FRACTAL ANALYSIS OF TEMPERATURE TIME SERIES FROM BATCH SUGARCANE CRYSTALLIZATION

Batch crystallization is an important process in many industries, for example, fine chemicals, foods, and pharmaceuticals. Monitoring of the main process variables is essential for process understanding, diagnosis, and for product quality control. It is known that temperature has a critical effect on crystallization. Temperature measurements from crystallization systems display fluctuations with apparently random and complex behavior. Fractal analysis of complex time series has received significant attention in the last few years due to its capability for extraction of hidden useful information of the underlying phenomena behind the time-series complexity. In this work, the potential of fractal analysis of time series for diagnosis of industrial crystallization processes is investigated using temperature measurements from a typical batch sugarcane crystallization system. The crystallizer was operated at different cooling profiles, finding that fractal index is directly related to crystal mean diameter dynamics. Thus, we establish that fractal analysis is a simple and robust alternative for the characterization of batch crystallization.