Abstract

Abstract In offshore structures there are instances where the application of passive fire protection (PFP) is not possible or desired on certain portions of a structural member's surface area with the most common cases being those where the top surface is left unprotected due to deck grating or plating and for skid beams. Standard methods of fire analysis are not appropriate for partially unprotected members because of the extreme thermal gradients that may occur in portions of the cross-section during a fire. This paper will demonstrate that the behavior is difficult to predict for such unique problems and that advanced finite element (FE) analysis is usually required to ensure the assessment does not lead to un-conservative design and conclusions. The analysis technique utilized was a sequentially coupled thermal-stress analysis in ABAQUS. The results showed that the top flange heats up rapidly and that the heat conducts very slowly down the web to the rest of the cross-section, with an extreme thermal gradient occurring over a small height. This causes such structures to be susceptible to local buckling and lateral-torsional buckling which cannot be predicted accurately with simplified methods. The study will show that for similar types of structural components to survive a fire event, sufficient redundancy and ductility is required to allow for catenary behavior. An FE solver capable of analyzing problems with large strains and deflections is therefore critical to predicting the response. Introduction In the offshore industry it is common for deck beam members to have the surfaces below the deck protected while the top surface is exposed to direct loading from extreme fire events. Current code and standard provisions on heat transfer and strength assessment of deck/floor members using simplified and advance methods are not directly applicable to these cases due to the large temperature differential that occurs between protected and unprotected elements of the cross section. There are simplified strength methods published by Eurocode3 [1] for building members that cover different types of typical loading/protection schemes. However, the loading/protection cases common for offshore deck members are not covered. Dwaikat and Kodur [2] have developed simplified performance based methods for restrained beams that can include thermal gradient. However, these were not developed for the type of extreme thermal gradient that will occur over a small portion of the cross-section in the cases described above. In addition, mostly compact rolled sections with lateral restraint throughout the fire have been researched and tested. Such assumptions are valid for most onshore applications, but are usually not for offshore structures where plate girders are used with non-compact or slender elements and where the deck components providing lateral support may lose their integrity. Thus conclusions from such studies on local buckling and lateral-torsional buckling are not applicable. The intent of this paper is to demonstrate how advanced FE analysis can be used to solve unique fire analysis problems such as this where the behavior is not well understood or easily predicted. However, testing and simplified analysis is always recommended to supplement the findings of such an analysis.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call