| Summary: | This work is mainly concerned with the development of sophisticated yet efficient methods for assessing the elastic buckling of steel beams with web openings, focussing on local buckling effects in the web region. A new computational method is proposed which extends the use of the Element Free Galerkin (EFG) method for the numerical discretisation combined with a simplified buckling assessment approach based on the Rotational Spring Analogy (RSA). The new approach considers several potential simplifications offering a balance between computational efficiency and accuracy in local buckling analysis. In the present EFG/RSA method, considerable advantage is established by separating the planar and out-of-plane responses. Planar analysis is further enhanced using modular concepts, where the beam is divided into unit cells, each of which resembles a super-element with a reduced number of freedoms, and solved using a standard discrete procedure. As for the out-of-plane analysis, the application of a ‘local region’ is adopted to significantly reduce the size of the original buckling problem. Finally, local buckling assessment is conducted using an effective approach that utilises an iterative procedure based on a rank 2 reduced eigenvalue problem along with a shifting local region. Several illustrative examples are provided which highlight the efficiency and accuracy of the developed approach in comparison with detailed nonlinear finite element analysis performed using ADAPTIC, and which demonstrate general applicability to local buckling analysis of steel beams with web openings of various shapes and sizes. This work also considers the development of a simplified design-oriented method which is presented particularly for web-post buckling. Towards this end, a simplified analytical model is proposed based on an analogy with equivalent rectangular thin beams (RTB), where a semi-empirical approach is used to calibrate the suggested formulation against the results obtained from the numerical work performed earlier.
|
|---|
