3. Electronic Theses and Dissertations (ETDs) - All submissions
Permanent URI for this communityhttps://wiredspace.wits.ac.za/handle/10539/45
Browse
3 results
Search Results
Item An expert system for selecting roofing types in developing areas(1993) Oliveira, Silvia Da Conceicao CunhaThis dissertation has as it's objective to develop an expert system for the purpose of investigating the suitability of using expert systems in developing areas. It is the first time to our knowledge that an expert system has been tested for usage in such a context. The housing and more specifically roofing domain is investigated for an application topic resulting in "the selection of suitable roofing types" being chosen. Potential end-users are identified, an appropriate context of such a system is outlined and an expert system tool is selected, The knowledge engineering and building of the knowledge base are described. Feedback from the parties involved in testing and validating BUILDROOF is documented. In addition the suitability of using expert systems in the developing areas context in which BUILDROOF was developed, is discussed. Finally, recommendations regarding potentially suitable areas of application for expert systems in developing areas are outlined.Item A parametric study on IBR sheeting supported by purlins(2016) Mlasi, Marope StellaThis dissertation presents a parametric study conducted on the dimensions of an inverted box rib (IBR) sheet. The objectives of the study were (i) to obtain the optimum dimensions that would be used to simulate new IBR profiles made from standard sheet input coils whilst meeting the stiffness criterion; and (ii) to perform a cost analysis to determine the most economic simulated profile. The dimensions which were varied in the parametric study were rib-height, rib-width, sheet thickness and the number of pans per 686 mm cover width of a single IBR 686 sheet. Numerical experiments were conducted using the Finite Elements Method and Abaqus/CAE software. The results were compared to the predictions obtained from Euler-Bernoulli beam theory. The outputs from the experiments were the sheet deflection from which stiffness was determined, and eigenvalues from which the profiles’ stability and buckling modes were calculated. This study found that sheet stiffness increased as the rib-height, sheet thickness and number of pans per 686 mm cover width increased. In contrast, the rib-width had little effect on the sheet stiffness. Hence, it was concluded that, for any IBR sheet profile, the rib-width should be kept at approximately 23 mm to avoid using more material in the sheet. The optimum dimensions found were rib-height of 34 mm, rib-width of 23 mm and four pans per 686 mm cover width. The commercially available IBR 686 sheet is made up of four pans and has a rib-height and a rib-width of 37 and 33 mm, respectively. It is manufactured from an input sheet coil of 925 mm. Reducing the rib-height from 37 mm to 34 mm and the rib-width from 33 mm to 23 mm resulted in increasing the cover width by 8 %. The optimum dimensions were further used to simulate profiles made from the 925, 940, 1000, 1175, 1219, 1225, 1250 and 1320 mm standard sheet coils in order to find the most economic IBR profiles that met the stiffness criterion. The 1250 mm coil yielded the most economical IBR profile, which has six 97 mm wide pans, and a 996 mm cover width. This profile resulted in a 10 % cost saving compared with the next closest profile.Item The snap-through stability of plastically designed steel pitched-roof portal frames(2015-01-12) Bryant, John SpencerRecent proposed rafter slenderness limits, to prevent snap-through of plastically designed pitched-roof portal frames, incorporate the elastic snap-through buckling load of such frames. It has been suggested that the elastic snap-through buckling load used in the proposals is over-estimated making these slenderness limits unconservative. This is supported by a more rigorous elastic analysis. To test the proposals, model frames lying on or close to the slenderness limits were tested to failure in the laboratory. Frame dimensions were chosen so that the frames were only susceptible to snap-through instability. Failure loads far lower than the expected plastic collapse loads were measured, showing that the elastic snap-through buckling load is over -estimated . Since plastic analysis is easily applied to portal frames, these slenderness limits are best replaced by a similar limit incorporating a more accurate elastic snap-through buckling load. A new limit is outlined which must still be tested by further research