Study on drying characteristics and moisture content prediction model of Panax notoginseng taproot by using segmented drying of microwave vacuum combined with hot air

Abstract

AbstractIn order to explore the drying characteristics of Panax notoginseng taproot (PNT) and realize the dynamic monitoring of moisture content in the segmented drying process of microwave vacuum combined with hot air, the PNT was conducted the experimental study under different convert moisture contents of wet basis (20, 30, 40 g•g−1), upper‐temperature limits (45, 50, 55°C), power densities (0.50, 0.75, 1.00, 1.50 W•g−1) and chamber pressures (1, 3, 5, 10 kPa). The dynamic changes in moisture ratio and drying rate under different drying conditions were studied, and a snake optimizer‐support vector regression (SO‐SVR) model was proposed to predict moisture content during the drying process. The results showed that the drying process mainly consisted of decreasing rate and constant rate, and the stage of constant rate occurred when the moisture content of the dry basis decreased from 1 to 0.5 g•g−1. In the stage of microwave vacuum, decreasing chamber pressure, increasing upper‐temperature limit and power density could increase the drying rate. In order to verify the prediction ability of SO‐SVR, this study carried out a comparative analysis between SO‐SVR and two famous prediction models (genetic algorithm‐support vector regression and particle swarm optimization‐support vector regression) through three evaluation indexes (coefficient of determination (R2), mean absolute error (MAE) and root mean square error (RMSE). The results showed that the prediction accuracy of the SO‐SVR model was the highest, and the R2, MAE, and RMSE of the prediction set were 0.9850, 0.0140, and 0.0173, respectively. Therefore, the SO‐SVR model is a feasible method to predict the moisture content of PNT during the segmented drying process of microwave vacuum combined with hot air.Practical ApplicationIn the condition of segmented drying of microwave vacuum combined with hot air, the drying technology and characteristics of PNT were explored and proposed a SO‐SVR model to dynamically monitor the moisture content. This method has important theoretical significance and practical value to improve the processing efficiency and quality in the process of Panax notoginseng production.