01The monthly surprise
Energy, grade and recovery estimated every heat.
- A physics model of your furnace — mass, energy and slag — calibrated on plant data
- Drift flagged with its cause: basicity, reductant, electrodes
Act in the shift, not the month
EnterpriseFurnace Twin · physics-based digital twin
Ask the furnace. See the proof.
For submerged-arc and other smelting furnaces, where energy is half the operating cost and the recipe lives in senior operators’ heads. A physics model of your furnace — mass and energy balance, slag chemistry, kinetics — calibrated on every heat, with cited answers your whole shift can ask for.
- Physics + data hybrid model
- Every heat calibrates it
- Cited answers, never a guess
WhySound familiar?
Three things we hear in every first conversation.
From plant heads, process engineers and furnace operators. Composite quotes from discovery calls — the problems are real; the names are left out.
-
“We learn the furnace used more power when the bill arrives.”
Energy is half the operating cost, but specific energy is only seen monthly — after the money is spent.
Plant headWeeks between cause and number
-
“The recipe lives in two senior operators’ heads.”
Reductant ratio and slag basicity are set by habit; when those operators are away, every heat is a gamble.
Process engineerKnowledge that walks out of the gate
-
“We can’t try a new blend without risking a heat.”
Every change is tested on the real furnace, so most good ideas are never tried.
Furnace in-chargeSavings left on the table
HowWhat changes with Furnace Twin
The same three problems, on the screens that solve them.
Product screens with illustrative data. In a briefing we model one of your furnaces.
02Knowledge in heads
Anyone on shift can ask the furnace.
- Plain-language questions answered from records, the model and literature
- Every answer cites its sources
Experience that stays in the plant
03Risky trials
Try the change on the twin first.
- Reductant, basicity, temperature and feed compared side by side
- Only the best scenario goes to a two-heat trial
Fewer gambles, faster savings
FieldFurnace Twin in the field
The people, the phone, the place.
The workplace it runs in, the person who uses it on a phone, the team at the desk and the detail it reads.
Photographs are illustrative — real customer sites are not shown.
- 40–60%Of operating costis energy in a submerged-arc furnace
- 3–8%Energy at staketypical digital-twin saving range
- 6Physics modelsmass · energy · slag · kinetics · electrical · zones
- ± rangeOn every estimatecalibrated against plant history
- CitedAnswersrecords · model · literature
- VoiceShift intakeHindi or English
01The screens
From the furnace floor to a decision you can defend.
The live overview, the scenario engine, cited answers and voice shift intake.
autowhat · Furnace Twin / Furnace overview
Illustrative data
Live estimates of energy, grade and recovery, the furnace zones, and what the twin is watching.
Product screens with illustrative data — furnace, ore and heat figures are placeholders.
02Before and after
Most furnaces run on habit and a monthly power bill.
Less than one in twenty Indian furnaces has any digital optimisation. The twin turns every heat into evidence.
| Area | Today | With Furnace Twin |
|---|---|---|
| Energy | SEC tracked monthly, after the bill | Specific energy estimated every heat, with the cause of drift |
| Recipe | Reductant and slag set by habit | Scenarios priced on the model before a heat is risked |
| Knowledge | In senior operators’ heads and old reports | Cited answers anyone on shift can ask for |
| Shift handover | A notebook and a phone call | Guided voice intake that becomes structured data |
| Models | A one-off study that goes stale | Calibrated on every heat, versioned and audited |
03Physics first
Eight steps, from plant data to a calibrated twin.
Metallurgy builds the model; plant history corrects it. Generic machine learning alone cannot tell you why.
-
01
Plant data in
Heats · lab · power
Heat logs, lab assays, electrical data and shift notes collected per heat.
-
02
Mass balance
Physics
Every tonne of ore, reductant and flux traced to metal, slag and gas.
-
03
Energy balance
Physics
Electrical input against reactions, heating and losses — specific energy explained.
-
04
Slag chemistry
Thermodynamics
Basicity, viscosity and how much iron or chrome is co-reduced.
-
05
Calibrate
Hybrid model
Physics corrected with plant history so estimates match the furnace you actually run.
