Microwave drying effects on modeling thin-layer drying kinetics, energy efficiency analysis, and physical properties of garlic.

Artificial neural network (ANN) models can reduce energy consumption by optimizing mass and heat transfer parameters. In the present work, an artificial neural network technique was used to model a microwave dryer during the drying process of garlic slices. It also investigated the effects of balancing the heating characteristics on the quality parameters such as flavor strength and allicin content. The results revealed that increasing the microwave power reduces the values of allicin content and flavor strength, but increases with airflow velocity. It was observed that the artificial neural network demonstrates the predictive capability of minimizing energy consumption, recording the lowest mean squared error (MSE = 0.0026) and the maximum correlation coefficient (R = 0.9994). The effective moisture diffusivity ranged from 3.34 to 5.27 × 10⁻¹⁰ m²/s, while the activation energy varied between 10.16 and 19.19 kJ/mol, indicating that moisture diffusion requires considerable energy input. The Midilli et al. equation provides the best fit for garlic slice drying curves under different operating conditions. Specific energy consumption strongly depended on drying time, ranging from 1.45 to 15.71 MJ/kg. However, the drying efficiency of 11.15% was achieved at a microwave power of 300 W and an air velocity of 0.3 m/s. More importantly, the proposed drying approach demonstrates the potential of advanced computational tools to save energy cost and enhance heating uniformity in microwave drying systems.