Drying characteristics of blueberries by blanching water combined with ultrasonic pretreatment and convolutional neural network modeling

The effects of ultrasound pretreatment (UP) and ultrasound combined blanching pretreatment (UBP) with five drying methods: hot air drying (HAD), infrared convection drying (ICD), long-wave infrared drying (FID), medium-wave infrared drying (MID), and short-wave infrared drying (SID) on the drying characteristics and quality of blueberries were investigated for the first time. The experimental results showed that both pretreatment methods promoted the drying process of blueberries. At the same drying temperature, the drying time was reduced by 10 %–40 %, and the specific energy consumption was reduced by more than 15 %. After UP and UBP, the D eff of blueberries increased by more than 20 % and 35 %, respectively, and the rehydration rate increased by more than 18 % and 30 %, respectively. The tissue structure of blueberries dried by infrared method was severely damaged after 4 h drying, resulting in a fast drying rate and a short drying time. A convolutional neural network model (CNN-LSTM) with superior performance was proposed to predict the changes in moisture ratio during blueberry drying under different drying conditions. These findings provide a theoretical foundation for optimizing industrial blueberry drying, with future work focusing on scaling pretreatment protocols and integrating real-time AI-driven process control to enhance efficiency and product quality. • UBP cuts drying time and energy use by up to 41.67 % and 42.38 %, respectively. • UBP boosts rehydration rate and water diffusion by 30.2 % and 39.28 %, respectively. • Short-wave infrared drying achieves fastest rates and highest rehydration rate. • Infrared convection drying best preserves color in dried blueberries. • CNN-LSTM model predicts moisture ratio variation with high accuracy.