Optimization and estimation processes of the exergetic parameters for onion drying in a multistage semi-industrial continuous dryer: A Comparative assessment of RSM and ANFIS modeling

Abstract

The aim of this research work is to evaluate the capability of RSM and ANFIS modeling to optimize and predict the exergy performance for onion drying process. The experiments were accomplished in a multi-stage semi-industrial continuous dryer at the ranges of the air temperatures (40-70°C), air velocities (0.50-1.50 m/s) and belt linear speeds (2.5-10.5 mm/s). A central composite design (CCD) in RSM with a second-order polynomial model was adopted to investigate the effects of independent variables on three response parameters: exergy loss (0.0182-0.0555 kJ/s) for minimizing, and exergy efficiency (62.33-89.25%) and exergetic improvement potential rate (0.025-0.061 kJ/s) for maximizing. The Takagi-Sugeno ANFIS model and the hybrid learning algorithm were applied to predict the exergy parameters. After variance statistical analysis, all the models obtained in RSM were significant, as well as all operating variables had a notable effect on all the responses. The results depicted that the highest coefficient of determination by applying ANFIS model for predicting the exergy loss, exergy efficiency and improvement potential rate achieved 0.9962, 0.9985 and 0.9924, respectively while the values of R2 for the prediction of these responses in RSM were 0.9479, 0.9688 and 0.9993, respectively. The corresponding optimal conditions for the combined responses were at the air temperature of 40 ?C, air velocity of 1 m/s and belt linear speed of 10.50 mm/s with the maximum desirability of 1.00, whereas the optimized values for exergy loss, exergy efficiency and improvement potential rate acquired 0.045 kJ/s, 76.756% and 0.025 kJ/s, respectively.