In the housing unit case, a structural unit concept based on recycled plastics was developed to address homelessness, disaster relief, and housing issues in developing countries. The Daedalus World Shelter concept considered proposes a solution to two world problems: waste plastics recycling and versatile shelters. ANSYS academic software simulated deflections and stresses under self-weight, snow, and wind loading. The prototype bridge deck was formed from square tubes sandwiched between two plates, all made from polyester/glass. Structural CAE models were created to study the response of the deck in a connected deck stringer configuration and allow for future parametric studies of the deck at other stringer spacings and deck thicknesses.
To facilitate the meshing of the large structure, the "extrusion" method was applied extensively, using a meshed 2-D cross-section as a template for generating brick elements. The model had 8,700 3-D elements with 28,700 nodes. Each node had three translational degrees of freedom. CAE predicted approximately 10 percent greater deflections than those measured during physical testing. However, modest overpredictions of the deck deflections were expected since minimum design material properties were used in lieu of actual material properties, overcompensating for the stiffening effect of the coarse mesh. Further, the longitudinal strains on the bottom surface of the deck under the load patch were within seven percent of those measured before the fatigue tests. Both test and CAE model values of transverse strains showed comparable bands of alternating compressive and tensile patches (Figure 1).
Figure 1: Longitudinal Strains across square tubes
The Daedalus structural housing unit employed a basic building block, a ribbed plate panel, for all panels in the plastic shelter: walls, roofing, flooring, doors, and windows. The structural analysis consisted of beam-theory-based parametric studies and analysis of the unit panel, in addition to a complete structural analysis. In all analyses, brick elements were used in place of shell or 2D elements. Bricks offered analysis speed-enhancing advantages while maintaining physically meaningful models for the primary users of the results: architects, building contractors, plastics extruders, and civil engineers.
The panel mesh was generated by extruding a unit cell and then replicating the cell to form a panel. The constraints on the panel were vertical, "pinned" supports around the perimeter of the panel. Self-weight was the first loading. A standard snow load was then applied and calculated deflections were marginally acceptable (0.76cm or Length/160). Significant roof deflections were not acceptable for flat roofs with snow or rain loading because roof deflection and the resultant "ponding" would grow out of control. Therefore, a pitched roof was recommended. For an assessment of deflection in wall, roof, and related connections, a wind speed of 45m/s (100mph) was selected with a mean recurrence interval of 50 years.
The structure's design was clearly deflection controlled. Standard snow loads gave acceptable deflections for the 3.5mm thick panels, suggesting the planned 5mm panels would be more than adequate. As with the bridge deck case, several key design decisions were made on the basis of the CAE models. There is a significant application pool for CAE, even where only limited hardware and limited-version software are available. The CAE procedure used for the deck and housing applications suggests a strategy for small companies and university environments. Additionally, where resources are not as limited, the procedure may be used to model much larger, more involved structures.
Figure 2: Deflection under wind loading