Intelligent analysis system and automatic conference of train formations at EFVM using computation vision

Autores

  • Diogo Henrique da Silva IFES - Instituto Federal de Educação, Ciência e Tecnologia do Espírito Santo
  • Gustavo Maia de Almeida IFES - Instituto Federal de Educação, Ciência e Tecnologia do Espírito Santo

DOI:

https://doi.org/10.55592/cilamce.v6i06.10156

Palavras-chave:

Railway, Computer Vision, Conference

Resumo

Every wagon and locomotive has an alphanumeric code on each of its longitudinal faces that identifies and distinguishes it from any other railway vehicle on the Brazilian network. The train formation process consists of joining one or more sets of wagons and locomotives, which will be used to transport ore, and is an inherent step in the railway logistics process, where resources (wagons and locomotives) are joined or separated due to physical limitations of ore loading terminals. At the end of a formation we have a sequence of vehicles ordered and sequenced by their codes, and from this we are able to identify the characteristics of the ore being transported. The activity of forming trains is very important in railway operations, and requires attention and care, as they involve operational, fiscal and commercial risks, which, if not mitigated, can bring irreparable negative results. Such results range from fines imposed by government agencies and even potentially catastrophic accidents, thus putting the lives of employees, communities, partners and the environment at risk. As an example, we have the possibility of an ore composition being delivered with the wrong wagon sequence to a customer, and it entering the production line of a steel mill, which could cause an explosion due to the use of a type of ore wrongly added to a mixture in the blast furnace. Faced with such high risks, the activity of checking the trains after the maneuvers to separate and join the wagons becomes even more important. Making a cut in the Iron Ore circuit on the Vitória a Minas Railway, around 7,800 wagons are allocated daily per train, making checking all formations/compositions a very laborious task. The present work consists of the development of an intelligent system, based on Computer Vision, which automatically checks the alphanumeric codes of train compositions, aiming to guarantee the correct sequence of wagons in each formation of ore trains in the Tubarão yard in Vitória a Minas Railway.

Downloads

Publicado

2024-12-02