GENETIC PROGRAMMING-BASED PREDICTIVE MODELLING OF MOISTURE RATIO IN DRYING OF GRAPES USING DESICCANT ROTARY DRYER

Abstract

In this study predictive models are developed for moisture ratio in the drying process of grapes using a desiccant rotary dryer. Response Surface Methodology (RSM) and Genetic Programming (GP) are employed to capture the relationship between critical drying parameters—temperature, airflow velocity and time—and the moisture ratio. The RSM model demonstrated high accuracy with a correlation coefficient of 0.992, whereas the GP model achieved a slightly lower correlation coefficient of 0.983. However, GP offered a simpler and interpretable structure. Comparative analysis reveals that both the models are in close proximity of experimental data and thus, are suitable for predicting drying parameters in food processing. This study highlights the effectiveness of using GP to enhance efficiency in grape drying— with implications for broader food dehydration applications.