Probabilistic safety analysis of existing infrastructure

Abstract

Existing infrastructures have significant economic value in the world. These infrastructure networks can also be subjected to hazard events mainly by nature or human. Besides, deterioration of material used overtime via such corrosion/fatigue can affect the structural conditions that may lead to failure. In order to avoid various failure mechanism of these structures, periodic maintenance is essential based on specific Risk-approaches, which would quantify the uncertainties and evaluate the safety. Therefore, this dissertation aims to investigate the response parameters of critical infrastructures (specifically bridges) and to consider the statistical evaluation of the past documented failures due to natural hazards. A literature review is developed covering bridges disasters over in the past 50 years, highlighting the primary failure causes. It could be noticed that Floods can considered one of the leading causes of bridge failures. In rainfall regions, flash flood events tend to occur more regularly compared to earthquakes. In many countries, damage caused by the floods and collapse constituted a large proportion of disasters. Although, the flood fragility has gained relatively little attention, in comparison to detailed seismic fragility research for civil infrastructure. In this dissertation, a novel approach to extract flood fragility curves for bridges has been proposed. Fragility curves are often obtained through structural analyses and failure modes, which are classified as excessive requirements of a bridge subjected to high water pressure (due to water level, velocity and structural degradation), and these known as the primary causes of bridge failures during floods. The introduced flood fragility model is analysed by using SeismoStruct software, to show a comprehensive approach measuring the probability of flood-induced bridge failure efficiently. The results of the numerical model were validated, and it is demonstrated that the fragility of the flood increased in a flood event as the water velocity increases. However, the results indicated that the probability of exceeding damage rates on a different water level is almost similar.