ANN Modelling and Economic Evaluation of Chimney-Assisted Greenhouse Solar Dryer under No Load Condition
Rajesh Kumar Mishra *
Department of Mechanical Engineering, AKS University, Satna (MP), India.
Shrihar Pandey
Department of Mechanical Engineering, AKS University, Satna (MP), India.
Shivbilas Maurya
Department of Agriculture Engineering and Food Technology, AKS University, Satna (MP), India.
Ankit Bharti
Department of Agriculture Engineering and Food Technology, AKS University, Satna (MP), India.
Vijay Singh
Department of Agriculture Engineering and Food Technology, AKS University, Satna (MP), India.
Ajeet Sarathe
Department of Agriculture Engineering and Food Technology, AKS University, Satna (MP), India.
*Author to whom correspondence should be addressed.
Abstract
The paper focuses on ANN-based prediction of dryer temperature in chimney-assisted greenhouse solar dryers (GSDs) used for crop drying, coupled with economic analysis to evaluate the feasibility for small-scale use. The paper aims to explore the ANN modelling and economic evaluation of the chimney-assisted greenhouse solar dryer under no load conditions. The increasing need for cost-effective and energy-efficient agricultural drying methods has driven innovation in solar drying technologies, with artificial intelligence (AI) playing a key role. This study implemented an Artificial Neural Network (ANN) to model the thermal performance of a chimney-assisted greenhouse solar dryer under no load conditions. A 3-2-1 ANN architecture, trained using the Levenberg-Marquardt algorithm, achieved a high predictive accuracy with a coefficient of determination (R²) of 0.99. Using input variables like ambient temperature, solar irradiance, and relative humidity, the model accurately predicted maximum internal dryer temperatures, 50.4 °C and minimum 30°C during the day, significantly higher than ambient conditions. The ANN model demonstrated reliability through low Mean Squared Error (MSE) values of 0.8107 (training), 0.4597 (validation), and 0.5324 (testing). These findings confirm the ANN’s effectiveness in capturing complex environmental interactions and predicting thermal behaviour with precision. The resulting model supports optimised dryer control for enhanced moisture removal efficiency. Furthermore, an economic analysis showed a short payback period of 8 months, emphasising the financial feasibility of such systems for small-scale farmers. The integration of AI modelling with economic assessment presents a comprehensive framework for advancing solar drying technologies, promoting wider adoption of renewable and intelligent solutions in the agricultural sector. In conclusion, the combined approach of ANN modelling and economic evaluation provides a valuable framework for optimising solar drying systems and promoting sustainable agricultural practices.
Keywords: Artificial Neural Network (ANN), chimney assisted greenhouse solar dryer, agricultural, Mean Squared Error (MSE), artificial intelligence