AI ‐driven optimization of solar drying for cooked rice: Experimental and thermal analysis

In order to predict and optimize forced‐convection solar drying of leftover cooked white rice under actual outdoor conditions, this study offers an AI‐driven framework. A multi‐tray drying chamber, a flat‐plate solar air collector, and Internet of Things‐based sensors for temperature, humidity, irradiance, and airflow make up the experimental setup. A 70/15/15 data split was used to assess eight predictive models, including ensemble, deep learning, and regression architectures. With R 2 = 0.97 and RMSE = 1.20%, the UNet–ConvLSTM model outperformed the others in accurately representing spatiotemporal drying dynamics. Temperature, irradiance, and drying time were found to be the most important affecting factors using explainable AI approaches (SHAP and PDP). Energy reductions of 1.2 kWh per batch and cost savings of ₹405 per batch resulted from operational optimization, which cut drying time by about 18%–22%. The system's ₹21,060 monthly economic gain indicates its viability. By combining AI prediction, XAI interpretability, and process optimization, drying efficiency is greatly increased, energy consumption is decreased, and sustainable waste valorization is supported. A scalable solution for intelligent sun drying systems in circular economy applications is offered by the suggested method.