Data-Driven Modeling and Process Optimization of Osmotic Dehydration in Acid-Pretreated Frozen Blueberries
摘要
The osmotic dehydration (OD) of blueberries is often limited by their waxy skin and low permeability, resulting in slow mass transfer, variable texture, and quality loss. Pretreatments such as organic acids and process conditions (temperature and solution ratio) have been proposed to overcome these barriers, yet their combined effects on frozen wild blueberries remain insufficiently understood. This study therefore investigated how citric and ascorbic acids alone and in combination, temperature, and fruit:osmotic solution influence OD performance in 60°Brix sucrose. Key responses included mass transfer (water loss, solid gain, and total soluble solids), color, texture, phytochemicals, and microstructure. A significant interaction among temperature, ratio, and pretreatment (p < 0.01) was observed for several parameters, notably texture. The greatest decline in solution total soluble solids occurred within the first 1–2 h, highlighting rapid early diffusion. Combined citric + ascorbic acid at 1:4 produced the highest dehydration efficiency, whereas higher ratios (1:7 and 1:10) and non-pretreated samples showed reduced mass transfer. Texture peaked in non-pretreated berries at 60 °C/1:10 (31.3 N mm⁻1), while citric acid alone yielded the lowest stiffness. Color retention was superior with citric acid at room temperature (ΔE < 4), and pigment degradation intensified at 70 °C. Scanning electron microscopy confirmed epidermal disruption and pore formation after acid pretreatment, consistent with improved permeability. Overall, citric acid, alone or combined with ascorbic acid, optimized OD by enhancing mass transfer and maintaining visual quality. Among predictive models, Support Vector Regression (SVR) gave the best individual accuracy (R2 = 0.73 for WL; 0.69 for WR), while the ensemble model achieved the highest overall performance (R2 = 0.93, RMSE = 1.20, MBE = − 0.01).