Quality changes and shelf-life prediction of far-infrared radiation heating assisted pulsed vacuum dried blueberries by SSA-ELM

Abstract

The effects of storage conditions on the changes in quality attributes of blueberries dried by far-infrared radiation heating assisted pulsed vacuum drying (FIR-PVD) during accelerated storage for 72 days were investigated. The total phenolics (TP), total monomeric anthocyanins (TMA) contents, and antioxidant activities greatly decreased with the increase of storage time. The highest retention rates of TP and TMA of FIR-PVD blueberries were obtained under the dark-vacuum condition after accelerated storage, i.e. 43.71 % and 30.38 %, respectively. Compared to hot air drying (HAD) samples, FIR-PVD samples exhibited slightly higher degradation rates of TP and TMA during storage. Light-air condition markedly promoted the decrease of antioxidant activities of FIR-PVD and HAD blueberries. The L* and b* values of dried samples decreased, while a* values increased slowly with the increase of accelerated storage time. Additionally, salp swarm algorithm (SSA) and extreme learning machine (ELM) were applied to establish a shelf-life prediction model. The modeling results showed that ELM model optimized by SSA (SSA-ELM) can be an effective predictive tool for the shelf life of dried blueberries under different conditions.

Keywords: Blueberries; Extreme learning machine; Physicochemical properties; Pulsed vacuum drying; Salp swarm algorithm; Storage.