Inelastic Capacity of Pipe Rack Structures

Sammanfattning: Steel structures in an offshore processing environment must be designed for an accidental explosion event. The current verification procedure implemented at Aker Solutions (AKSO) enforces an elastic single degree of freedom (SDOF) analogy to implicitly account for dynamic effects in linear static FE analyses. The implication of this approach is that structures are designed to remain within their elastic range during an explosion event, generating weight and cost inefficient designs. The objective of this thesis was therefore to determine the inelastic capacity of a typical pipe rack structure subjected to blast. A modified verification procedure was also evaluated to determine if this unutilized capacity may be accounted for via incorporation of an inelastic SDOF-response model. Research on the blast phenomena showed that the dominant load effects from interaction between a pipe rack structure and a blast pulse were governed by the dynamic blast-pressure, idealized as a symmetric triangular pulse-excitation to characterize the typical deflagration type explosion. This pressure was converted to drag loads and assigned to structural members in the conducted finite element analysis (FEA) study, aiming to reveal the true dynamic structural response to pulse excitations of varying magnitudes. A specific pipe rack design from a previous AKSO project was chosen for evaluation. The associated piping configuration was accounted for implicitly in the FEA study via simplified calculations of mass contribution and transfer of blast loading to the analyzed rack structure. Failure criteria were also established to prevent rupture of safety critical pipe lines according to the defined post-blast functionality requirements. Results from full nonlinear FE analyses showed that the original pipe rack design remained completely elastic during simulations of the design blast-pulse scenario. The validity of the current verification procedure was thereby confirmed and the associated elastic SDOF-model was found to prevail with great accuracy. Failure criteria were not exceeded until the original design-pulse had been magnified by a factor of 4.8 to a corresponding peak pressure of approximately 1 barg and pulse duration of 50 milliseconds. The amount of inelastic capacity possessed by the pipe rack at this load level was quantified by the observed ductility ratio of 2.25. The expanded SDOF analogy failed to capture the inelastic dynamic response with a satisfactory level of precision. However, conservative estimates of a pipe rack structures full capacity were obtained when this inelastic SDOF-model was incorporated into the modified verification procedure.