Repository logo

A multi-physics methodology for the simulation of reactive flow and elastoplastic structural response

Published version

Change log



We propose a numerical methodology for the numerical simulation of distinct, interacting physical processes described by a combination of compressible, inert and reactive forms of the Euler equations, multiphase equations and elastoplastic equations. These systems of equations are usually solved by coupling finite element and CFD models. Here we solve them simultaneously, by recasting all the equations in the same, hyperbolic form and solving them on the same grid with the same finite-volume numerical schemes. The proposed compressible, multiphase, hydrodynamic formulation can employ a hierarchy of five reactive and non-reactive flow models, which allows simple to more involved applications to be directly described by the appropriate selection. The communication between the hydrodynamic and elastoplastic systems is facilitated by means of mixed-material Riemann solvers at the boundaries of the systems, which represent physical material boundaries. To this end we derive approximate mixed Riemann solvers for each pair of the above models based on characteristic equations. The components for reactive flow and elastoplastic solid modelling are validated separately before presenting validation for the full, coupled systems. Multi-dimensional use cases demonstrate the suitability of the reactive flow-solid interaction methodology in the context of impact-driven initiation of reactive flow and structural response due to violent reaction in automotive (e.g. car crash) or defence (e.g. explosive reactive armour) applications. Several types of explosives (C4, Detasheet, nitromethane, gaseous fuel) in gaseous, liquid and solid state are considered.



multiphase and multi-material, structural response, impact, condensed-phase, elastoplastic, ignition

Journal Title

Journal of Computational Physics

Conference Name

Journal ISSN


Volume Title



Engineering and Physical Sciences Research Council (EP/K014188/1)
This work was supported by Jaguar Land Rover and the UK-EPSRC grant EP/K014188/1 as part of the jointly funded Programme for Simulation Innovation.