EQUIP YOUR ENGINEERS WITH THE POWER OF MACHINE LEARNING
Made with steel sector and needs in mind
Run quick (<1 min) evaluations before training full fledged models
Train models with years of data in minutes on an 8-core CPU
Apply to any numerical/classification problem. Train based on customized metrics (R2, MSE, etc)
Use trained models for on-line prediction within QMOS suite
Create, store and schedule a large array of models
Alloy Calculator & Lab Assistant
Streamline melt shop workflows from liquid steel refining to finished product quality. QMOS Data Science helps operators and lab technicians make faster, data-driven decisions as chemistry samples become available. It supports chemistry-based and machine-learning alloy optimization, physical quality targets, what-if regrading with MES specification retrieval, transition chemistry fine-tuning, and real-time QMOS feedback.
Utilize our product/location matching algorithm to streamline your product warehouse. Reduce forklift/crane travel time and get material out of the plant and onto the road faster. Our approach can generate simple best practice instructions all the way to detailed location management based on a daily production schedule.
In addition to running predictive analytics on your local network, you can harness the processing power of Google Compute Engine. Dramatically reduce training time and access your models anywhere with an internet connection. Feel at ease with the latest authentication and authorization protocols in place. Your model data will be securely encrypted and stored on Google Cloud.
Integrate your best models in QMOS workflow for real-time prediction. View accurate physicals predictions for each chemistry sample directly in QMOS screens and disposition your heats accordingly. Set up automatic retraining of models to guarantee ongoing accuracy.
Scrap Planner
Optimize scrap recipes using chemistry concentrations, scrap constraints, inventory levels, residual elements, and the upcoming melt schedule.
• Creates optimized cost-per-ton recipes
• Adjusts chemistry using current production results
• Supports multiple charges and layered charge components
• Identifies missing commodities when inventory is limited
Scrap Vision –Density
Use image recognition to support scrap management during truck check-in. Integrated with QMOS Scrap Scale, the AI model detects truck container types and volumes, estimates scrap volume, and calculates scrap density.
- Works with existing scale cameras
- No additional hardware required
- Alerts operators when further investigation is needed
- Models can be maintained by plant personnel as data changes
Proven Melt Shop Results
QMOS Data Science helps melt shops turn production data into measurable improvements across alloy optimization, scrap planning, and scrap density tracking.
• Cut ladle refining time by 2 minutes and added 300 heats/year
• Reduced alloy cost by $2.9/ton
• Regraded 4 heats/month to avoid downgrades or scrap
• Reduced scrap cost by $7/ton through optimized recipes
• Improved vendor scrap density consistency by 50%
