Skip to main content
The ARC Training Centre for
Transforming Maintenance through Data Science
Publications

Developing and evaluating predictive conveyor belt wear models

Callum Webb, Joanna Sikorska, Ramzan Nazim Khan and Melinda Hodkiewicz

Conveyor belts are critical components in global supply chains in mining, power, and manufacturing industries.

The belt is often the most expensive component of a conveyor system and downtime is costly. Predicting belt wear rate from conveyor design and operational parameters is useful because it allows operators to accurately

forecast belt replacements on new conveyors, estimate wear rate on conveyors without adequate thickness data, and improve their understanding of how different variables influence or relate to belt wear. Our work demonstrates how such predictive models can be developed and evaluated.


Publication: Data-Centric Engineering
DOI: 10.1017/dce.2020.1