Yield loss is spread across melting, casting and rolling, and the cause is often clear only after the scrap or rework figure appears. INSUS connects process, material, quality and equipment data to flag yield-loss risk earlier so operators and engineers can review interventions while material is still recoverable.
01
Yield loss is spread across the production line.
02
The cause is often visible only after the loss.
03
Recommendations stay with operators and engineers.
The problem
The gap between material entering a steel plant and saleable steel leaving it comes from many interacting causes: raw-material variation, chemistry, casting conditions, rolling defects, equipment problems and rework. Monitoring often relies on zone checks, periodic sampling and reactive adjustments. By the time final yield or scrap is visible, the conditions that caused the loss may already be several process steps behind the team.
The shift
Operators and engineers stay responsible for every process change. What changes is whether yield-loss risk is visible from production and quality data before material becomes scrap.
Today
With INSUS
How it works
This fits OptiStudio as quality and process decision support across selected melting, casting or rolling stages. Day one is advisory. Existing control systems stay in charge.
Connect
Read-only links to DCS/SCADA, historians, MES, LIMS and available quality or equipment records across the selected melting, casting or rolling stages.
Understand
Analyse process variables, material conditions, quality results and equipment signals to identify combinations historically associated with scrap, defects or rework on this plant.
Predict
Estimate the likelihood of yield-loss events such as defect-related waste, scrap or rework based on current and recent production conditions.
Advise
Show contributing factors and recommended operating adjustments for operator and engineer review. No automatic writes to process control on day one.
Verify
Log predictions against actual scrap, rework and quality results. Track accuracy openly and refine the model for the selected line or yield-loss mechanism.
Getting started
For steel plants that already hold process historian, MES and quality data for a selected line or yield-loss problem. Delivered through OptiStudio as yield-risk prediction and decision support. Expert Services can help with data assessment and validation. Advisory first. Computer vision is used only where suitable image data already exists.
Related for steel
Yield prediction sits beside slag and emissions work on the same steelmaking operation.
Yield, scrap and rework are often understood only after material has moved through the line. This use case predicts yield-loss risk from process, quality and material data so teams can review interventions earlier.
Across melting, steelmaking, continuous casting, rolling and finishing. A pilot usually focuses on one line or mechanism where business impact and data availability are strongest.
DCS/SCADA, historians, MES, LIMS, material tracking, quality records and available equipment information for the selected process area. No new sensors by default.
No. Day one is advisory. AI provides risk flags, contributing factors and recommendations. Operators and engineers retain decision-making authority.
Only where suitable digital image or video data already exists, for example for surface-defect detection on rolling lines. It is not assumed for every pilot.
Through OptiStudio as quality and process decision support. Expert Services can support data assessment, integration and validation with your production and quality teams.
Models are plant- and line-specific. Prediction quality depends on historical coverage of scrap, rework and quality outcomes. Results are measured against your own baseline, not promised industry percentages.
Steel · Yield loss prediction
Use the production and quality data you already record to see yield-loss risk earlier. Operators and engineers keep the decision.
