THE USE OF COUPLED AND UNCOUPLED ANALYSIS TECHNIQUES IN THE ASSESSMENT OF BLAST WALL RESPONSE TO EXPLOSIONS

摘要

Floating oil and gas production facilities are highly weight sensitive. Yet, certain safety critical structural elements of these facilities have to be designed to withstand high explosion loads. This is particularly the case for blast walls. The situation is made more acute as some such walls are free standing with no lateral support at the top edge of the wall making them prone to high deflection at the top and large strain at their supports. It is therefore important to use analysis techniques that enable a realistic assessment of such structures without being over-conservative so that their weight may be controlled. In the present study, three analysis techniques were used with a view to comparing the effect this has on the response of the blast wall: (a) Lagrangian, (b) uncoupled Eulerian-Lagrangian (UEL) and (c) coupled Eulerian-Lagrangian (CEL). In the first Lagrangian approach, the blast loading is approximated as a reflected pressure-time curve (normally obtained from a CFD simulation of the blast); this is applied to the surface of the wall. The UEL approach requires an Eulerian simulation to be performed using as input the overpressure parameters obtained from a CFD analysis. The Eulerian analysis can be used to give a better estimate of the overpressure distribution on the wall) and is followed by a Lagrangian analysis. In the Eulerian analysis, the blast wall is treated as a rigid object and the analysis is used to determine the blast load distribution over the wall surface. This is followed by a Lagrangian analysis to determine the structural response under the pre-determined blast load distribution. Finally, CEL involves coupling of the Eulerian (blast wave) and the Lagrangian (structure) to describe the interaction between the blast wave and the structural response throughout the response time. Both loading and response can be described more accurately by this approach even when deformations are large. The paper demonstrates how coupled analysis allows the effect of the interaction between the load and the response of the wall to be accounted for. The results are compared with those from Lagrangian and uncoupled analysis and the differences are reconciled through intermediate analysis steps. It is shown how deflection and strain performance criteria can be satisfied using the coupled analysis whilst avoiding unnecessary increases in the weight of the structure. As the complexity of the analysis increases from Lagrangian to UEL and CEL, the computational demand increases significantly. A comparison of the times needed for the analysis in this study is also given.