Neuro-control adjustment of conveyor belt tension

Abstract

Goal.Identification of ways to improve the quality indicators of conveyor operation during their design and construction using modern automation elements, automatic control systems, and the application of neural controller systems for automatic support of optimal technological processes during material transportation. Research methodology. The proposed methodology involves supervisory training of a neural controller system for automatic tensioning of the conveyor belt using the MATLAB software package. Research results. A sequence for building and training an imitation neural control system for regulating (supporting) the optimal tension of the conveyor belt has been developed and investigated using the MATLAB software package. It was established that for systems containing elements with significant time constants, it is advisable to experimentally determine the ratio of the number of training stages to the number of segments during supervisory training of the neural controller. Scientific novelty. The article proposes the construction of an imitation model of a tension control system for a conveyor belt using artificial intelligence and the implementation of a neural controller. This approach allows for preliminary tuning of the developed control system and its utilization in the design phase of similar objects. Practical significance. The sequence for constructing an imitation model of an optimal tension control neural system for a conveyor belt can be applied in the design and investigation of similar mechanisms. The system is adaptable and enhances reliability while maintaining desired performance parameters. The model can be recommended for preliminary determination of neural controller parameters, regulator system tuning, obtaining predicted transient processes, and improving work productivity.