Prescriptive Optimization as a Core Pillar for Integrated Decision-Making Brazil’s Largest Steel Plants

The Challenge: Increasingly Complex Decisions in a Highly Integrated Environment  The Tubarão unit of ArcelorMittal holds a strategic position within the Brazilian steelmaking system. It is a large scale, highly integrated operation, with multiple production routes, interdependent decisions, and an operational environment where small variations in choices generate significant impacts on cost, energy, productivity, and process stability. 

In this context, ArcelorMittal already operated with a high level of technical and operational maturity. The challenge was not to introduce basic optimization concepts, but rather to answer a more sophisticated question:  How to elevate Decision-making to a new level by integrating raw materials, energy, and process into a single prescriptive model capable of supporting both strategic and recurring plant decisions?  Additionally, complexity increased as:  decisions regarding raw material blends directly impacted energy consumption;   the plant’s energy balance became increasingly critical for efficiency and competitiveness;   different areas needed to converge toward consistent, quantitative decisions aligned with global operational objectives.  

The Solution: An Integrated Prescriptive Model Based on the Real Plant Process  Cassotis was selected to expand its collaboration with ArcelorMittal by developing an advanced mathematical optimization model, specifically designed to represent the operational reality of the Tubarão unit.  The project was built upon three fundamental pillars:  Prescriptive mathematical modeling capable of handling multiple variables, constraints, and nonlinear trade-offs;   Faithful representation of the steelmaking process, incorporating physical, operational, and energy aspects of the plant;   Decision integration, connecting raw materials, energy consumption, and overall operational performance.   More than suggesting lower cost combinations, the model was designed to answer strategic day to day questions of the plant, evaluating scenarios, anticipating impacts, and supporting decisions based on both technical and economic criteria simultaneously.  Implementation: From Model to Operational Decision Making  The project involved deep work in model construction, calibration, and validation, always anchored in real operational data.  Key steps included:  detailed mapping of process constraints and variables;   development of an integrated optimization model for raw materials and energy;   calibration using historical data and validation with plant experts;   extensive testing of operational and strategic scenarios.   The result was a robust, reliable decision environment aligned with the reality of ArcelorMittal Tubarão, capable of supporting recurring decisions with a high level of confidence. 

Results: Beyond Cost Savings — Predictability and Consistency  While economic gains are a fundamental part of the project, the impact goes beyond simple cost reduction.  Cassotis’ solution now provides ArcelorMittal Tubarão with:  greater visibility of trade-offs between cost, energy, and process;   more consistent and reproducible decisions, reducing reliance on isolated analyses;   a unified framework to align different areas around the same assumptions;   a solid foundation for strategic decisions in an increasingly volatile environment.   The model consolidates itself as a central instrument of decision governance, elevating the analytical maturity of the operation.