Original title: Application of advanced statistical methods in assessment of the late phase of a nuclear accident
Translated title: Aplikace pokročilých statistických metod v odhadu pozdních fázý nukleárních nehod
Authors: Hofman, Radek
Document type: Papers
Conference/Event: Dny radiacni ochrany /30/., Liptovsky Jan (SK), 2008-11-10 / 2008-11-14
Year: 2008
Language: eng
Abstract: [eng] [cze]
The paper presents a new methodology for improving of estimates of radiological situation on terrain in the late phase of a nuclear accident. Methods of Bayesian filtering are applied to the problem. The estimates are based on combination of modeled and measured data provided by responsible authorities. Exploiting information on uncertainty of both the data sources, we are able to produce improved estimate of the true situation on terrain. We also attempt to account for model error, which is unknown and plays crucial role in accuracy of the estimates. The main contribution of this paper is application of an approach based on advanced statistical methods, which allows for estimating of model error covariance structure upon measurements. Model error is estimated on basis of measured-minus-observed residuals evaluated upon measured and modeled values. The methodology is demonstrated on a sample scenario with simulated measurements. Článek presentuje novou metodologii pro zlepšení předpovídání radiačních situací v terénu v pozdních fázích nukleárních havárií.
Keywords: radiation protection
Project no.: CEZ:AV0Z10750506 (CEP), GA102/07/1596 (CEP)
Funding provider: GA ČR
Host item entry: Dny radiační ochrany /30/

Institution: Institute of Information Theory and Automation AS ČR (web)
Document availability information: Fulltext is available at external website.
External URL: http://library.utia.cas.cz/separaty/2008/AS/hofman-application of advanced statistical methods in assessment of the late phase of a nuclear accident.pdf
Original record: http://hdl.handle.net/11104/0165112

Permalink: http://www.nusl.cz/ntk/nusl-39353


The record appears in these collections:
Research > Institutes ASCR > Institute of Information Theory and Automation
Conference materials > Papers