DEVELOPMENT OF A DRYING PROCESS CONTROL SYSTEM BASED ON INTELLIGENT MANAGEMENT USING ARTIFICIAL NEURAL NETWORKS

Abstract

Solar dryers are among the environmentally friendly and energy-efficient drying systems designed for effective dehydration of agricultural products under natural conditions. These systems, by utilizing solar energy, optimize the moisture content of products and help preserve their quality. However, the efficiency of solar dryers depends on external factors such as ambient temperature, solar radiation, and airflow, making real-time control of these parameters essential. Traditional PI controllers are not sufficiently effective under such variable conditions due to their long tuning time and signiicant overshoot. In this study, a predictive control system based on artificial neural networks (ANN) was developed to enhance the performance of solar dryers and was compared with a PI-controller-based system. A mathematical model of the drying process was created in the MATLAB R2014a environment using the Simulink software package, and computer simulations were carried out for various control methods. The results showed that the system controlled by the predictive neuro-controller achieved a settling time of 160 seconds, which is 36% faster compared to the PI-controlled system (settling time of 250 seconds). Additionally, the neural control system maintained temperature stability with an accuracy of ±1.2°C, demonstrating significantly higher precision compared to the PI-controller. The results confirm that a control system based on artificial neural networks plays a crucial role in ensuring the stable operation of solar dryers, optimizing energy consumption, and improving product quality. This approach enables the automation of agricultural drying technologies and ensures their environmentally sustainable implementation. The findings indicate promising prospects for the large-scale industrial application of this system.