Tong Yuan, Xinjian Du, Ran Li, Na Wang, Guochen Zhang, Qian Zhang

College of Mechanical and Power Engineering, Dalian Ocean University

Abstract：

Fresh kelp has high moisture content, and drying is commonly employed to extend its shelf life. Real-time online detection of moisture content during drying is crucial for optimizing drying processes and enhancing product quality. This paper designs a PLC-based BP neural network moisture content prediction system for kelp drying, achieving real-time acquisition of environmental parameters and accurate moisture content prediction. The system was integrated into a multi-energy kelp drying device for experimental validation. Results indicate that under solar-heat pump drying mode and heat pump drying mode, the R², RMSE, and MAE of the moisture content prediction model were 0.982, 3.659, and 3.403, and 0.978, 3.241, and 3.026, respectively, meeting the requirements for moisture content prediction during kelp drying.

Key Words：

kelp; moisture content prediction; BP neural network; PLC