REACTPOOL: a code implementing a new multi-compound pool model that accounts for chemical reactions and changing...

All chemicals that react violently with water or in contact with water liberate toxic gas are included in the list of substances covered by the majority of the international legislation on major hazards. This category includes a large number of chemicals that are used widely in the process industries. A survey of accidents that occurred in the last 10 years in the USA shows numerous major incidents that involved spillages of these substances. Even so, there are almost no experimental data on the behaviour of these chemicals on release. Furthermore, there are very few published studies on modelling the behaviour of such spillages, except in the case of hydrogen fluoride. In previous work we reported a new theoretical model [J. Haz. Mat. 62 (1998) 101-129, J. Haz. Mat. 62 (1998) 131-142, J. Haz. Mat. A67 (1999) 9-40], that describes accidental spills of SO(3) and oleum, which are substances with very complex behaviour that belong to this category. It describes both the pool [J. Haz. Mat. 62 (1998) 101-129, J. Haz. Mat. 62 (1998) 131-142] and the cloud behaviour [J. Haz. Mat. A67 (1999) 9-40]. In the work reported here the pool model was modified in a generic form in order to include other water reactive chemicals. REACTPOOL is a new code that can be used for both instantaneous and continuous liquid releases under a wide range of input parameters (steady or varying). It can be used for all liquids irrespective of their volatility and reactivity, and it describes pools consisting of more than one liquid that can have changing composition and properties. The purpose of this paper is to present the general procedure followed in REACTPOOL and to show how the new model has been modified and implemented for substances other than SO(3) and oleum. The modelling procedure has been implemented in a computer code written in Visual Basic, and results of the model have been generated using this code. It should be noted that this model requires validation data, but that the availability of such data awaits the performance of suitable experimental investigations.