CHEMICAL ENGINEERING TRANSACTIONS VOL. 35, 2013 A publication of The Italian Association of Chemical Engineering www.aidic.it/cet Guest Editors: Petar Varbanov, Jiří Klemeš, Panos Seferlis, Athanasios I. Papadopoulos, Spyros Voutetakis Copyright © 2013, AIDIC Servizi S.r.l., ISBN 978-88-95608-26-6; ISSN 1974-9791 Uncertainty analysis of industrial fire effects simulation Samia Chettouh*, Rachida Hamzi, Fares Innal, Djamel Haddad Laboratory of Research in Industrial Prevention, Institute of health and industrial safety, University of Batna, Algeria samia.chettouh@yahoo.fr Smoke dispersion prediction systems are becoming increasingly valuable tools in smoke management. Numerical models for dispersion and chemical transport, also known as air quality models, can be used to investigate the fire plume evolution and the smoke impacts (e.g. concentration, temperature). However, all prediction systems include some level of uncertainty, which may occur from the meteorological inputs, diffusion assumptions, plume dynamics, or emission production. Uncertainty analysis enables to avoid as much as possible bad decisions that may have a large impact in a field such as safety. In this study, we are interested in the uncertainty propagation related to NO2 atmospheric dispersion resulting from a crude oil tank fire. Uncertainties were defined a priori in each of the following input parameters: wind speed, pollutant emission rate and its diffusivity coefficient. For that purpose, a Monte Carlo approach has been used. 1. Introduction Due to the complex nature of fire, mathematical prediction models used in fire safety engineering are often simplified and based on a number of assumptions. Even when very sophisticated models are available, a trade-off is often necessary between accuracy, cost and time for design engineers (Lundin, 1999). Many years of research have made it possible to model a wide range of fire phenomena with fire and smoke transport models. Accuracy of results from mathematical models is often complicated by the presence of uncertainties in their inputs data. Therefore, to be used in effective decision making process, the uncertainty in model predictions must be quantified (Refsgaard, 2007). Uncertainty analysis investigates the effects of lack of knowledge and other potential sources of error in the model (e.g., the uncertainty associated with model parameter values) (EPA, 2009). When carried out, uncertainty analysis allows model users to be more informed about the confidence that can be placed in model results and hence becomes a quality insurance factor. Within the framework of industrial fire effects, uncertainties in fuel loads, fuel consumption, and emission factors limit our ability to provide the models with accurate emissions inputs. There are also various other uncertainties in meteorological inputs, and parameters related to modelling of smoke transport and dispersion. In addition, there are uncertainties in chemical reactions and phase transformations (gas to particle and vice versa) during the modelling of ozone and secondary particulate matter formation. In this paper, we study the uncertainty propagation of input parameters of NO2 atmospheric dispersion model on the variation of its output (NO2 concentration). In particular, three input parameters are considered as variables: wind speed, pollutant emission rate and its diffusivity coefficient. Each of them is modeled through a probability density function (pdf). The uncertainty propagation has been conducted using the Monte Carlo sampling. All the results are presented in terms of mean values and confidence interval (lower and upper) bounds. The remainder of this paper is organized as follow. Section 2 is devoted to the presentation of the developed numeric dispersion model. Section 3 gives the general scheme of uncertainty analysis process. Also, therein are given the different probability distributions with respect to the considered uncertain input parameters. Section 4 provides the study results in terms of NO2 plume dispersion and its concentration at a given threshold distance (defined with regard to target elements). Finally, section 5 summarizes our concluding remarks. DOI: 10.3303/CET1335237 Please cite this article as: Chettouh S., Hamzi R., Innal F., Haddad D., 2013, Uncertainty analysis of industrial fire effects simulation, Chemical Engineering Transactions, 35, 1423-1428 DOI:10.3303/CET1335237 1423