Synthesis and research of the intelligent automatic control system for a fruit drying apparatus

Modern drying technologies are essential for extending the shelf life of fruits and vegetables while minimizing energy consumption and maintaining product quality. This study presents the development and implementation of an intelligent automatic control system for a conveyor-type fruit drying apparatus, integrating fuzzy logic and artificial neural network (ANN) methodologies. The novelty of the work lies in the first real-time application of a NARMA L2 neural network controller for regulating the air relative humidity in the drying chamber. Ripe fruits (apples, pears, apricots, and plums) with high moisture content were selected and processed under industrial conditions using a G4-KSK-15 conveyor-type dryer. A dynamic mathematical model of the drying process was constructed by approximating transient characteristics via a second-order transfer function. This model served as the foundation for simulating three control strategies in MATLAB/Simulink: classic PID, PID-Fuzzy, and PID-NARMA L2. Experimental validation demonstrated that the NARMA L2 controller significantly outperformed the other methods, achieving a 57% reduction in settling time and a 74% reduction in overshoot compared to the conventional PID regulator. The improved control response directly contributed to reduced energy consumption during the drying phase. This work confirms that intelligent control systems—particularly those incorporating ANN—enhance process stability, product quality, and energy efficiency in industrial fruit drying operations. The findings provide a foundation for broader application of AI-driven control systems in the food industry.