COMPARISON OF HEAVY GAS DISPERSION MODELS FOR INSTANTANEOUS RELEASES

In recent years several new heavy-gas dispersion codes have been devel oped. Some of these were specifically based on an analysis of the Thor ney Island Heavy Gas Dispersion Trials, others were developed independ ently of these trials. A code-comparison exercise was considered to be a valuable contribution to the validation of these codes, though no d irect comparison with experimental data has been made. Results for the following codes have been obtained: DENZ (a number of versions), CIGA LE3, DEGADIS, DRIFT, EOLE, GASTAR and SLAB. Also, results from the Bri tter and McQuaid Workbook (BMW) have been included. Twenty-five cases have been considered; five release conditions in five meteorological c onditions. All the releases are isothermal and near-instantaneous, wit h an initial gas density twice that of air. The most significant findi ng from this work is that there are still substantial differences betw een model codes. The major differences occur for releases at low wind speed, in Pasquill F stability and with a large roughness length. It i s precisely for these conditions that experimental data are lacking. N o statement can be made regarding which model is 'best', since no comp arison with experimental data has been made. The question of independe nt model evaluation is an important one and is being addressed elsewhe re.