Experimental Study and Mathematical Modeling of Starch Extraction from Pollen Rice by Alkali Solvent

In this study, leaching during extraction of starch from pollen rice using sodium hydroxide as the solvent was investigated. Leaching experiments were carried out at 20, 30, 40 ºC and 50 ºC and amount of starch extracted from the samples were evaluated at different process duration times. The effect of temperature and time on the concentration of extracted starch has been studied. According to the results, increasing the temperature and time of the leaching increased the amount of starch extracted from the rice, and also the concentration of starch extracted by the multilayer perceptron network is predicted. Three-layer neural network structure 2-13-1 consists of input data and an intermediate layer and an outlet layer with 13 neurons in the middle layer, which is the most suitable structure for optimizing the neural network. The total data regression is 0.99646, which focuses on the 45 ° line of data well-defined. The mean square error (MSE) for the total data is 0.000066047 and the root error square (RMSE) for the total data is 0.0081269.