a) to quantify the spread in results of CFD predictions due to the way in which a CFD code is applied;
b) to quantify the validity (i.e. accuracy) of CFD predictions in large, complex gas dispersion situations.
To achieve these objectives, the proposed programme of work will consider a number of test cases using CFD and wind tunnel modelling, together with an evaluation of the uncertainty in the CFD predictions.
The research will provide better understanding of the accuracy of CFD techniques obtained in practice, and hence information on how to evaluate, with greater confidence, the overall quality of modelling results.
The consequences of major industrial accidents involving releases of toxic and hazardous gases are increasingly being analyzed using computational fluid dynamics (CFD) techniques.
The problem of near-field gas dispersion in the immediate vicinity of buildings is of great importance when the rate of spread and build-up of gas concentrations must be predicted for selected accident scenarios. The availability and capabilities of sophisticated commercial CFD codes and powerful computing workstations has enabled user to tackle increasingly large and complex scenarios. However, the validity of these predictions is generally uncertain for two important reasons :
a) the way in which a CFD code is applied to a specific problem can have a critical impact on the final results;
b) there is a considerable disparity between the scope of current model validation and the complexity of the actual industrial scenarios.
The first point relates closely to the expertise of the CFD user in specifying the "best possible solution" within the usual constraints of time, cost and computer resources, while the second growing extent of CFD usage in industry, there is an urgent need for both these aspects to be examined in depth; that is the aim of this project.
Funding SchemeCSC - Cost-sharing contracts
GU2 5XH Guildford