This paper describes simplified and advanced methods which can be used in the evaluation of the blast resistance of offshore structures and individual components. For component analysis, improvements to single degree of freedom procedures are suggested, which enable the efficient modelling of the strain-rate effect and the beam-column action. For assemblages, adaptive nonlinear analysis procedures supplement the accuracy of the nonlinear finite element method with considerable computational and modelling savings, rendering its application practicable. Several comparative examples are presented in this paper which illustrate the applicability and relative accuracy of the various methods.
The resistance of offshore structures to explosions represents a major design consideration which is mostly addressed on two fronts; the first is concerned with minimising blast overpressures through improved design details, whereas the second involves the evaluation of the local and global structural response to such overpressures. In the latter context, appropriate consideration must be given to the geometric and material nonlinearities in the structural response, since the behaviour of structures subjected to explosion is usually associated with large displacements and considerable material inelasticity. This paper discusses two approaches often employed in the evaluation of member and overall structural response under explosion conditions: equivalent single degree of freedom analysis and nonlinear finite element analysis. The applicability of the equivalent single degree of freedom approach to the blast assessment of individual components is appraised, and improvements allowing the modelling of the strain-rate effect and the beam-column action are suggested. The nonlinear finite element approach is then discussed, and established as an accurate, albeit computationally expensive, method for predicting the large displacement inelastic response of structures subject to explosion loading. Previous work by the first author has shown, however, that modelling and computational savings in excess of 90% can be achieved at no loss in accuracy.
