This study systematically revealed the effects of air-impingement drying based on parameter intelligent control (ADPIC) on drying kinetics, internal temperature, moisture migration, microstructure, lignin content, texture, color attributes and moisture content prediction of bamboo shoots. Notably, increasing ADPIC relative humidity (RH) from 20% to 50% enhanced the internal heating rate, reducing the time required to reach the target temperature from 15 min to 5 min. Characterized by the shortest drying time (95 min), the lowest lignin content (5.37 g/100g) and color difference (ΔE) of 16.97, the 5-min high-humidity group simultaneously achieved the highest brittleness (1.71 N) and chewiness (0.55 N). Furthermore, microstructural analysis revealed that microchannel distribution density decreased with elevated ADPIC temperature but increased with prolonged high-humidity phase durations. The backpropagation neural network enhanced by dung beetle optimization algorithm (DBO-BP) model was developed for moisture content prediction and demonstrated superior performance, achieving the determination coefficient (R 2 ) of 0.994 and significantly lower prediction errors (mean absolute error (MAE) and root mean square error (RMSE)) compared to BP neural network. This indicated higher predictive accuracy and robustness throughout the drying process. Therefore, the optimized DBO-BP neural network was recommended for precise moisture content prediction during the ADPIC of bamboo shoots. • Effects of ADPIC parameters on bamboo shoots quality were firstly revealed. • A novel predictive DBO-BP neural network model was developed. • The two-stage ADPIC strategy proved suitable for drying bamboo shoots.
Dai et al. (Fri,) studied this question.