Seismic risk assessment of industrial plants is of paramount importance to ensure adequate design against earthquake hazards. Seismic vulnerability of industrial plant components is often evaluated through a fragility analysis to conform to structural safety requirements. Fragility curves of single components are usually developed by neglecting the effect of actual boundary conditions. Thus, an incorrect evaluation of individual fragility curves can affect the overall fragility curve of a system. This may lead to “erroneous” seismic risk evaluation for a plant in comparison with its real state. Hence, it is important to study the effect of uncertainties, introduced at the boundaries when coupling effects are neglected, on the dynamic characteristics of a system.

Along this line, this paper investigates the effects of uncertain boundary conditions on the probability distributions of the dynamic properties of a simple chain-like system with increasing number of degrees of freedom. In order to describe the uncertain boundary condition, a modified version of the well-known β distribution is proposed. Subsequently, the Analytical Moment Expansion (AME) method is employed to estimate the statistical moments of the output random variables as an alternative to more computationally-demanding Monte Carlo simulations. Finally, a preliminary extension of the proposed approach to a realistic piping system connected to a class of broad/slender tanks is discussed.

This content is only available via PDF.
You do not currently have access to this content.