Fluid–structure coupled simulation framework for lightweight explosion containment structures under large deformations

Abstract

Lightweight, single-use explosion containment structures provide an effective solution for neutralizing rogue explosives, combining affordability with ease of transport. This paper introduces a three-stage simulation framework that captures the distinct physical processes and time scales involved in detonation, shock propagation, and large, plastic structural deformations. A working hypothesis is that as the structure becomes lighter and more flexible, its dynamic interaction with the gaseous explosion products becomes increasingly significant. Unlike previous studies that rely on empirical models to approximate pressure loads, this framework employs a partitioned procedure to couple a finite volume compressible fluid dynamics solver with a finite element structural dynamics solver. Given the rapid expansion of explosion products and the large structural deformation, the level set and embedded boundary methods are utilized to track the fluid-fluid and fluid–structure interfaces. The interfacial mass, momentum, and energy fluxes are computed by locally constructing and solving one-dimensional bi-material Riemann problems. A case study is presented involving a thin-walled steel chamber subjected to an internal explosion of $250\text{g}$ TNT. The result shows a 30% increase in the chamber volume due to plastic deformation, with its strains remaining below the fracture limit. Although the incident shock pulse carries the highest pressure, the subsequent pulses from wave reflections also contribute significantly to structural deformation. The high energy and compressibility of the explosion products lead to highly nonlinear fluid dynamics, with shock speeds varying across both space and time. Comparisons with simpler simulation methods reveal that decoupling the fluid and structural dynamics overestimates the plastic strain by 43.75%, while modeling the fluid dynamics as a transient pressure load fitted to the first shock pulse underestimates the plastic strain by 31.25%.