Prediction Of Sugar Yield From Sugar Cane Using Process Modeling

ABSTRACTA study was undertaken to developed a model that will be used to predict sugar and theby-products from sugar cane. The model developed from MATLAB was used to predictthe sugar, bagasse, filter cake and molasses yield from sugar cane. The predicted valuesfrom the model were compared to yield data obtained from the production of sugar canefrom the Savannah Sugar Company, Numan, Nigeria for 90 days. The analysis ofvariance (ANOVA) at p ≤ 0.01 was used to determine if there were significant differencein the yield predicted by the model and the measured factory yield. The F-LSD at p ≤0.01 was used to separate the means. The model is validated where there was nosignificant difference between its predicted yield and the factory-obtained yield. Thesugar cane input of 2,150.52 MT was obtained from the Savannah Sugar factory. Thecorresponding imbibitions water pumped into the mixed juice was 673.12 MT. Thepredicted sugar, bagasse, molasses and filter cake yield using the MATLAB model was279.5MT (13%), 1,049.46MT(48%), 111.828MT(5.2%) and 101.1MT(4.7%)respectively. The ANOVA showed that there was no significant difference between theMATLAB model and the factory-based model. It is concluded that the ANOVA validatedMATLAB model for sugar yield prediction. Consequently, this model is recommendedfor use in predicting sugar and by-products yields from sugar cane. Statistical analysisusing the F-LSD showed that the model satisfied such an invariant confidence in themodel, and provided support for its validity.