PSI - Issue 13

Uroš Tatić et al. / Procedia Structural Integrity 13 (2018) 496– 502 Author name / Structural Integrity Procedia 00 (2018) 000–000

500

5

conditions that window can be exposed and different reinforcement designs. Additional physical specimens were not tested. All of the models were based on the same geometry used for experimental testing (M Sora Natura 3 window profile). 5.1. Model 1 Numerical model was defined as an “assembly” (two subassemblies: window frame and window casement) of the eight 3d deformable parts, with use of tie constrain:  Window frame o Middle beam  Window casement Numerical model was created as a half of physical experiment, due to a presence of symmetry plane. Difference in the profiles on the top and bottom did not allow use of the quarter of the model (Maneski T., 1998). During the definition of the geometry it was required to replicate real life parameters as close as possible with an introduction of the small approximations (in the form of small filets, chamfers etc.) on the locations where it is not of significant influence on the model structural behaviour and stress strain distribution. All of the profile elements were created as extrude feature of the main profile sketch with an additional cut feature used to define end cut profiles (profiles used to ensure contact of side and middle beam with top and bottom beam). Window frame was used to connect whole model to the ground and restrain it from movement during the calculation. Outer surfaces of the assembly (windows frame) were considered as encasture due to a testing conditions (window frame was positioned within testing rig and bolted to it during the test). Middle beam profile was connected to the outer frame with the use of tie constrain. Symmetry represents only additional restrain other the direct contact between elements. Positioning of the individual elements within the numerical model were defined with linear translation and rotation rather than the constraint (point to point, surface to surface, etc.). Experimental load used for the real experiment was simulated as scaled pressure load distributed only to a middle beam. 5.2. Model 2 Numerical model was defined as an “assembly” in a same way as model 1 but with added dowels used for mutual contact between elements. Positioning of the individual elements within the numerical model were defined with linear translation and rotation rather than the constraint (point to point, surface to surface, etc.). Experimental load used for the real experiment was simulated as scaled pressure load distributed only to a middle beam. Loading and boundary conditions were defined within two individual steps: Initial and Load. Tie constrain was used to simulate glued contact between dowels and each of the wood segments, but the mutual contact between them was simulated using contact restrains. o Side beam o Top beam o Bottom beam o Middle profile o Side profile o Top profile o Bottom profile

Fig. 5 Results of a deformation on the Model 1with tie constrain (left); Results of a deformation on the Model 2 with contact definition between individual elements (right)