NUCOR STEEL

By: Mayukh, Sami, Twesha

Quantifying the Value & Saleability of Intermix (Non-standard characteristics) coils

​

NUCOR STEEL

• Largest steel producer in the U.S. and a leader in sustainable manufacturing
• Uses electric arc furnaces (EAF) with recycled scrap to reduce environmental impact
• Decentralized structure allows plants to operate efficiently and adapt quickly
• Strong culture of safety, incentives, and innovation driving consistent growth

STEEL MANUFACTURING & EAF

• Steel is produced in multiple grades (defined by physical and chemical characteristics) to meet customer requirements
• Production requires switching between grades
• EAF:
• melts recycled scrap using electric arcs
• Enables flexible, on-demand production
• Allows faster switching between grades than traditional methods
• Flexible production leads to frequent grade transitions

WHAT ARE INTERMIX COILS

• Non-standard coils produced during transitions between steel grades
• Contain mixed chemical properties, leading to various classifications
• Harder to classify so often sold at a discount or reprocessed

CURRENT TONNAGE BASED APPROACH

Current Approach:

• Penalty is based solely on the amount of intermix tonnage produced during a transition
• Assumes a direct, linear relationship between tonnage and cost
• Treats all intermix material as having equal value regardless of quality
• Simple and easy to implement within existing scheduling systems

PROBLEM

​

• Tonnage-based penalty:
• Treats all intermix tonnage as having equal value
• Does not capture quality differences across transitions
• Ignores variability and consistency of outcomes
• Does not reflect demand or saleability
• Leads to miscalculated penalty of intermix and inefficient decisions

DATA COLLECTION

Dataset provided by Nucor Steel with historical records of intermix coils from grade transitions

The dataset includes information on coil dimensions, production timestamps, intermix coil customer identifiers, and grade transitions.

The variables allow us to analyze patterns in coil production and identify factors that may influence the saleability and quality of intermix coils

DATA OVERVIEW

9 variables · 0 missing values · 6,400 coils

QUALITY & RISK

VARY

ACROSS TRANSITIONS

DEMAND

IS

NOT

UNIFORM

TRANSITIONS

PRODUCE MULTIPLE

COILS

GEOMETRY INSIGHT: DEMAND-DRIVEN PERSPECTIVE

• Sellable coils (Prime + Excess Prime) cluster in specific ranges
• Certain width/thickness combinations appear more frequently
• Geometry aligns with what the market demands
• Patterns vary across grades
• Matching preferred geometry improves saleability

​

WEIGHTED PENALTY MODEL

RESULTS

OUR MODEL REASSIGNS PENALTIES MORE ACCURATELY

NUCOR APP

WHAT ‘S NEXT

• Integrate the model into production scheduling to reduce costly grade transitions
• Expand the Nucor App into a real-time, company-wide decision support system
• Incorporate demand forecasting and market-driven pricing inputs
• Develop a dynamic learning component that continuously updates cost and pricing based on observed outcomes and historical performance

RECOMMENDATIONS

Adopt a data-driven penalty framework (quality, variability, coil count, demand) since tonnage alone mispenalizes transitions and misses true cost drivers

Penalize low quality, high-variability, low-demand, and high coil count transitions more heavily due to increased risk, inefficiency, and operational burden

Use coil count as the primary measure of operational burden, leveraging observed average coils per transition rather than tonnage-based approximations.

CONCLUSION

We identified gaps in Nucor’s tonnage-based approach, which overlooks quality, consistency, and demand in valuing intermix coils

We built an interactive decision-support tool to make the model practical and easy to use

Developed a data-driven penalty framework for more accurate assessments

This project enables better changeover cost estimation, more efficient scheduling, and improved production planning.

THANK

YOU!