blastFoam | Cityscape Explosive Modeling & Validation
In this short article we present comparisons of blastFoam simulation outputs with scaled explosive experiments conducted at Cranfield University (Brittle 2004) to investigate blast loading and interaction in geometrically complex urban environments.
Background
Current approaches to urban cityscape terrorism modeling, are stymied by numerous technical limitations related to: 1) the scale and fidelity of the urban terrain and, 2) the relative complexity of the modeling approaches required to calculate accurate airblast loading on structures and subsequent response/damage subject to these loads.
The key phenomenology associated with calculating airblast loads within an ‘urban canyon’ environment are related to the shielding, channeling, and reflection of the shock front as it propagates away from the point of detonation. These all occur in 3-dimensions and provide a significant technical challenge in terms of the detail and discretization of the analytical domain required to capture these phenomena, and the huge amounts of data that must be handled for both the domain and the calculation.
To make such modeling tractable, approaches have been developed which grossly simplify the key complexities and remove much of the physical basis to the calculations. A typical simplification approximates the extent and severity of structural damage from a pre-defined explosive scenario using concentric circles, based on a linear ground plane distance from the point of detonation. This approach might be appropriate for a very rapid assessment with a paucity of alternative options however, the technology and data is now available to conduct more complex assessments which include more of the inherent physics. This more detailed analysis can provide more accurate results across a broader set of problem domains and with less uncertainty inherent to the modeling results.
The employment of physics-based methods such as Computational Fluid Dynamics (CFD) to calculate airblast loads in urban environments is very much the current state-of-the-art. The approach explicitly models: 1) the detonation of the explosive material, 2) the expansion of the detonation products, 3) the development and propagation of a shock front in air, and 4) the reflections and dissipation of that shock front within the urban landscape. This has clear benefits over empirical, semi-empirical and closed form solutions, in that each calculation can account for the unique geometric conditions present within the computational domain.
Figure : An example of blastFoam explosive threat scenario modeling conducted by Synthetik. Scenario: 2,000kg TNT VBIED in the vicinity of London Bridge, UK
As with any modeling approach, the conduct of validation through comparison with experimental results is critical to assess the accuracy and reliability of simulation outputs. Building on extensive validation work already conducted by the Synthetik team (Brewer et al. 2021), this short article presents comparisons of blastFoam simulation outputs with scaled explosive experiments conducted at Cranfield University (Brittle 2004) to investigate blast loading and interaction in geometrically complex urban environments.
Experimental Set-up
CITYSCAPE GEOMETRY
The notional street configuration developed by Brittle (2004) incorporated buildings of varying geometry and orientation, set within a realistic road system with off-set junctions to provide complicated routes for blast propagation with varying degrees of confinement.
The street layout was developed from studying urban center street configurations that did not contain regular shaped buildings. A street width of 15m was selected as this represented a two-vehicle lane street with pedestrian sidewalks on either side.
Building heights were selected to provide a mixture of three- and four-story buildings as these are common to virtually all urban cityscapes.
Scale model blast experiments were designed and conducted at 1/50th scale for an equivalent 2,000kg truck bomb. A series of 18 firings were completed at the RMCS Explosives Research and Demonstration Area (ERDA), UK in May 2004.
The building models were constructed using un-reinforced concrete cast in wooden molds. Pressure transducers were accommodated by setting hollow plastic pipe in the concrete blocks at the relevant locations.
EXPLOSIVE CHARGES
At full-scale the explosive threat was intended to simulate a 2,000kg TNT truck bomb, with a height of burst of 2m above the ground. The equivalent scaled explosive charge required was 16g TNT, including the detonator. All firings were conducted using a spherical charge of PE4, with a scaled height of burst of 40mm.
blastFoam Modeling
As the experimental work was conducted at 1/50th scale, the blastFoam simulation was also conducted using scaled dimensions. The three-dimensional street configuration was generated by the Synthetik team and then imported into blastFoam.
