Portfolio
Portfolio

CFD Simulation of Explosion

5b518cd12a56d.png


This is an introduction to computational fluid dynamics (CFD) analysis techniques for supersonic rapid combustion reactions (detonation), such as gas explosions and solid propellant explosions.


Explosion simulation using CFD presents significant challenges in terms of physical modeling, numerical analysis, and computational resources.


From the perspective of physical modeling, challenges include the extreme scale difference between the reaction zone (several μm to mm) and the explosion’s influence range (several m to tens of m), analysis instability caused by rapid changes in density/temperature/pressure before and after shock waves and detonations, complex chemical reactions, and reproducing reflection and refraction phenomena when shock waves pass through material boundaries.


Numerically, it requires shock capture techniques, high-order accuracy methods, boundary tracking techniques for explosives and surrounding media, and complex chemical reaction models.


Computational resources face challenges due to the need for extreme grid refinement in shock and reaction regions, leading to an explosive increase in total grid cells. Small grid sizes and high speeds demand very short time steps, making it difficult to secure sufficient total simulation time.


NEXTFOAM has developed a CFD analysis code capable of overcoming these challenges and analyzing supersonic combustion (detonation) phenomena, such as gas explosions and solid propellant detonations, at high resolution. It incorporates state equations and reaction rate equations specific to each explosive type to model complex chemical reactions and calculates scalar (λ – reaction progress variable) transport equations.


Explosion analysis must handle an extremely wide scale, from reaction regions in the μm range to the entire explosion domain in the m range. To address this, an Octree-based Anisotropic Adaptive Mesh Refinement (AMR) is applied to optimize computational resources. Grids are concentrated only in regions of interest, enhancing accuracy.


To ensure analysis accuracy, a third-order or higher ENO (Weighted Essentially Non-Oscillatory) spatial discretization technique and Runge-Kutta time integration are applied to stably capture rapid pressure and density changes, such as shock waves and detonation waves.


For boundary tracking, Level-set and Ghost Fluid techniques are utilized to enhance the precision of boundary condition handling. In parallel computing, Dynamic Load Balancing maintains efficient computational performance even in large-scale analyses.


 


The efficiency of this technology has also been demonstrated in the 2D Riemann problem benchmark. Compared to traditional structured grid analysis, it achieved up to 98% memory savings and a 75-fold improvement in computational speed, while also showing very high agreement with experimental data.

 


CaseMeshTimeRAM Usage

Structured code1,000,000223.84 sec11.04 GB
AMR level 3155,00039.23 sec1.6 GB
AMR level 712,5002.97 sec0.01 GB

There are two actual interpretation cases.


First, in the ‘Confined Ratestick’ interpretation, the grid size was automatically adjusted within the range of 31.25 to 500 μm to precisely match the shock wave velocity with the experimental results.

 


In open-field explosion analysis, the grid size was automatically adjusted between 0.3 and 320 mm, and the time interval between 1e-8 and 1e-5 seconds. The shock wave pressure at each distance was successfully verified against experimental data.


 


This explosion CFD analysis technology can be applied across diverse fields, including military and defense sectors for explosive performance analysis and calculating safe blast distances, industrial sectors for gas explosion risk assessment, and R&D stages for designing new propellants and detonators. 


Above all, our strengths lie in our capability to simultaneously analyze extreme scale differences, our experimentally validated methodologies, and our extensive experience in various explosion types—enabling us to deliver the analyses our customers require quickly and accurately.