Level up for Mechanica AI: steel, aluminium and chemical production AI successful cases

Alexander Khaytin
We are pleased to share with you some results that our Mechanica AI team had gained through the last year.
Last year Mechanica AI had successfully passed a crucial stage of its development, strengthening its footprint in the steel and aluminum production industries.
Advancing in Steel
Our ferroalloys optimization solution is a software module that prescripts the optimal amount of ferroalloys to add at a given stage of the process. Despite many challenges of uncertainty and fluctuating variables, our machine learning model comes up with optimal decisions based on historical data and has been tested in a full-scale everyday use.
The moded has already helped our client to save about $1M within three months. For the project, the expected annual savings in ferroalloy costs are up to $5M.
The last step to this full-scale integration was challenging, however, it helped us gain critical insights on how to embed the AI into the complex environment of the big steelmaking plant.
Currently, we are providing one of the most mature AI solutions in steelmaking.
Optimization of Ferroalloy use case study:
Developing in Aluminum
In 3 months, Mechanica AI has managed to create and test an AI solution predicting losses caused by magnetohydrodynamic instability in aluminum electrolysis production. The project is now being deployed to full-scale production.

For another client, Mechanica AI has developed a model predicting the electrolytic cell yield loss. The predictive system is able to pinpoint electrolytic cells that are likely to have an efficiency decrease in the next several days. That means an on-site technician can target and prevent possible issues before they affect aluminum output. The project is currently in the factory-wide test
Once the system is fully rolled out at the first chosen plant, the potential effect is estimated to be over $10m in yearly revenue.
Improving Aluminum Yield at RUSAL Plants case study:
Emerging in Chemical Production
Our experience is not limited to steel and aluminum. We developed a family of predictive models for the complex cyclic chemical production process. Our goal is to predict the key parameters of the process, in order to optimize the process parameters settings. In 3 months, the modeling and testing were successfully finished, and the next stage of the solution development is currently under discussion.
Predictive Models for Chemicals Production case study
The market demands many different AI solutions
It is impossible to use a "custom development for any task" approach every time.

That's why our team has been working hard to develop the AI Mill platform.

AI Mill 's goal is to provide its customers with the ability to create, test, deploy, and run an extensive set of predictive models without the need to employ data science or DevOps experts. The beta test of AI Mill will start in March 2020. Further updates on the project will be posted on our blog later this summer.

Interested in how AI can be implemented in your business? Reach out to us to explore the possible optimization solutions for your industry.

More posts you may like