Optimization of Sugarcane Process Production using Response Surface Methodology (RSM) and Artificial Neural Networks (ANNs)

Abstract

The research aims are to determine the effect of several independent variables on the quantity of sugarcane juice production (mj) as the dependent variable. The independent variables in this study include the top roller rotation (n), clearance of sugarcane (cs), and clearance between the top roller and the rear roller (c2). The methods used in this research are Response Surface Methodology (RSM) and Artificial Neural Networks (ANNs). As a result of the central composite design, 20 experiments of optimum conditions were obtained at n = 13.49 rpm, cs = 2.68 cm, and c2 = 2.06 cm, with the quantity of sugarcane juice produced being 0.342 kg. Meanwhile, the results obtained with the backpropagation algorithm (n = 15 rpm, cs = 2.8 cm, and c2 = 1.74 cm) showed that the quantity of sugarcane juice produced was 0.350 kg.