Progressive freeze concentration performance prediction based on polynomial curve model for star fruit juice concentration

Abstract

Progressive freeze concentration (PFC) is a simpler freeze concentration process of removing water content in fruit juice through ice crystal formation in order to concentrate a solution. Vertical finned crystallizer (VFC) was used in the PFC system as the ice crystallizer in this study. A mathematical model is highly needed to be developed so that theories can be validated and to understand the system developed better with minimal risk and cost. Mathematical modelling is also essential to analyze the performance of the system. In this work, the use of mathematical model was explored based on a polynomial regression in analysing and predicting the performance of PFC system. The polynomials curve fitting were first performed to develop the models followed by the simulations to predict the target variables of effective partition constant (K value) and solute recovery (Y value). The relationship of operating parameters including coolant temperature and operation time on the PFC performance values were also discovered via the correlated polynomial regression models. Based on simulations result, the highest efficiencies of PFC process were achieved at approximately - coolant temperature of 10oC and operation time of 55 minutes. To validate the models’ accuracy, the statistical assessment parameters of R-squared and Absolute Average Relative Deviation (AARD) were determined. The findings of this study conferred satisfactory results of the prediction performance of polynomial regression model, in which the least analysis error of AARD (i.e., below 10%) and the highest R-squared (i.e., above 0.97) were successfully achieved. It is concluded that polynomials-based predictive models are promising alternatives to replace time-consuming and expensive experimental evaluation of PFC process for fruit juices.