Classification of critical heat flux patterns using entropy minimax principle

Abstract

A new approach for classification of the critical heat flux (CHF) patterns is presented in this paper. The patterns are classified according to the parametric trends using the entropy minimax principle which is an information theoretic method of multivariate statistical modeling. As the reference CHF data for application of the methodology, the 1986 AECL-UO CHF Look-Up Table is used. The table represents the averages of CHF values corrected for a 8mm round tube based on the local conditions hypothesis. After pattern classification, the characteristics of resulted patterns are discussed with emphasis on the parametric trends and physical mechanisms.