A fuzzy inference system for two-phase flow regime identification from radiography images

Abstract

Swiftly identifying the two-phase flow that occurs in coolant channels is crucial for monitoring energy producing installations such as boiling water reactors. In this piece of research, a Sugeno-type fuzzy inference system is implemented for online, non-invasive flow regime identification. The proposed system is predominantly efficient in its construction and operation: a single directly computable input is employed and as many fuzzy inference outputs and rules are used as there are flow regimes to be identified. Noninvasiveness is accomplished through the utilisation of radiography images. Compactness notwithstanding, the fuzzy inference system successfully and reliably identifies the flow regime of sequences of frames from neutron radiography videos.