Far-infrared drying influence on machine learning algorithms in improving corn drying process with graphene irradiation heating plates

Drying methods often suffer from appropriate drying process modeling, low heating efficiency, and high-energy consumption. However, graphene may provide significant improvements during the drying process due to its ability to save energy and convert electro-thermal energy effectively. In this study, random forest (RF), surface vector machine (SVM), artificial neural network (ANN), and k-nearest neighbor (kNN) were the four machine-learning algorithms employed to analyze the drying process of two types of corn (ZD88 and RP909) at varying temperatures of 35, 45, and 55 degrees C in a graphene far-infrared drying. The study found that the drying process of corn showed a decrease in moisture ratio during the falling rate period from 35 to 55 degrees C. The moisture diffusivity increased from 35 to 55 degrees C with the rise in the temperature, ranging from 2.74 to 4.36 x 10-8 m2/s for ZD88 and 2.05 to 3.07 x 10-8 m2/s for RP909. The ZD88 showed a heat efficiency of 81.62% at 55 degrees C. The results of the machine learning algorithms revealed that 1-6 ANN was the best architecture. Normalized PUK was the best for SVM filters. A k = 3 and s-fold = 5 were the best values for k-NN and RF, respectively. Among the computed algorithms, the tested data for normalized Pearson universal kernel SVM filters proved to be the most effective in computing the process of drying corn. Overall, a graphene far-infrared dryer showed improved heating efficiency and less energy consumption, providing a new concept for developing a drying process through machine learning algorithms strategies.Practical applicationsDrying methods often suffer from appropriate drying process modeling, low heating efficiency, and high-energy consumption. This study provides significant improvements due to the ability of graphene irradiation heating plates to save energy and convert electro-thermal energy effectively. The study achieved a heating efficiency of 81.62% at an infrared temperature of 55 degrees C. Machine learning algorithms are intelligent models that minimize the limitations in describing the drying process of agricultural materials. The study found the normalized Pearson universal kernel was the most effective computation SVM filter. Therefore, using a graphene far-infrared dryer showed a better drying alternative for industrial applications, as it has the potential to improve heating efficiency and consume less energy, providing a new concept for developing drying equipment through machine learning algorithms modeling strategies. The normalized Pearson universal kernel was the most effective computation surface vector machine (SVM) filter for predicting the drying process of corn. The SVM model proved to be the most optimal in modeling the drying process of corn compared to the other finding model. image