Figure : Flythrough of the computational cityscape model generated by Synthetik for the blastFoam simulations, indicating the explosive charge (red sphere) and pressure gage locations (numbered).
A blastFoam simulation was then executed with pressure-time histories recorded for all surfaces (buildings and ground) and gage locations.
BLASTFOAM SIMULATION OUTPUTS
Figure : 2D plan and 3D blastFoam simulation outputs. Top: Numerical schlieren visualization to highlight density gradients (i.e., shocks). Bottom: Overpressure fringe plots.
Comparison of Experiment and blastFoam Simulation Data
The pressure- and impulse-time histories from the experimental results were compared to the numerical simulation outputs to assess the validity of using blastFoam to analyze blast wave propagation in complex geometrical situations.
Correlation between the experimental data and the blastFoam simulation were generally very good. The correlation tended to be better for gage locations where the blast waves were of low pressure and long duration, or waves had coalesced. This is not unexpected as resolution of pressure is cell size dependent.
The related time histories are generally of similar shapes and offer a good degree of confidence in the validity of blastFoam simulations for modeling explosive threats in cityscape environments.
References:
[1] M. Brittle, “Blast Propagation in a Geometrically Complex Environment”, Thesis (M.Sc.). Cranfield University, 2004.
[2] T. Brewer, J. Heylmun, B. Shields, and P. Vonk, “Validation of the blastFoam Computational Fluid Dynamics (CFD) Solver,” Synthetik Applied Technologies, Austin, Texas, USA, 2021.
[3] J. Heylmun, P. Vonk, P., and T. Brewer, "blastFoam: An OpenFOAM Solver for Compressible Multi-Fluid Flow with Application to High-Explosive Detonation." Synthetik Applied Technologies, LLC., 2019.
[4] I. G. Cullis, N. Nikiforakis, P. Frankl, P. Blakely, P. Bennett, and P. Greenwood, “Simulating geometrically complex blast scenarios,” Defence Technology, vol. 12, no. 2, pp. 134–146, Apr. 2016, doi: 10.1016/j.dt.2016.01.005.
[5] N. N. Fedorova, S. A. Valger, and A. V. Fedorov, “Simulation of blast action on civil structures using ANSYS Autodyn,” Perm, Russia, 2016, p. 020016. doi: 10.1063/1.4963939.
[6] W. Drazin, “Blast Propagation and Damage in Urban Topographies”, Thesis (PhD). University of Cambridge, 2017
[7] J. Tang, “Development of a Parallel Adaptive Cartesian Cell Code to Simulate Blast in Complex Geometries.” Dissertation. University of Queensland, 2008.
Acknowledgement
We would like to thank and acknowledge Mr. Stephen Hudson for his continued support and guidance in the development of our cityscape modeling capability.
About blastFoam:
blastFoam is an opensource solver for multi-component compressible flow with application to high-explosive detonation, explosive safety and airblast.
The solver is based on the OpenFOAM® framework and provides solutions to highly compressible systems including single and multi-phase compressible flow, and single- and multi-velocity systems.
blastFoam provides implementations of the essential numerical methods (e.g. 2nd and 3rd order schemes), equations of state (e.g. ideal gas, stiffened gas, Jones-Wilkins-Lee, etc.), run-time selectable flux schemes (e.g. HLL, HLLC, AUSM+, Kurganov/Tadmor) and high-order explicit time integration (e.g. 2nd, 3rd, and 4th order).
blastFoam provides activation and explosive burn models to simulate the initiation and expansion of energetic materials, as well as afterburn models to simulate under-oxygenated explosives that exhibit delayed energy release.
Synthetik's opensource CFD airblast code based on the OpenFOAM framework is available on GitHub: https://github.com/synthetik-technologies/blastfoam
blastFoam is developed by Synthetik Applied Technologies:
https://www.synthetik-technologies.com/
Disclaimer:
This offering is not approved or endorsed by OpenCFD Limited, producer and distributor of the OpenFOAM software via www.openfoam.com, and owner of the OPENFOAM and OpenCFD trade marks.