Open AccessEVALUATION OF CANE SUGAR PRODUCTION USING MULTIVARIATE STATISTICAL METHODS
Author(s) -
Bruno José Chiaramonte de Castro,
André Bernardo
Publication year - 2019
Publication title -
the journal of engineering and exact sciences
Language(s) - English
Resource type - Journals
ISSN - 2527-1075
DOI - 10.18540/jcecvl5iss3pp0228-0237
Subject(s) - sugar , principal component analysis , multivariate statistics , partial least squares regression , cane , ethanol fuel , principal component regression , mathematics , sugar cane , starch , multivariate analysis , statistics , production (economics) , sugar production , food science , microbiology and biotechnology , environmental science , chemistry , agricultural science , biology , economics , fermentation , macroeconomics
In sugarcane industries, process monitoring has the main purpose of maximizing sugar and ethanol production, meeting the quality parameters demanded by customers. The aim of this work was to identify industrial process variables that presented the greatest impacts on the quantity and quality of the produced sugar, by applying principal component analysis (PCA) and partial least squares regression (PLS) to the process data of a sugar and ethanol industry. The PCA correlation matrix highlighted the correlation between the presence of alcoholic flocs in sugar and the concentrations of starch and dextran in it. Both PCA and PLS showed that the color of the sugar was highly correlated to its moisture content. The first three principal components accounted for 40.92% of the total data variability.