Effects of different drying methods on Camellia sinensis: Investigation of quality parameters and drying kinetics using artificial neural networks

This study aimed to compare the drying kinetics and quality outcomes of tea leaves subjected to four different drying methods—freeze drying (FD), hot air drying (HAD), infrared drying (ID), and microwave drying (MWD). Six thin-layer drying models (Alibas, Demir et al. Henderson & Pabis, Improved Midilli-Kucuk, Logarithmic, and Weibull) were fitted to the experimental data. Artificial neural network (ANN) models were also developed to predict the dimensionless moisture ratio (MR) using drying time and process parameters as inputs. The ANN model showed high prediction performance, with R² values reaching up to 0.9999. In addition, the ANN model achieved strong generalization performance, with Rc² = 0.9967, Rp² = 0.9132, and RPD = 3.3936, confirming its excellent predictive ability. Quality assessments revealed that FD preserved the highest antioxidant capacity (up to 94.7 ± 0.1 %), followed by MWD, HAD, and ID. The lowest water activity, enhancing shelf life, was observed in FD (0.29 ± 0.01 to 0.34 ± 0.01), while MWD showed the highest (0.41 ± 0.04 to 0.64 ± 0.01). Color analysis indicated the least change in FD and the most in ID. Overall, FD produced the highest quality tea, while MWD offered faster drying. ANN models effectively captured nonlinear drying behaviors. This integrated modeling and evaluation approach can support future optimization and quality control strategies in tea drying processes. Although the unified ANN yielded high accuracy (ALL R = 0.9999), model generalization is presently limited to laboratory-scale trials on a single tea cultivar. Further validation on industrial dryers and diverse leaf grades is required, and the ‘black-box’ nature of ANNs complicates direct physico-chemical interpretation. This is the first known study to integrate both artificial neural network (ANN) and mathematical modeling approaches to comprehensively assess the drying kinetics and quality attributes of tea leaves subjected to four different drying methods.