Abstract
This document provides a framework for the assessment of gas explosion impact on structures. It shows how the probability of a structure achieving a predefined level of success in withstanding a defined gas explosion scenario may be calculated. The thesis opens with a general introduction to the need for assessing the impact of gas explosions in offshore installations. In the second chapter relevant literature regarding the existing theory and approaches used in assessment of gas explosions is reviewed. In particular, limitations of existing techniques are highlighted. The experimental data that is available to the author is also discussed. In the third chapter a set of objectives is defined. Meeting these objectives would allow many of the outstanding issues in gas explosion assessment, highlighted in chapter two, to be addressed. These objectives essentially define the need for assessment of explosions to relate to the impact upon the structure they engulf and that the impact is dependent not only upon the explosion but upon the structure so effected. The fourth chapter extends an existing technique of applying time-averaging windows to pressure-time histories from explosions. In this chapter, the importance of an approach to analysing explosions that relates directly to structural response is first established. An improved understanding of the effect of explosion parameters upon explosion magnitude is also made. The fifth chapter makes use of conventional signal analysis techniques to address weaknesses of time-averaging approach. These techniques allow a better relationship between identified structural response frequency and quantification of loading to be reached. These techniques also allow analysis of the explosion time history to be made in the frequency domain. Our understanding of the explosion parameters on explosion magnitude is again advanced. Whereas the fifth chapter allows analysis to be made in the frequency pressure domain, the sixth chapter shows how analysis can be made in the frequency-loading domain. This is achieved by assessing the impact of pressure-time histories on a simple oscillating structure, a single degree of freedom object. Of all the techniques developed, this is believed to be the most appropriate. In the seventh chapter we begin to apply the techniques developed by assessing, among other things, the inherent repeatability of an explosion and our ability to accurately predict the magnitude of the explosion. We start to generate a body of statistical data relating to accidental gas explosions. Chapter eight shows how the knowledge gained in previous chapters might be applied within a framework that is relevant to current practices in risk assessment. The framework selected is known as the limit state approach. It is shown how performance criteria for the structure undergoing gas explosions might be set and how the probability of the structure meeting these conditions might be assessed.