Citation

Sarkar C, Abbasi SA. J. Hazard. Mater. 2006; 137(1): 8-30.

Affiliation

Center for Pollution Control and Energy Technology, Pondicherry University, Pondicherry, 605 014, India

Copyright

(Copyright © 2006, Elsevier Publishing)

DOI

unavailable

PMID

unavailable

Abstract

The strategies to prevent accidents from occurring in a process industry, or to minimize the harm if an accident does take place, always revolve around forecasting the likely accidents and their impacts. Based on the likely frequency and severity of the accidents, resources are committed towards preventing the accidents. Nearly all techniques of ranking hazardous units, be it the hazard and operability studies, fault tree analysis, hazard indice, etc. - qualitative as well as quantitative - depend essentially on the assessment of the likely frequency and the likely harm accidents in different units may cause. This fact makes it exceedingly important that forecasting the accidents and their likely impact is done as accurately as possible. In the present study we introduce a new approach to accident forecasting based on the discrete modeling paradigm of cellular automata. In this treatment an accident is modeled as a self-evolving phenomena, the impact of which is strongly influenced by the size, nature, and position of the environmental components which lie in the vicinity of the accident site. The outward propagation of the mass, energy and momentum from the accident epicenter is modeled as a fast diffusion process occurring in discrete space-time coordinates. The quantum of energy and material that would flow into each discrete space element (cell) due to the accidental release is evaluated and the degree of vulnerability posed to the receptors if present in the cell is measured at the end of each time element. This approach is able to effectively take into account the modifications in the flux of energy and material which occur as a result of the heterogeneous environment prevailing between the accident epicenter and the receptor. Consequently, more realistic accident scenarios are generated than possible with the prevailing techniques. The efficacy of the approach has been illustrated with case studies.

Language: en