence of the absence of clear and distinct guidelines in the nominal stress based fatigue design codes such as BS 7608:1993; ECCS 6: 1985 and IIW XIII-1965-03 on weld modeling and nominal stress extraction procedures in conjunction with FEM. Explicit guidelines for finite element modeling and fatigue relevant stress determination do exist in the IIW fatigue design recommendations on top of the nominal stress guidelines, but focus primarily on the implementation of the hot spot stress fatigue assessment procedures. This dissertation consequently entails the development of a nominal stress extraction procedure for fatigue design and analysis of plate-like structures, utilizing shell elements. Firstly, the integrity of shell elements as concerned with the accurate capturing of the stiffness properties and stress distribution in the vicinity of welds are investigated, with the aim of establishing a set of guidelines and recommendations for the correct meshing and modeling procedure of welds in plate-like structures. Secondly, an extensive numerical investigation into the stress concentration characteristics of various T-piece and stiffener configurations is performed, resulting in a nominal stress extraction procedure. The developed methodology is applied on a complex plate-like structure for verification purposes. The structure is modeled by means of a finite element model, compiled according to the meshing recommendations developed. The stress distribution due to static loading is investigated and compared with measured values. Furthermore, the stress response of the structure due to stochastic dynamic loading is investigated and also validated in terms of the suitability for assessment by static equivalent design criteria, in particular the Fatigue Equivalent Static Load (FESL) methodology. A nominal stress and hot spot stress fatigue life prediction under stochastic loading is made, based on measured stresses in conjunction with the developed stress extraction methodology and the IIW guidelines respectively. Furthermore the finite element stresses are implemented in conjunction with the FESL procedure to repeat the nominal stress and hot spot stress life predictions. The viability and integrity of the FESL methodology is also critically assessed. The actual fatigue life of the structure under the particular loading characteristics is then determined and compared to the predicted lives.
Dissertation (MEng (Mechanical Engineering))--University of Pretoria, 2006.