Digital Twin — JSW Steel (Bellary)
Visual: Plant digital twin.
Client Overview
JSW Steel runs integrated steel production lines with large furnaces, continuous casting, and rolling mills. The client needed to model complex plant interactions to reduce energy consumption and improve throughput without risking production stability.
Challenge
Energy costs and material losses during cast-to-roll transitions impacted margins. The plant lacked a way to simulate “what-if” adjustments at scale or predict the energy impact of operational changes before implementing them on the floor.
Goal: Build a validated digital twin to simulate operational changes and identify energy savings and throughput gains safely.
Solution — Digital Twin Platform
We delivered a multi-layer digital twin linking process models, telemetry, and control logic to provide pre-deployment simulation and optimization recommendations.
Core modules
Approach
- Build physics-aligned process models validated against historical telemetry.
- Calibrate twin using 3 months of high-resolution sensor and energy meter data.
- Run optimization scenarios (e.g., furnace heating schedules, rolling speed adjustments).
- Deliver plant-side recommendations with expected energy and yield impacts.
Technology stack
Implementation — How We Executed
Phase 1 — Data Ingestion & Model Scoping (Weeks 1–6)
Cataloged sensors, validated timestamps, and defined model boundaries with process engineers.
Phase 2 — Model Development (Weeks 7–20)
Built and iterated process models; integrated energy meters and control logic for realistic simulations.
Phase 3 — Calibration & Validation (Weeks 21–28)
Compared simulation outputs to real production runs, adjusted parameters, and improved fidelity.
Phase 4 — Scenario Testing & Handover (Weeks 29–44)
Ran multiple operational scenarios, quantified energy tradeoffs and throughput impacts, and produced implementation playbooks for operations teams.
Impact & Results
15%
Reduction in energy consumption
8%
Increase in throughput during optimized windows
20%
Reduction in material losses during transitions
9 months
Time to validated operational improvements
Qualitative outcomes
- Operations team gained confidence to apply changes during low-risk windows.
- Improved planning and sharper visibility for long runs.
- Framework for continual model refinement as plant conditions evolve.
Client Testimonial
Key Highlights & Learnings
- Digital twins are most valuable when validated against quality historical data.
- Cross-team workshops ensure scenario outputs are operationally actionable.
- Start with high-leverage subsystems (energy or bottleneck points) for fastest ROI.