Teses #dimensionamento

The issue of this Master Dissertation is the design of steel beams and columns in fire situation. The formulation for the design in fire was presented, explaining its origin. The main objective is the development of a computational tool, which was called AçoInc, able to calculate the resistances of steel elements at room temperature according to ABNT NBR 8800:2008 and in case of fire according to ABNT NBR 14323:2013. The temperature of the steel element in fire is also obtained through the AçoInc, following the heat transfer model of ABNT NBR 14323:2013 and basing on fire curves. These curves are the standard fire curve associated with the required fire resistance time and parametric curves for natural fire model of Eurocode 1 (2002). With the support of spreadsheets, some studies were performed. The first one is probabilistic study of failure of column and beam in fire situation considering the parametric fire and by means of the Monte Carlo method. Other study is parametric sensitivity analyses, which were performed to verify how the resistances are influenced by the dimensions of cross section, the variables of the natural fire and the properties associated with heat transfer. For this study, in addition to spreadsheets, we used the Software Statgraphics Centurion XV (2007), which contains resources of statistical analysis. Applying the spreadsheets, the equivalence between parametric curves and fire standard curve in the obtention of the steel temperature in the fire situation was verified. For the use of parametric fire, partial factors, based on the equivalent time method detailed on ABNT NBR 14323:2013 were proposed. At last, with the assistance of spreadsheets, simple non-computational tools were created for determining the critical temperatures for columns and beams in some constructive situations of practical interest. Some conclusions from these studies are that the steel temperature is the most important parameter in determining the resistance of steel structures in a fire situation, that the current formulation for parametric fire modeling leads to a variation of the fire temperature in function of the opening factor with undesirable discontinuity and that the data found in the international literature about probability of fire and flashover lead to divergent values of structural failure probability.