Background
In 2023, EMerald Geomodelling was hired by VALE S/A, the largest mining company in South America, to accurately characterize the real geological conditions of a planned railway route. The Central-West Integration Railway (FICO) is part of a logistical development policy in Brazil to connect the grain producing hubs of the Central-West to the North-South Railway. The project had three main objectives: using airborne geoscanning to understand large-scale trends in soil conditions; map different types of soils and rocks, both in terms of soil type and mechanical properties; and mitigate risks during the execution of the work.
EMerald Geomodelling was able to classify three different categories of underground materials, differentiating them by their mechanical resistance, which is directly correlated with the ease of excavation. The first category comprised more fragile sediments, which could be handled without the need for complex planning and extra costs. The second category consisted of a combination of smaller sediments and larger rocks, generally posing a more significant challenge, and often resulting in additional expense. The third and final category covered hard rocks that would require blasting to excavate. The materials were also classified according to grain size, with sand, silt and clay being differentiated, as well as rocky material.
How it was done
In June 2023, EMerald Geomodelling conducted airborne geoscanning surveys on three distinct segments of the railway. During the data collection process, EMerald Geomodelling closely supervised the operation, ensuring data quality for its geomodelling applications.
During these field surveys, a total of 365 km of the railway alignment was covered, resulting in the collection of 1911 line kilometers of flight data. In this initial phase of the project, surveys had already been carried out to collect geotechnical data, allowing the application of EMerald Geomodelling's machine learning algorithms. Geophysical models, specifically electrical resistivity models, were then generated and interpreted to create soil models.
Deliveries to the client
