Intelligence on top of the automation you already paid for
FermentDT is an AI supervisory control system — a digital twin — that layers neural-network soft sensors, predictive modelling, and model-predictive control on top of the PLC/SCADA already running an industrial fermentation. It doesn't replace your control system; it sits on top of it, tells you what's about to go wrong before it does, then steers the process toward a target curve instead of a static setpoint.
Closed-loop control, not monitoring — and fermentation-native, not a horizontal analytics toolkit you build yourself.
How it works
Four moves, around the clock
FermentDT runs a continuous supervisory loop over every tank. It reads, predicts, decides and steers — then reads again, a few seconds later, for as long as the batch is alive.
The core difference
A target curve, not a static setpoint
Holding a fixed number is why the same vintage drifts every vintage. FermentDT steers the whole trajectory.
One number to hold. The batch wanders around it and nobody knows the end-point is wrong until it's too late to act.
● A small early correction keeps the batch tracking the ideal path — and lands it on the same end-point, every run.
No new probes
Measure what you can't put a sensor on
Neural soft sensors turn the cheap signals you already log into the lab values you'd otherwise sample by hand — continuously, with nothing new in the tank.
+ digital twin
Where it sits
On top of the stack — not in place of it
An optimization layer over the control system you already run. No rip-and-replace, no plant downtime, vendor-neutral by design.
01—06 · What's in the box
The supervisory stack
Six capabilities, one engine — the full loop, packaged.
Built for your tanks
One closed-loop engine, spoken in your language
Once-a-year ferments where a stuck or spoiled tank is a vintage you can't redo — and nobody can sample every tank every hour.
Year-round, high-throughput fermentation where batch-to-batch consistency and tank turns are the whole economics.
Emerging precision-fermentation and food processes where repeatable yield and process control gate scale-up.
Larger bioprocess operations running on PLC/SCADA that want closed-loop intelligence without an enterprise rip-and-replace.
On the floor
What it changes
catch contamination and temperature excursions before they spread across the cellar
predictive control drives every batch toward the same trajectory
fewer 2 a.m. cellar checks, fewer overnight surprises
software on top of automation you already paid for
See it run
The live demo lab
Step inside the FermentDT live demo — a simulated digital-twin dashboard that shows the system the way an operator would see it. Watch live fermentation telemetry stream from a virtual tank while the soft sensors infer what the hardware can't, an anomaly flag lights up as a temperature excursion pulls the batch off-trajectory, and the model-predictive controller proposes the correction that brings it back to the target curve.
Enter the live demo →