Uma nova metodologia para avaliação da vulnerabilidade estrutural e robustez

Abstract

The design of structures with the adoption of normative safety coefficients is not always adequate to guarantee the stability of constructions, especially in cases where unforeseen actions lead to localized failures in supporting elements. As examples, one can mention the cases of the World Trade Center, in the United States, where airplane collisions caused the collapse of both towers, and the Ronan Point Building, in England, where an explosion of a gas cylinder caused the failure of an entire section of the building. Thus, the consideration of the random nature of loads is important to design a safe structure. There are several methods of structural design that consider uncertainties in the construction’s calculation and execution processes. The present work specifically focuses on the use of probabilistic techniques. Probabilistic methods always admit a tiny probability of failure and often demand high computational efforts. In this sense, the current challenge is to develop a purely probabilistic structural analysis method with acceptable computational processing costs. Hence, this dissertation is based on a probabilistic methodology that proposes a new strategy to evaluate the global probability of failure and its respective robustness index, considering different types of uncertainties. The proposed methodology begins with the construction of an event tree, generating random scenarios through a Monte Carlo simulation. Two benchmark case studies are considered, yielding promising results with reasonable computational processing time.