Company Description:
This multinational corporation is one of the largest mining operators in Brazil. It specialized in minoring iron ore and nickel. It averages $34B in revenue a year and employees more than 195,000 people worldwide.
Challenge:
A slurry leak is a critical problem in the mining industry with many environmental consequences. Monitoring pipelines with costly sensors take a significant and expensive investment. The company was looking for a different solution that would predict leaks at a much lower cost.
Strategy:
Stefanini started with a full failure analysis of the existing sensors along the pipeline, capitalizing on statistical models. We then developed a dashboard of the automation for operational teams to view data and support their decisions. With the implementation of an open-source analytics platform and powerful machine learning techniques, we were able to model and predict possible leaks, better understand plant behavior, and notify parties of any abnormalities.
Transformation:
We implemented the prediction system in 3 pipelines with a total of 25 prediction models monitoring 24/7. It successfully detected a pipeline leak, preventing environment fines around millions of dollars. The open-source platform helped reduce the overall project cost. Along with cost savings, we improved user experience by providing easy access and control of the data for the client.
See more cases here.
