Modelling vapour transport in indoor environments for improved detection of explosives using dogs

Abstract

Air movement in indoor spaces can be complex due to large regions with no dominant flow direction and low mean velocities. Therefore, vapour released from an explosive indoors would be expected to result in a high degree of temporal and spatial variability in concentration. To improve canine detection capability, specifically training equipment, training methods and concepts of use, the science of vapour signatures in enclosed spaces needs to be improved. Large-eddy simulation has been used to study the vapour field in a benchmark test room. The work provides insight into vapour behaviour within indoor spaces and results have been interpreted in relation to vapour detection using dogs. For the test room, it was shown that vapour concentrations reduce rapidly within a short distance from the source. However, the concentration fluctuations, which occur at frequencies that a dog should be able to detect, can be significantly greater than the mean concentration. Due to the low volatility of many explosives, the vapour they produce will readily partition onto surfaces altering the concentrations in the room. A multi-layer vapour sorption model based on computational fluid dynamics (CFD) was validated. The CFD sorption model and a well-mixed sorption model were applied to the benchmark test room. It was shown, for a moderately high volatility explosive, that absorption had little effect on the well-mixed concentration but could have a significant effect on concentrations in the vicinity of the absorbing surface. When it is not possible/practical to build a CFD model, eddy diffusion models can be used to rapidly predict the spatially resolved concentration field indoors. However, there is uncertainty over the parameter that governs mixing, the eddy diffusion coefficient, De. Work has been carried out to develop a method to predict De for mechanically ventilated, isothermal rooms. It was found that De is a function of the air flow rate, room volume and number of air supply vents only. This will enable eddy diffusion modelling to be used with more confidence in the future to plan canine training experiments or to interpret detection results.

More information

Identifiers

ORCID for Timothy, Graham Foat: orcid.org/0000-0001-7514-9385

ORCID for Zhengtong Xie: orcid.org/0000-0002-8119-7532

Catalogue record

Export record