Stochastic Analysis of Shape Imperfection in RC Cooling Tower Shells

Abstract

In this paper, the response variability of a reinforced concrete cooling tower shell due to shape imperfections is investigated, based on statistical stochastic analysis. In general, shape imperfections are modeled as ideal shapes, such as axisymmetric cosine shape and/or cyclic shape imperfections. However, these kinds of idealized imperfections cannot well enough represent the real shape imperfections in actual cooling tower shells. In this study, the shape imperfection of the cooling tower is assumed to be globally distributed stochastic fields rather than localized ones. The two geometrical parameters, i.e., thickness and radius, of a cooling tower shell are chosen as the spatially varying stochastic fields with predetermined statistical terms. The amount of additional responses revealed that the coefficients of variation (COVs) of stresses are larger than those of displacements, and that the additional response due to randomness in the cooling tower radius is larger than that in the shell thickness. The COVs of these two parameters show quite opposite trends. As the type of stochastic field approaches white noise, the response variability is decreased in the case of randomness in thickness, but increased in the case of radius.