ABSTRACT
The railway connections between the most economically prominent states located in Southeastern Brazil were merged under the administration of a single company around 1996. In this way, railways constructed between the 1890s and 1980s, located near the mines of Minas Gerais State and connected to São Paulo and Rio de Janeiro’s coastlines were merged mainly to enable transportation of iron ore from reserves of Minas Gerais to the main Brazil ports in São Paulo and Rio de Janeiro. This company has under its administration 631 bridges, from 10 to 130 years old, with different structural arrangements, i.e. a large and heterogeneous railway bridge system. The first bridge management system was implemented around 2015, with the main objective of registering geometry of the bridges. In addition, since 2015 many visual inspections and structural monitoring have been performed, generating a large quantity of data, demanding a new Bridge Management System (BMS). In this new version of the BMS, the system is expected to support the well-planned maintenance by determining bridge condition, predicting future deterioration, and selecting optimal timing. Since the integrated maintenance database was built such as for inventory data and the inspection records and repair/improvement history evaluation data, the number of deteriorated bridges which need repair and/or reinforcement was determined by using several kinds of formula based on fuzzy logic.
