<![CDATA[Agricultural drying strongly affects post-harvest product quality and storage stability. This study develops a smart solar drying system integrating multi-parameter sensing, electrical storage, water-based thermal storage, an Industrial IoT V-box, and a Siemens S7-1200 PLC. Temperature, relative humidity, solar irradiance, airflow, and COâ‚‚ concentration were monitored to characterize the drying environment. A hierarchical Random Forest framework was developed to estimate the chamber temperature, relative humidity, and coffee moisture content. Leave-one-day-out validation yielded pooled R2 values of 0.909 and 0.890 for indoor temperature and relative humidity, respectively. The moisture model achieved R2=0.882, MAE = 0.472 percentage points, and RMSE = 0.740 percentage points against the sensor reference. The Random Forest models were executed on the V-box, while the PLC acquired sensor data and controlled the actuators. The results demonstrate the feasibility of integrating data-driven estimation with PLC-based supervision for solar coffee drying.]]>
Copyrights © 2026
