Application of Even Span Greenhouse Solar Dryer (ESGSD) for drying Persian shallot; Kinetic analysis, machine learning modeling and quality evaluation

Abstract

This paper presents an experimental analysis of the Even Span Greenhouse Solar Dryer (ESGSD) for drying Persian shallot. The results are compared with three different oven drying methods at three temperatures including: 70 °C, 60 °C and 50 °C. Additionally, the study includes color analysis, kinetic modeling and mass prediction of dried shallot using Gaussian Process Regression (GPR) model. The results showed that for the oven drying methods at 50 °C and 60 °C, the Modified Midilli model had the highest accuracy, while for 70 °C the Hasibuan and Daud was the most accurate model. The RMSE indexes were 0.024, 0.014 and 0.027 for the oven methods at 50 °C, 60 °C and 70 °C, respectively. For Solar 1 and Solar 2 (both in ESGSD), the Two-term and Hasibuan and Daud models were the best, both with an RMSE of 0.019. Color analysis indicated no significant difference in the ΔE index between the oven method at 50 °C and ESGSD drying. However, significant differences were observed for the oven methods at 60 °C and 70 °C compared to other methods. The GPR model, using the K-fold cross-validation, showed MAPE of 2.77 %, 3.91 % and 3.34 % for training, testing and total phases.

Keywords

• Sustainability
• Gaussian process regression (GPR)
• Color analysis
• Modeling