Designing a software system to monitor and control aggregates crushing process with machine vision
Kokkonen, Jesse (2024)
2024
Tietotekniikan DI-ohjelma - Master's Programme in Information Technology
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2024-08-09
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202406267422
https://urn.fi/URN:NBN:fi:tuni-202406267422
Tiivistelmä
Crushing is a process that produces aggregate materials used widely in the construction of our infrastructure. Although some of the products are subject to precise quality requirements, the industry lacks modern quality measuring techniques. Academic work researching the use of machine vision in aggregates measurement exists, and this thesis aims to instrumentalize this knowledge into a low-cost system that can be used for monitoring and controlling the aggregates crushing and screening process.
A literature review was conducted to discover the best practices for measuring aggregate particles with machine vision. It was found out that best accuracy is achieved with expensive 3Dcameras or laser instruments. The methods for calculating particle properties were instead chosen from research using strictly 2D-cameras. The best compromise for size calculations in this system was found to be best-ft rectangle.
The system itself was based on an earlier prototype version. One of the main goals was to connect the existing system into crusher automation and use the machine vision data as control input to the crusher. The design of this system was approached through design science, and the fnal design was described as a design science artefact.
A modular software architecture consisting of image inference, database, Modbus server, web server, and advanced process control modules was designed as the solution. This design was based on functional and non-functional, interface, and data management requirements derived from user stories from customers and product management.
The conclusion of this thesis is that a low-cost machine vision system can be used to detect changes in aggregate particle size when crusher setting is adjusted. The designed software architecture is a working solution for using machine vision as control input to a crusher or quality monitoring. The control logic could not be tested in a real crushing circuit, but the control is labeled a proof of concept based on testing with crusher simulator environment.
A literature review was conducted to discover the best practices for measuring aggregate particles with machine vision. It was found out that best accuracy is achieved with expensive 3Dcameras or laser instruments. The methods for calculating particle properties were instead chosen from research using strictly 2D-cameras. The best compromise for size calculations in this system was found to be best-ft rectangle.
The system itself was based on an earlier prototype version. One of the main goals was to connect the existing system into crusher automation and use the machine vision data as control input to the crusher. The design of this system was approached through design science, and the fnal design was described as a design science artefact.
A modular software architecture consisting of image inference, database, Modbus server, web server, and advanced process control modules was designed as the solution. This design was based on functional and non-functional, interface, and data management requirements derived from user stories from customers and product management.
The conclusion of this thesis is that a low-cost machine vision system can be used to detect changes in aggregate particle size when crusher setting is adjusted. The designed software architecture is a working solution for using machine vision as control input to a crusher or quality monitoring. The control logic could not be tested in a real crushing circuit, but the control is labeled a proof of concept based on testing with crusher simulator environment.