Hany S. El‐Mesery, Ahmed H. ElMesiry, Zicheng Hu, Amer Ali Mahdi, Xinai Zhang, Mansuur Husein, Tilahun Seyoum Workneh
This study optimised the drying performance and quality of okra slices by using a Microwave-Assisted Convective Drying (MACD) system that was integrated with the Artificial Neural Network (ANN) and Self-Organising Map (SOM) algorithms. Drying experiments were conducted at an air temperature that ranged from 40 to 60 °C, an airflow rate of 0.5–1.5 m/s, and a microwave power of 100–400 W. The ANN model, which was trained on 90 experimental datasets, accurately predicted the physicochemical and bioactive parameters, including the Total Phenolic Content (TPC), the Total Flavonoid Content (TFC), Vitamin C, the rehydration ratio, colour change ( δ E), shrinkage and water activity. The model validation yielded an R 2 = 0.985 and RMSE = 0.012, which outperformed the traditional regression model (R 2 = 0.912). The SOM algorithm identified five optimisation zones, with the optimal zone being characterised by 60 °C, 0.3 m/s, and 400 W, thereby achieving high-quality retention: Vitamin C = 23.9 g/kg, TPC = 14.1 g/kg, TFC = 4.5 g/kg, and rehydration ratio = 4.9. These results demonstrate that a hybrid microwave hot-air heating system and AI modelling significantly enhance the drying efficiency, while preserving the integrity of the nutritional quality. The proposed ANN–SOM framework provides a quantitative and transferrable control strategy for industrial dryers, which enables the intelligent and energy-efficient processing of thermo-sensitive food products. • This study explored the improvement in okra-drying with a hybrid microwave dryer. • ANN was used to predict the optimal process conditions for drying okra slices. • Microwave power, airflow and temperature influenced the drying properties of okra. • Lower temperatures, IR and higher airflows provided optimal conditions. • A hybrid microwave heating system produces high-quality okra.