Prediction of Apple Slices Drying Kinetic during Infrared-Assisted-Hot Air Drying by Deep Neural Networks

The effects of temperature, air velocity, and infrared radiation distances on the drying characteristics and quality of apple slices were investigated using infrared-assisted-hot air drying (IRAHAD). Drying temperature and air velocity had remarkable effects on the drying kinetics, color, total phenol content, total flavonoid content, and vitamin C content (VCC) of apple slices. Infrared radiation distance demonstrated similar results, other than for VCC and color. The shortest drying time was obtained at 70 degrees C, air velocity of 3 m/s and infrared radiation distance of 10 cm. A deep neural network (DNN) was developed, based on 4526 groups of apple slice drying data, and was applied to predict changes in moisture ratio (MR) and dry basis moisture content (DBMC) of apple slices during drying. DNN predicted that the coefficient of determination (R-2) was 0.9975 and 1.0000, and the mean absolute error (MAE) was 0.001100 and 0.000127, for MR and DBMC, respectively. Furthermore, DNN obtained the highest R-2 and lowest MAE values when compared with multilayer perceptron (MLP) and support vector regression (SVR). Therefore, DNN can provide new ideas for the rapid detection of apple moisture and guide apple processing in order to improve quality and intelligent control in the drying process.

Automated summary

The effects of temperature, air velocity, and infrared radiation distances on the drying characteristics and quality of apple slices were investigated. A deep neural network (DNN) was developed to predict changes in moisture ratio and dry basis moisture content of apple slices during drying. DNN outperformed other modeling methods in terms of accuracy and can provide new ideas for rapid detection and intelligent control of apple moisture in the drying process.