Utilization of measurements, machine learning, and analytical calculation for preventing belt flip over on conveyor belts

• Published in: Measurement : journal of the International Measurement Confederation 2023-08, Vol.218, p.113157, Article 113157
• Main Authors: Rumin, Przemysław, Kotowicz, Janusz, Hogg, Daniel, Zastawna-Rumin, Anna
• Format: Article
• Language: English
• Subjects: Anomaly detection; Belt conveyor; Machine learning; Predictive maintenance
• ISSN: 0263-2241
• DOI: 10.1016/j.measurement.2023.113157

•An innovative algorithm is proposed to prevent belt flip over on belt conveyors.•An novel solution as part of the idea of industry 4.0 and the concept of a digital twin.•Conveyor geometry and measurements were used to optimize the machine learning models.•Based on measurements and simulations the accuracy of the method was demonstrated. The use of information technology in modern industry is becoming increasingly important to ensure the failure-free operation of devices. Innovatively, this article describes methods to predict belt flip over. The article describes the tested installation and the acquisition and processing of a unique data set gathered from hilly and very long belt conveyors. Values are predicted using novel analytical algorithms with a combination of machine learning algorithms. A simulation study is presented to demonstrate the effectiveness of the proposed method. The simulation results and case studies show that the proposed method generates accurate results and can be used to predict belt flip over for large-scale belt conveyors.