Mining the mine waste data: AI applied to tailings safety

Project ID: 2531bd1624

(You will need this ID for your application)

Research Area(s): Engineering
Artificial Inteligence and Robotics

UCL Lead department: Civil, Environmental and Geomatic Engineering (CEGE)

Project Summary:

Mining for metals and rare earth elements generates 5–7 billion tonnes of waste material called tailings every year. Tailings are finely ground rock particles, similar in size to sand and silt, but with unusual mineralogy and particle shapes, stored in large lagoons behind tailings dams. Globally, there are around 8,000 active and legacy tailings storage facilities, holding an estimated 217 km³ of material, and around 10% show instability signs. These are wrongly named after the ore being retrieved yet tailings properties are more related to the extraction process. Limited understanding of the engineering properties of tailings are the reason of failures that can release destructive flows of sludge, causing catastrophic damage. Since 2011, 39 major failures have been recorded, resulting in 620 fatalities and severe environmental impacts.

You will work with the Geotechnical group and AI experts at UCL, analysing datasets from mining companies (including Vale, one of the world’s largest). You will use AI algorithms to search the database and understand how the extraction processes and initial grading of the ore correlates to the waste material (tailings) and ultimately how these key factors control the mechanical behaviour of tailings. Where data gaps exist, you will perform high quality laboratory tests in our laboratory. At the end, we aim to use AI to allow us to predict the mechanical properties of the tailings knowing the mineralogic properties of the ore and the type of beneficiation process and how it impacts the stability of waste structures.

We seek a highly motivated student with a background in geotechnical engineering or related disciplines and an interest in data analytics and machine learning. Skills in laboratory testing or computational modelling are desirable but not essential. Curiosity, problem-solving ability, and enthusiasm for sustainability and risk reduction are important.