No abstract available

[en] Buckling analysis methods and acceptance criteria for single shells of various configurations are well documented and adequately covered by many codes. There are, however, no guidelines or criteria for large Double-Shell-Stiffened (DSS) structures, which have been used recently in nuclear power plant applications. The existing codes for buckling analysis cannot be directly utilized because of the uniqueness of structural configuration and complexity of loading. This paper discusses a method for determining the critical buckling loads for this type of structure under a multitude load and suggests buckling criteria for the design of DSS structures. The method commonly used to determine the critical buckling loads for a single shell with or without stiffeners applies reduction factors to the theoretical results. The capacity reduction factors, which are often obtained from experimental results, include plasticity corrections and account for the difference between actual and theoretical buckling loads resulting from the effects of imperfections and nonlinearities. The interaction formulas derived from experimental results can be used to compute the interaction effects of three stress components. This paper extends these concepts and discusses their applicability to a DSS cylindrical structure. (orig./HP)

[en] To perform dynamic analysis a complicated system is frequently replaced with an appropriate equivalent simplified member. The most common way of finding such an equivalent member is to place lumped mass and mass moment of inertia at the corresponding floor level of a beam such that the significant frequencies of this beam would match with those of the original structural system. The effects due to the linked members between wall and column as well as the unsymmetrical geometric effects can be taken into consideration by making a 3-D static analysis. Nevertheless, the rotational stiffness is in general difficult to be determined. The simplest way of finding such a floor characteristics could be the least square method or the collocation method. (orig.)

[en] To study the load time history variation in the frequency domain, both Fourier transformation and response spectra were considered. Normal approximation (Central Limit Theorem) is employed to derive statistics for various confidence levels. Correlation factors on responses from reduced case applications are presented for various confidence levels. It is concluded that normal approximation with response spectral approach is the better method for determining the correlation factor for a pre-established confidence level, and that sufficient reliability on the structural responses may be achieved when reduced case applications are considered. This paper discusses methods of establishing correlation factors required for the adjustment of structural responses to achieve in acceptable confidence level. (orig./HP)