A review on machine learning assisted solar drying system with phase change material

Abstract

Solar energy is one of the most suitable renewable sources for drying crops and generating heat and electricity, replacing the need for fossil fuels. It provides clean and sustainable energy that benefits the environment. Nanoparticle PCM is used to reduce drying time and enhance the thermal conductivity of solar drying during the drying process. The mathematical model analyzes moisture loss, drying time, and crop quality. Machine learning tools simplify the modeling, design, forecasting, simulation, optimization, and problem detection of any drying system with high accuracy in less time than conventional methods. ML analyzes data from many sensors, such as humidity, temperature, solar radiation, and temperature, to make real-time adjustments to improve product quality and efficiency. Radial Basis function (RBF), Multilayered Perceptron (MPL), Artificial Neural Network (ANN), and Support Vector Machine (SVM) are algorithms used in ML to control and optimize the drying process in solar dryers are cost-effective, reduce the time required for accurate results, and enhance the energy efficiency, product quality, and overall sustainability in solar drying. This review paper shows that an ML-based system makes the dryer more sustainable and economical. RBF's lowest RMSE (0.99 and 33.67 for dried temperature and mass) provided better accuracy. The study found a net annual benefit of $ 3000 for the ML-based system and $ 1000 for the conventional system. The experimental value of MR and ANN forecasts nearly matched, with a maximum variation of 0.001, and experimental DR decreased by 0.02 d.b/h, with the variance predicted by ANN being 0.04.