In heat treatment, the furnace temperature is only part of the story. The oven can report zone temperatures, setpoints, and air measurements, but it does not directly tell how the core, surface, and colder regions of the product are evolving. For large coils, billets, castings, molds, and other massive geometries, that difference can decide whether the cycle is complete, too short, too long, or drifting away from the intended thermal path.
Why Real-Time Follow-Up Matters
Even when a thermal process has been carefully planned, real furnaces rarely behave exactly as expected. Temperature oscillations, airflow changes, varying loads, door openings, sensor deviations, production interruptions, and equipment aging can all change the actual thermal history experienced by the product.
A furnace sensor tells what the equipment is doing at a measured location. ThermoTwin is designed to answer the next question: what does that mean for the temperature distribution inside the product geometry right now?
What ThermoTwin Does
ThermoTwin connects measured process data to a physical thermal model. As new boundary measurements arrive, the model updates the estimated product state and presents the result in terms that are useful for operations, engineering, and quality decisions.
- Track the estimated internal temperature field while the cycle is running.
- See core, surface, minimum, average, and maximum temperatures from one live view.
- Understand whether the product is actually following the intended thermal path.
- Detect drift, interruptions, uneven boundary heating, and setpoint mismatch earlier.
- Support release, soak, hold, and cooling decisions with product-focused indicators.
Industrial Connection
Every plant has a different data reality. Some sites already have reliable furnace sensors, PLC tags, historian records, and network access. Others need a small, dedicated acquisition layer before a live model can be trusted. ThermoTwin can be configured for both cases.
- TCP/IP, Modbus TCP, Modbus RTU, OPC UA, PLC interfaces, historians, SCADA systems, and industrial databases.
- Reuse of available furnace sensors, zone thermocouples, air-temperature signals, or existing plant data where they are reliable.
- Optional ODE-supplied hardware for new acquisition points, edge communication, and practical signal integration.
- A communication layer adapted to the plant instead of forcing the plant to adapt to the software.
How It Works
The concept is simple: the furnace provides measured boundary conditions, the reduced thermal model updates the product state, and the operator sees the internal thermal estimate without waiting for a full offline simulation.
In the replay below, the recipe and four furnace measurements sit on the left. The coil temperature history is larger on the right, because that is the primary engineering result: how the minimum, average, and maximum product temperatures move during the cycle.
On-demand implementation
ThermoTwin is built around the process that exists in your plant. We define the monitored geometry, available furnace signals, data acquisition route, operator view, and validation steps for each installation.
Physical model with minimal sensors
The system does not require sensors inside every product location. It uses a physics-based thermal model to convert a small number of reliable boundary measurements into an estimated internal temperature field.
Reduced-order live calculation
For real-time use, the detailed physics are reduced into a fast model that can update continuously as new measurements arrive. This keeps the calculation light enough for live process follow-up while preserving the thermal behavior that matters for decisions.
Example: Aluminum Coil Annealing
In aluminum coil annealing, ThermoTwin receives real furnace temperatures during the cycle and continuously estimates the internal temperature field across the coil. The model follows the real boundary behavior, so top and bottom heating differences, side effects, unexpected door openings, different heating rates, and interruptions are reflected in the calculated product state.
The demo above replays a precomputed cycle from a static data file. The browser is not solving the thermal model live; it is replaying saved frames at 1000x speed so the page stays simple and fast.
Beyond Coils
Although aluminum coil annealing is a clear use case, the same architecture can be adapted to any product where measured furnace conditions drive an internal thermal state that operators cannot directly observe.
From Furnace Data to Product Intelligence
ThermoTwin turns ordinary furnace measurements into a product-focused thermal estimate. The operator still sees the oven behavior, but also receives a clearer view of what the product is likely experiencing internally.
This creates a practical bridge between plant data, thermal engineering, and quality control: fewer blind decisions, better cycle visibility, and a stronger foundation for data-driven heat-treatment optimization.
FAQ
Can ThermoTwin be built on demand for a specific furnace or product?
Yes. ThermoTwin is an on-demand engineering software project. The geometry, thermal model, communication layer, replay database, hardware needs, and operator interface are adapted to the furnace, product, and data infrastructure available in the plant.
Do we need many sensors inside the product?
Usually no. The goal is to use a minimal practical sensor set, such as furnace zones, top and bottom boundary temperatures, sidewall or return-air signals, and existing PLC or historian data. The physical model estimates the internal product state from those boundary conditions.
Can ODE provide the hardware connection?
Yes. When existing sensors and plant data are enough, ThermoTwin can reuse them. When the installation needs additional acquisition points, edge communication, or signal conditioning, ODE can specify and provide the right hardware around the model.
How can the model run fast enough for live follow-up?
The live system uses reduced-order physical models. The model keeps the important heat-transfer behavior, but is simplified enough to update quickly when new plant measurements arrive.
Is the browser demo calculating the twin live?
No. The site demo replays precomputed frames stored in the site data folder. A real ThermoTwin installation can run the reduced-order model live against plant data, but the marketing page stays lightweight by replaying saved data.
Apply ThermoTwin to your process
Share your geometry, furnace configuration, and available plant data. We can adapt the communication layer, hardware acquisition, reduced model, and operator view around your process.