Modelling and optimization of high-pressure homogenization of not-from-concentrate juice: Achieving better juice quality using sustainable production

The present work optimized high-pressure homogenization (HPH) parameters for not-from-concentrate combined peach and carrot juices, based on a two-step comprehensive model using factor analysis and analytic hierarchy process methods. Treating combined juice with pressures over 200 MPa retained more amounts of the bioactive compounds (carotenoids and polyphenols) than non-homogenization. Nutrition-oriented optimization, with higher judgement weight on nutritional properties, and sense-oriented optimization, with higher weight on sensory properties, were set up. Combined juice (250 MPa, 1 pass and 25 degrees C) had the best quality, based on the nutrition- and sense-oriented models. Back propagation neural network (BPNN) models could predict antioxidant capacities of the combined juice with greater accuracy compared with stepwise linear regression. The relative errors of BPNN prediction model were <= 5%.

Automated summary

The study optimized high-pressure homogenization parameters for not-from-concentrate combined peach and carrot juices using factor analysis and analytic hierarchy process methods. The best quality combined juice was at 250 MPa, 1 pass, and 25 degrees C, based on nutrition- and sense-oriented models. Back propagation neural network (BPNN) models were more accurate in predicting antioxidant capacities of the combined juice compared to stepwise linear regression, with <= 5% relative errors in BPNN prediction model.