(81bp) An Evaluation of Denser Than Air Dispersion Models

The modeling of denser than air dispersion models constitutes an important aspect of risk analysis in chemical industries, considering the evaluation of production systems, storage and the transport of hazardous materials.

Many dispersion models have been developed for lighter than air materials, but these are not suitable for the modeling and simulation of the dispersion of gases such as ammonia, chlorine or propane, which emphasizes the need to study particular models for this type of substances. Moreover, currently there is no knowledge or undergoing research with the capability of studying the dispersion of denser than air gas clouds in Bogotá D. C. or Colombia. This type of models have no computational tools easily obtained in contexts such as the Latin-American and this is why an instrument based on internationally known models was developed as a joint project with the local government. What is more, most tools are not user friendly and are not practical when trying to perform a simulation.

This tool will facilitate the evaluation of toxic effects because of denser than air gas dispersion, with the aim of offering credible scenarios as a consequence of an accidental emission and thus evaluate the impact on the medium. In addition, the objective is to introduce a public policy for risk analysis and evaluation in Bogotá D. C. and include this type of studies in engineering programs (undergraduate and graduate) nationwide. The development of new simulation tools must aim for computational efficiency and include enough phenomenological elements that allow the construction of believable scenarios. (Crowl 2002) explains that to fully understand the behavior of a heavy gas dispersion, a release incident must first be selected (rupture or break in a pipeline, hole in a tank or pipeline, runaway reaction, etc.). Then, a source model to describe the release incident must be chosen, with which the total quantity of released material, the release rate and the release duration may be calculated. Finally, a proper dispersion model must be selected to describe the denser than air gas dispersion.

From these models the downwind concentration, duration and other parameters may be calculated. Thus, looking to provide Bogotá D. C. with a fully functional computer tool, two dispersion models were programmed as well as seven source models. An initial analysis studying the existing gas dispersion models was performed. Box models, plume models, the model of Britter-McQuaid (Eisner 1989) and the SLAB model (Ermak 1990) were all examined. However, box models and plume models perform too many assumptions and some of these do not predict accurately the physical phenomena describing the dispersion of heavier than air gases. Since the goal was to obtain accurate results describing the effects surrounding gas dispersion, such models were discarded.

The Britter-McQuaid model is based on a set of nomograms which represent the concentration decay in releases from a continuous point source and an instantaneous cubic source, as a function of down-wind distance (C. J. H. van den Bosch 2005). The model consists of empirical correlations between a set of independent variables that determine the gross properties of the dense gas dispersion process. On the other hand, the SLAB model is designed for advanced dispersion studies. It describes the concentration in a cloud or plume by solving a 1-dimensional set of conservation equations for mass, chemical compound concentration, energy and the three components of momentum (C. J. H. van den Bosch 2005). In addition to the Britter-McQuaid and SLAB models, it is known that computational tools such as CFD's can be extremely expensive for countries like Colombia where the budget is limited, but the risks related with the daily operation in chemical industries are basically the same as in other countries. That is why the development of computational tools based on mathematical analysis that would simulate the dispersion of gases by using models like SLAB and Britter-McQuaid, provide a good approach compared to the results that CFD tools offer.

Two different programs based on the mentioned models were created using Java as the programming language. Furthermore, the programs were built along with friendly user interfaces for the data input. However, with the goal of validating such results, a popular CFD model among the petrochemical industry was used. CFD tools solve the compressible Navier-Stokes equations on a three dimensional grid using the finite volume method. The CFD tool has been extensively validated for the dispersion of denser than air gases (Hanna, Hansen et al. 2009) in urban and industrial areas, which is why given a proper use of the tool, appropriate and accurate results predicting gas dispersion may be obtained and used to validate the developed software.

Once the programs were built, a typical scenario was constructed in both of the developed programs and compared with the results obtained with the CFD tool. It was found that even though these three computational tools were based on different models, they all provide an estimation of the consequences of an event simulated with the same characteristics. Finally, the obtained results were overlapped with an ArcGis system to provide the local government not only with the description of the dense gas cloud dispersion, but the number of affected people, effect on infrastructure, etc. The aim is to provide the city a public policy for risk analysis and evaluation.

These results proved that both programs created using a regular computer and knowledge could offer an amazing benefit to make risk analysis in places where they are needed at a really low cost. Also, this project presents itself as a start for the development of additional tools that using differential equations and numerical methods could simulate other dispersion processes, and be the kick-off for many other method improvement projects that would change the way that risk analysis is made in places with limited resources.

References 

C. J. H. van den Bosch, R. A. P. M. W. (2005). Methods for the calculation of physical effects - due to releases of hazardous materials (liquids and gases) - 'Yellow Book'. The Hague, Committee for the Prevention of Disasters. Crowl, D. A. (2002). Chemical Process Safety: Fundamentals with Applications. Upper Saddle River, N.J., Prentice Hall. Eisner, H. S. (1989). "Workbook on the dispersion of dense gases : by R.E. Britter and J. McQuaid, Health and Safety Executive, U.K., 1988, HSE Contract Research Report No. 17/1988, ISBN 071760327-X, iii + 129 pages, £35.00." Journal of Occupational Accidents 11(2): 149-149. Ermak, D. L. (1990). User's Manual for SLAB: An atmospheric dispersion model for denser-than-air releases. Livermore, California, Lawrence Livermore National Laboratory. Gexcon. (2009). "CFD for consequence prediction." from http://www.gexcon.com/index.cfm?id=268677. Hanna, S. R., O. R. Hansen, et al. (2009). "CFD model simulation of dispersion from chlorine railcar releases in industrial and urban areas." Atmospheric Environment 43(2): 262-270.