Prediction of Toxic Emissions from Chemical Fire and Explosion

Abstract

The prediction of toxic emissions from industrial and warehouse fires and explosions involving reactive chemicals has eluded the hazard analysis community for quite some time. To address this issue, we developed a model called ADORA for the time evolution of toxic emissions and their dispersion in the atmosphere. At each time step, the conservation of mass, energy and momentum are solved while invoking thermochemical equilibrium or a constrained version thereof to determine the species composition in the cloud. During the initial stages of cloud evolution, the temperatures are usually high and thermochemical equilibrium applies. As the cloud cools down later due to air entrainment, the composition is governed by reaction kinetics. We use a computationally efficient approach called “constrained equilibrium” which is essentially an approximate way of accounting for temperature dependent reaction kinetics. In this approach, as air and moisture are entrained into the plume, the species concentrations are updated until the temperature decreases sufficiently to “freeze out” the toxic species of concern. The freeze-out temperatures for the toxic species of greatest concern are determined by examining the temperature dependent reaction kinetic rates. The temperatures below which the kinetics are too slow relative to cloud dynamics are selected as freeze out temperatures. This approach allows us to calculate as a function of time the cloud combustion rate, temperature, species composition, size, rise and travel distance downwind from the release location. Sample calculations for the detonation of explosives and for fires involving several reactive chemicals are given. For the former, the model predictions of major species such as carbon dioxide, carbon monoxide, nitrogen oxides and total non-methane hydrocarbons agree well with the limited available data. The model predicts the time dependent consumption of reactants and formation of reaction intermediates as well as stable end products. The concentration contours for toxic species such as hydrogen fluoride are presented and the trends discussed. The predictions of our model can be used to improve preparedness and emergency response planning in order to minimize the consequences of accidents involving reactive and energetic materials.