Abstract
Reliable assessment of hazardous distances, such as emissions originating from a battery thermal runaway, depends critically on the numerical modelling of the source term, but also on the modelling of the near and open field. For toxic releases, the effect distances can be of the order of hundreds of meters and accurate modelling of the turbulence properties of the atmospheric boundary layer in the usually-unobstructed open field is required. Integral dispersion models remain widely used due to their computational efficiency and proven performance in open field environments. In this domain, integral models provide relatively accurate solutions that allow for numerous parameter variations and/or scenarios due to the low computational costs. Nevertheless, the dispersion process can strongly depend on the near field of the release, which is usually obstructed. This holds for toxic and flammable releases, where the influence of buildings or installations can significantly increase the entrainment of ambient air. In this domain, dominated by flow separation, wake formation, and recirculation effects that can significantly alter the concentration fields, Computational Fluid Dynamics (CFD) generally provides superior fidelity but requires significantly increased computational effort, compared to integral models.
To address this gap, building wake models have been developed in the past for integral models. In the dispersion model in Gexcon’s consequence modelling software, EFFECTS, a building wake model is incorporated into the integral dispersion model. This coupling enables rapid estimation of wake-driven concentration enhancements without requiring detailed geometric meshing or time intensive CFD simulations, while still capturing the main physical aspects of the problem. In the present work, the implemented building wake model is compared with an experimental data set for dispersion around an isolated building, including additional comparison with CFD data for the same dataset. It is demonstrated that the integral model based approach captures the dominant physical mechanisms governing near field dilution as measured by the experiments. Benchmarking against the CFD simulation shows reasonable agreement in key metrics such as maximum concentration levels, plume size and location, while achieving computational speeds that are several orders of magnitude faster. In addition to the validation experiment, a real world application is presented involving a large-scale double wall storage tank reservoir. In the case of a hazard, the outer container, featuring an open top side, results in an elevated-pool evaporation scenario. The integral building wake model reproduces the main dispersion features which were observed in CFD predictions for this moderately complex geometry, supporting its relevance for practical industrial assessments.
