Feasibility Study of Neural Network-based Classification of Conveyor Belt Damage Using Partial DiagBelt Data

• Published in: Civil and environmental engineering reports Vol. 35; no. 3; pp. 313 - 325
• Main Authors: Rzeszowska, Aleksandra, Jurdziak, Leszek, Błażej, Ryszard
• Format: Journal Article
• Language: English
• Published: 28.07.2025
• ISSN: 2080-5187; 2450-8594; 2450-8594
• DOI: 10.59440/ceer/205939

Non-invasive methods for diagnosing conveyor belts enable effective detection of damage, significantly reducing the costs associated with belt replacement. Additionally, they allow for continuous monitoring of the belts’ technical condition and degree of wear over extended periods of operation. Such solutions also enhance safety in environments where conveyor systems are used. While belt wear is an inevitable process, its rate can vary depending on specific operational conditions, such as the conveyor’s location, its length, the type of material being transported, and the belt’s operating speed. This article discusses an artificial intelligence-based approach to classifying conveyor belt damage. A two-layer neural network was implemented in the MATLAB environment using the Deep Learning Toolbox. By optimizing the network, a high level of operational efficiency was achieved, reaching an accuracy range of 80–90%. This solution opens new possibilities for precise diagnostics and monitoring of conveyor belts’ technical state, contributing to improved durability and reduced operational costs.