Name: LUIZ GUILHERME VIANNA FRACALOSSI
Publication date: 14/03/2025
Advisor:
Name |
Role |
|---|---|
| CELSO JOSE MUNARO | Advisor |
Examining board:
| Name |
Role |
|---|---|
| CELSO JOSE MUNARO | Presidente |
| EVANDRO OTTONI TEATINI SALLES | Coorientador |
| MARCO ANTONIO DE SOUZA LEITE CUADROS | Examinador Externo |
| PATRICK MARQUES CIARELLI | Examinador Interno |
Summary: The Continuous Casting process aims to transform liquid steel from the Steel Mill, the previous stage, into a final product, with the focus of this study being steel slabs. The manufacture of parts with dimensions outside the specified limits, especially with regard to length, represents a significant problem, as it can directly compromise customer satisfaction and the production flow of the subsequent stages responsible for processing the steel slabs. In this context, measuring the dimensions of the steel slabs produced is essential to ensure quality control and minimize impacts on the production chain. The use of image-based systems for measuring dimensions is a widely disseminated practice in several sectors. However, its application in industrial environments, characterized by the presence of adverse conditions, such as high levels of dirt, intense heat and the presence of steam, represents a considerable challenge. The requirement for a high degree of precision in this context makes the process even more complex, demanding the development of specific technological solutions to overcome such limitations and ensure the effectiveness of the measurements. This dissertation describes the development of an image-based system for the automatic
measurement of the length of steel plates produced in the Continuous Casting process. For this application, the images of the plates are captured by a camera and processed in an industrial computer, using advanced digital image processing and computer vision techniques developed using the Python language. After data processing, the system makes the information generated by the software available to field operators, allowing production monitoring and control. In addition, the system emits audible and visual alarms whenever abnormalities are detected, contributing to the maintenance of quality and efficiency of the production process. The tests performed on the sample sets show a result of 92% of the measurements within the established quality control limit, demonstrating the technical feasibility of the application.
