Seismic Risk to Major Industrial Facilities

Seismic risk to major industrial facilities has been studied primarily in relation to nuclear power plants. Recently, however, attention has been focused on other large industrial plants due to the increasing awareness of the potential of damage of such plants and the consequences from failures at such plants.

The overall goal of this study is to develop a general methodology for seismic risk evaluation at major industrial facilities. Such a methodology includes (1) methods for estimating the damage to structures and equipment, (2) methods for evaluating the performance of the overall system, and (3) formulations for estimating the economic and other losses. This report focuses on the first of these aspects of the general risk methodology.

Currently existing methods for evaluating the performance of structures at major industrial facilities when subjected to earthquake ground motion were reviewed. The advanced first order reliability approach be used for that purpose. The main advantages of this approach is that reliability computations are mathematically invariant and uncertainties included in the analysis relate to physical parameters which can be quantified more readily. Simplified formulations for the limit state equations are presented which can be standardized for classes of structures with similar design configurations. The methods for failure probability (or reliability) estimates are illustrated through specific examples. For the purpose of presenting application examples, typical structures found at oil refineries are identified. Reliability estimates are obtained specifically for tall isomerization columns of uniform and non-uniform cross sections. Various degrees of complexity are introduced in the structural response formulations to assess the adequacy of the simplified approach.

Currently, methods for system performance are being investigated for potential use in the general seismic risk methodology. The attention is focused on incorporating component dependencies within the system and multiple performance states of components. The systems analysis method will be tested through an application to a simplified hypothetical oil refinery.