Possibilities of Using a Virtual Source Model to Refine the Prediction of the Impact of Injurious Effects in Massive Releases of Toxic Gases

Abstract

Massive leaks of toxic substances occur not only during accidents connected with the operation of industrial enterprises, but also during their transfers by the means of transport. These events pose a serious threat to both people and the environment. Particularly dangerous are situations where dense gas clouds are formed after the release of the given substance. These spread very quickly, while they tend to remain at the earth's surface for a relatively long time and flow into various depressions. In a few minutes, the toxic gas can reach a large area, as confirmed by the conclusions from the Jack Rabbit field tests. Knowledge of the behavior of heavy gas, as well as knowledge of events influencing its dispersion in real conditions, thus provide important information needed for effective management of the resulting accident. The key data is the extent of the harmful effects of the given hazardous substance, which can be obtained by simulating the predicted emergency situation using modeling software (e.g. ALOHA). However, the fundamental shortcoming of this approach is that these software usually generate overestimated results. The fact that dangerous concentrations in the real environment do not reach such distances from the source of leakage has been repeatedly proven by past accidents. The reason for this discrepancy is that the computer programs used do not allow to model with sufficient accuracy all the simultaneous physico-chemical processes that are applied during the dispersion of dense gas clouds. This task is far too complicated and only a part of it can be solved with the necessary precision. However, for the needs of emergency planning, these inaccuracies represent a relatively fundamental limitation. One way how to deal with this problem is to use a virtual resource (sometimes called a virtual point) model. It is successfully used in various areas where it is necessary to simplify the otherwise demanding physical modeling of the temporal and spatial distribution of matter emitted from a surface or volume source. The main idea of a virtual source is that the real primary emission source is approximated to an imaginary source that is located at a different location but acts and appears outwardly as if it were a real source of leakage. The key task of this solution is to determine the parameters of this source, because this is the only way to obtain results with significantly higher reliability during the subsequent simulation of the considered accident scenario.