Contents lists available at ScienceDirect Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes Two-phase flow bubble detection method applied to natural circulation system using fuzzy image processing R.C. Bueno a, ⁎ , P.H.F. Masotti b , J.F. Justo c , D.A. Andrade b , M.S. Rocha b , W.M. Torres b , R.N. de Mesquita b, ⁎ a Industrial Process Control, Federal Institute of Education, Science and Technology, IFSP/CPI, Av. Mogi das Cruzes, 1501 Parque Suzano, CEP 08673-010 Suzano, SP, Brazil b Nuclear Engineering Center, Instituto de Pesquisas Energéticas e Nucleares, IPEN/CNEN, Av. Professor Lineu Prestes, 2242 Cidade Universitária, CEP 05508-000 São Paulo, Brazil c Department of Electronic Systems Engineering at the Polytechnic School of the University of São Paulo, CEP 05508-900 São Paulo, SP, Brazil GRAPHICAL ABSTRACT ARTICLE INFO Keywords: Two-phase flow Edge detection Fuzzy system Image processing Natural circulation ABSTRACT Natural circulation cooling systems are currently used in new nuclear reactors. Over the last decades, research in these systems has focused in the study of flow and heat transfer parameters. A particular area of interest is the estimation of two-phase flow parameters by image processing and pattern recognition using intelligent pro- cessing. Several methods have been proposed to identify objects of interest in bubbly two-phase images. Edge detection is an important task to estimate flow parameters, in which the bubbles are segmented to obtain several features, such as void fraction, area, and diameter. However, current methods face difficulties in determining those parameters in high bubble-density two-phase flow images. Here, a new edge detection method is proposed to segment bubbles in natural circulation instability images. The new method (Fuzzy Contrast Standard Deviation – FUZCON) uses Fuzzy Logic and image standard deviation estimates of locally measured contrast levels. Images were obtained through an experimental circuit made of glass, which enables imaging flow patterns of natural circulation cycles at ambient pressure. The results indicated important improvements on edge detection effi- ciency for high void fraction estimation on high-density two-phase flow bubble images, when compared to classical detectors, without the need to use smoothing algorithms or human intervention. https://doi.org/10.1016/j.nucengdes.2018.05.026 Received 30 May 2017; Received in revised form 9 May 2018; Accepted 25 May 2018 ⁎ Corresponding authors. E-mail addresses: regiscb@ifsp.edu.br (R.C. Bueno), rnavarro@ipen.br (R.N. de Mesquita). Nuclear Engineering and Design 335 (2018) 255–264 0029-5493/ © 2018 Elsevier B.V. All rights reserved. T