-
06
Run scenarios
What-if
Reductant, basicity, temperature and feed changes compared before a trial.
-
07
Answer questions
Cited
Engineers ask in plain words; answers cite records, model runs and literature.
-
08
Learn from every heat
Model lab
New heats re-calibrate the model; every version is evaluated and logged.
04The platform
Twin, knowledge, models and a physics engine underneath.
Start with one furnace; more furnaces and furnace types run on the same platform.
Twin · 4
- Furnace overview
- Smelting model studio
- Scenario engine
- Transient simulation
Knowledge · 4
- Ask the furnace · cited
- Literature & PDF library
- Shift intake by voice
- Heat & lab forms
Models · 4
- Model lab · calibration
- Evaluation reports
- Model registry
- Governance & audit
Physics engine · 6
- Mass balance
- Energy balance
- Slag chemistry
- Reduction kinetics
- Electrical model
- Furnace zones
05How we deliver
Baseline, calibrate, then operate.
Built with metallurgists and software engineers together, around your furnace and your ore.
- Months 1–2 · BaselinePlant data connected, physics model set up for your furnace and ore, a baseline report.
- Months 2–4 · CalibrateThe hybrid model tuned on your heat history until estimates match the plant.
- Months 4–6 · OperateScenarios, cited answers and shift intake in daily use; first recipe trials.
- Then · ScaleMore furnaces and furnace types on the same platform, with every model versioned.
FAQQuestions we get about Furnace Twin
Straight answers, before the call.
Furnace Twin is a physics-based digital twin of the furnace — mass and energy balance, slag chemistry and kinetics — calibrated on every heat, with scenarios to try a change before the real furnace does, and cited answers for anyone on shift. Short, factual answers for buyers, procurement and the assistants people ask first.
In short. Furnace Twin is a physics-based digital twin of the furnace — mass and energy balance, slag chemistry and kinetics — calibrated on every heat, with scenarios to try a change before the real furnace does, and cited answers for anyone on shift. It is built for plant heads, process engineers and furnace operators. It exists to remove the problems those teams describe in almost every first conversation: specific energy seen monthly, after the money is spent; and reductant ratio and basicity set by habit, lost when the operator is away. With Furnace Twin, teams get energy, grade and recovery estimated every heat, and plain-language questions answered from records, the model and literature. It connects to Plant historian / PLC data, Thermodynamic model, Local LLM with citations and Scenario engine, is used by Process · operations · plant head, and runs on web. Like every Autowhat AI product, Furnace Twin goes live on one site, lane or process first, usually within weeks, on managed cloud, in your own VPC or on-premises, and shares one operational record with the other products, so the second deployment reuses the data and integrations of the first.
Who is Furnace Twin for?
Plant heads, process engineers and furnace operators. A physics-based digital twin of the furnace — mass and energy balance, slag chemistry and kinetics — calibrated on every heat, with scenarios to try a change before the real furnace does, and cited answers for anyone on shift.
What problems does Furnace Twin remove?
specific energy seen monthly, after the money is spent; reductant ratio and basicity set by habit, lost when the operator is away; good ideas never tried because every trial risks the furnace.
What do we get with Furnace Twin?
energy, grade and recovery estimated every heat; plain-language questions answered from records, the model and literature; reductant, basicity, temperature and feed compared side by side.
What does Furnace Twin connect to, and who uses it?
It connects to Plant historian / PLC data, Thermodynamic model, Local LLM with citations, Scenario engine. Users: Process · operations · plant head. Channel: Web.
How do we start with Furnace Twin?
With a 30-minute working session on your own data — an engineer from the relevant sector, not a sales queue. The first wave is scoped in that call; most products go live on one site, lane or process within weeks and then scale.
Where does Furnace Twin run, and who owns the data?
On managed cloud, in your own VPC (AWS, Azure or GCP) or on-premises, with your keys. SOC 2 Type II and ISO 27001 controls, India data residency when required, and local or open-weight models when policy demands it. The data stays yours.
06Briefing
Bring three months of heat data. We’ll show you where the energy goes.
A working session: your furnace modelled, the mass and energy balance closed, and the first scenarios priced.