دانلود رایگان مقاله آنالیز مقایسه ای از مدل عددی بخش ورق قوسی فولاد

abstract

The modern construction industry makes use of innovative production and assembly technologies, whose purpose is to implement reliable and simple structures. One product of such technologies is an arch-shaped steel sheet section that might be used as a selfsupporting covering for general construction and industrial building construction. Considering complex geometry and boundary conditions, the FEM model of this structure is sophisticated and may contain errors. This paper presents numerical approaches for the calculation of sheet metal section elements of such coverings, namely two numerical approaches that differ in detailing of the properties of considered physical object. The first approach is based on a model that is characterized by simplified geometry and boundary conditions. The second scenario concerns a detailed FEM model with actual geometry captured by a laser triangulation method, experimentally determined material stress–strain relationship, and load conditions measured on an experimental stand. The results obtained with the use of computer simulations based on both approaches described above and experimental results are compared. The errors caused by simplification of the first numerical model are discussed. Finally, an acceptable reduction of FEM model complexity is proposed for the analyzed structure.

5. Conclusion and further works

The analysis of arch-shaped covering sections is difficult as both an experimental test and a computational task because of the complex geometry and load conditions. The section geometry results from the conditions of forming by cold rolling, and loads are a representation ofthe work of a discrete piece of the self-supporting system of a roof covering. The tests concerned the impact ofthe geometrical parameters of complex numerical models of section and boundary conditions (the method of the model load on the material) on the searched value, which is the reaction of the kinematic extortionthroughoutthe load range, i.e.fromthe zero starting point to the destruction of the test piece. Two scenarios were analyzed. The first concerned a simplified geometry model (our own model), a single parameter extortion (axial displacement) and a bilinear material model. The second model was characterized by precisely mapped geometry obtained by 3D scanning, multi-parameter geometric extortions (axial displacements, angles of rotation of supports), which mapped the test conditions, and different material stress–strain relationship assigned to respective areas on the surface of the model. Computational models were compared with the test results in relation to the same parameter, i.e., the reaction resulting from kinematic extortions. In the process of the analyses, it was demonstrated that the divergences of the test and computational results of the two models over the entire area of sought values vary within the range of 0.3% to 13%. At the peak load value, the model analyzed in scenario 1 shows a 5.5% divergence of computational results compared to the tests, and 1.1% in scenario 2. When analyzing the selected displacement field, the computational results of both models are presented only with respect to the elastic strain. Plastic strains diverge considerably.