Bioelectrical measurement for sugar recovery of sugarcane prediction using artificial neural network

Abstract

One of the problems in the sugar industry is lack of low cost, simple and accurate measurement techniques for sugar recovery of sugarcane in the field or laboratory. This study investigated the potential using of bioelectrical properties as a non-destructive technique for this purpose. A parallel plate capacitor was developed to measure the bioelectric properties of sugarcane in a lateral and longitudinal position of the samples. Eighteen internode samples from 3 sugarcane varieties were measured within 0.1-10 kHz frequency range of LCR meter and then was analyzed sugar recovery in the laboratory. The result showed that in the lateral position are more capacitive and resistive than the longitudinal position. Artificial neural network (ANN) was developed for prediction of sugar recovery as a function of bioelectrical properties. The best ANN model produces a high accuracy in the lateral bioelectrical measurement position with a correlation coefficient (R) > 0.90 and mean square error (MSE) < 0.05. It showed that the ANN model based on bioelectrical properties had the potential to be developed as a simple technique to predict the sugar recovery of sugarcane.