PREDIÇÃO DA NECESSIDADE DE IMPOSIÇÃO DE RESTRIÇÃO TEMPORÁRIA DE VELOCIDADE NA VIA PERMANENTE UTILIZANDO ALGORITMOS DE MACHINE LEARNING

• Ano de defesa: 2021
• Autor(a) principal: Neves, Bruna dos Santos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Mestrado em Engenharia Civil; Centro Tecnológico; UFES; Programa de Pós-Graduação em Engenharia Civil
• Palavras-chave em Português: Restrição temporária de velocidade; ferrovia; XGBoost; Base de dados desbalanceada; manutenção preventiva; subject.br-rjbn; Engenharia Civil
• Link de acesso: http://repositorio.ufes.br/handle/10/14655

In Brazil, rail transport is essential for the country's economy. The railroad is a rail transport system, consisting of permanent way and other fixed installations, traffic equipment and other essential accessories for the safe and efficient handling of passengers and/or cargo. According to the infrastructure and superstructure of the permanent way, the Maximum Authorized Speed (MAS) of a railway is defined on its geometric project. The MAS is the maximum speed of the project for each section of the permanent way, which can be changed by the conditions of the railway. When an anomaly is identified on the track, the maximum train speed in the section that presents an anomaly must be lower than the MAS to ensure the safety of the operation, in which case a Temporary Speed Restriction is imposed. One of the challenges encountered in railway maintenance is to reduce the number of sections of the permanent way with temporary speed restrictions. The restrictions remain on the section until a maintenance team checks and corrects an existing anomaly. Thus, restrictions can have negative impacts in the capacity of the route. A tool that predicts the occurrence of speed restrictions would be fundamental to head the maintenance teams. For the development of this tool, we used data collected from the Vitória - Minas Railway (EFVM) in its entire length and during the period of 1 year and 10 months. By associating the historical data of permanent way restrictions with the historical data of track geometry parameters, the weather conditions and the gross ton on the tracks, a total database was created to make a prediction. These data were used as input for the training of a machine learning algorithm: XGBoost, which aim to predict the need to impose a temporary speed restriction in the following 30 days or 60 days.