Affiliations: Department of Mathematics, Gauhati University, Guwahati, Assam, India | Department of Statistics, Gauhati University, Guwahati, Assam, India | Health Physics Division, Bhabha Atomic Research Centre, Mumbai, India
Note: [] Corresponding author. Rituparna Chutia, Department of Mathematics, Gauhati University, Guwahati-781014, Assam, India. Tel: +91 98540 27940; E-mail: [email protected]
Abstract: Analytical investigation of any atmospheric dispersion model for nuclear industry is very much essential from the point that the model always provides the knowledge of air quality and it guides the decision maker to play their role in applying the various protective measures to mitigate the consequence of any radiation emergency if at all it occurs. On the basis of the air quality study, health risk of any member of the public or any occupational worker can be assessed. In the context of this air quality study, deterministic analysis does not provide the correct estimate of time integrated air concentration result as an outcome of any atmospheric dispersion model because parametric values of the model under concerned are uncertain. Uncertainty of these parameters being random as well as fuzzy, it is always essential to investigate the model with respect to both the types of the uncertainties viz. (a) aleatory uncertainty due to randomness and (b) epistemic uncertainty due to vagueness or lack of information. This article will explore an approach of uncertainty analysis of atmospheric dispersion wherein uncertainty of some parameters is represented by fuzzy set and some parameters are addressed as probabilistic and their combination is finally represented by fuzzy random variable. Due to this admixture new formalism of computing uncertainty is given a name as imprecise-probability fuzzy uncertainty modelling. New methodology of uncertainty quantification is demonstrated by a case study, in which the concentration of contaminant air during the leakage of ammonia through some industrial facility is selected as the target model.
Keywords: Uncertainty, imprecise probability, cumulative distribution function, membership function, percentile
