Indirect Monitoring Cane Sugar Crystallization via Image Fractal Analysis

Abstract

Due to the increasing demands of quality products, efficient monitoring systems in the current control and operation of industrial processes are essentials. However, in particulate processes as cane sugar crystallization, accurate, inexpensive and suitable sensors for the online monitoring of key process variables are not available. In this work, an alternative using the image analysis of micrographs captured in batch cooling crystallizer is presented. The propose is based on a combined treatment between fractal analysis and conventional binarization techniques, obtaining a normalized fractal index (NFI) that allow the dynamic monitoring of crystal mean diameter, D(4,3). In order to evaluate the monitoring system, the crystallizer was operated at different cooling profiles, finding that the methodology proposed can be used as an alternative technique, inexpensive and easy to implement, for monitoring crystal growth.