'Circling the square', or the generalisation of a rapid predictor for blast wave shielding

An accurate, fast-running engineering model (FREM) has been developed to estimate the extent and magnitude of a lone obstacle’s mitigation of peak specific impulse, serving as a surrogate for CFD. Its full-field predictions offer a 26,000-fold reduction in computation time relative to the CFD software used in its development, with a conservative point-to-point predictive error of 1.25%. However, its accuracy has only been verified for annular sector obstacles – a geometric simplification eliminating second-order flow-field interferences – significantly limiting its practical applicability. This work extends the FREM to more realistic urban geometries, namely rectilinear obstacles, via a ‘curvilinear transform’ that draws approximate equivalency to the simplified geometry necessary for applying the predictor. A preliminary transform, derived from engineering judgement, was evaluated against Viper::Blast CFD simulations of 273 unique rectilinear obstacles. It achieved a mean point-to-point error of 1.4% on average. Evolutionary optimisation of the analogues’ dimensions produced no meaningful improvement. Therefore, the proposed approach is found to be an effective, accurate and conservative extension of the FREM to more practical urban blast scenarios. The surrogate’s accuracy implies its idealisation of a uniform arrival to be representative of the physics of the true blast-obstacle interaction, demonstrating the assumption of a planar shock to be approximately valid for the range of geometries considered. The analytical framework can therefore be used to ascertain the scaled obstacle dimensions and stand-off distances bounding the validity of the planar shock assumption to inform design guidance.