An integral model for continuous HF releases

This report describes the development of an integral model for the dispersion of HF clouds, which is part of the work done by Risø in the URAHFREP project. URAHFREP, Understanding dispersion of industrial Releases of Anhydrous Hydrogen Fluoride and the associated Risk to the Environment and People, is a project sponsored by the European Commission under contract ENV4-CT970630.

The main objective has been to model the possible in uence of HF thermodynamics on the dispersion of atmospheric HF clouds. Both negative buoyancy (heavy gas) effects and positive buo yancy effects are possible depending on concentration, humidity and other factors. A main question is under which conditions these effects are strong enough to dominate naturally occurring fluctuations and produce plume lift-off.

The URAHFREP field trials showed only weak signs of positive buoyancy in plumes produced for 0.1 kg/s liquid spray releases. HF can form polymers in the gas phase and it forms highly non-ideal liquid mixtures with water. It is demonstrated that the HF thermodynamics needed for the dispersion model can be described by exact thermodynamical relations.

The treatment is based on the fugacity concept, which is explained in some detail emphasizing the link to measurements. Existing experimental data are scarce and of varying quality. The best data have been selected and analysed in order to obtain properties on the saturation curve. A relatively simple rings{and{chains model for the self-association in the gas phase is proposed, and it is demonstrated that the model is capable of reproducing independent measurements (not used to tune the model), in particular it predicts the enthalpy and the anomalous specific heat of HF very satisfactorily.

Exact relations describing phase equilibria for the water-HF system are set up, and the role of the mixing enthalpy is demonstrated. This is used to derive a simple four parameter model for the mixture. Finally the model is successfully tested against fog chamber experiments. The dispersion model is a more-or-less standard integral model with some additional features.

The ideas and assumptions of integral models is explained and the various scaling regimes for cloud growth are discussed. Re-analyzing the Prairie Grass data set it is found that boundary layer scaling is superior to mixed layer scaling, and hence the height of the boundary layer has no direct impact on the dispersion in the lower part of the boundary layer (including lift-off due to natural convection).

The model is tested against data from the URAHFREP field trials. Reasonable agreement is found. In most cases (all except one) the experiments showed no signs of buoyancy effects which is in agreement with model predictions. In Trial 12 reduced ground level concentrations were observed as well as an elevation of the cloud centroid.

This behaviour is captured by the model. A case study is made in order to determine the conditions necessary for HF induced buoyancy to have an effects on ground level concentrations. The possibility of including added mass is discussed and dynamic equations compatible with the level of complexity of an integral model are derived in the appendix.