Measure what your analysers can't.
Machine-learning virtual sensors estimate quality, composition, emissions and efficiency in real time — where a physical analyser is slow, costly or unavailable — trained on your historian and validated against your lab.
Use the instruments you already have to estimate the variables you can't measure continuously — and prove it against the lab before anything goes live.
The lab samples every four hours. The soft sensor never stops.
Orange dots are lab results — slow and sparse. The teal line is the soft sensor, inferring the same quality every second from temperatures, pressures and flows. Take the online analyser offline and see what's left.
- Inferred from instruments you already have
- Band widens when inputs drift out of the training range
- Lab results recalibrate the model
A whole solution set on one capability.
The Autopilot modules are built from a common analytics and control capability. The same capability is offered directly, scoped against a plant's own historian data.
Machine-learning virtual sensors estimating quality, composition, emissions and efficiency in real time.
Multivariable controllers that predict, optimise and respect process constraints on every cycle.
Degradation detected from vibration, temperature and process data — before the downtime, not after the failure.
Continuous monitoring with root-cause guidance written for the operator on shift.
Real-time replicas for what-if simulation, predictive insight and operator training.
Plant-wide optimisation computing the optimal targets that the advanced controllers then hold.
What changes when the estimate is continuous.
Stop waiting hours for the lab. A validated soft sensor publishes lab-grade estimates every few seconds, from instruments you already have.
| Lab sample | Soft sensor | |
|---|---|---|
| Latency | Hours — sample, transport, analyse | Seconds — published continuously |
| Coverage | A few samples per shift | Every cycle, every unit |
| Cost per value | Technician time, reagents, lab load | Existing instruments, no consumables |
| Use in control | Too late to act on | Direct input to MPC and RTO |
| Validation | Is the reference | Benchmarked against the lab before go-live |
Historian, DCS and lab data — not a vendor's generic model.
Inferred values checked against lab results on a holdout before an operator ever sees them.
Each estimate carries a band, so the operator knows when to trust it.
The soft sensor appears in the DCS like any other instrument.
Versus a lab sample every few hours.
On every estimate.
From soft sensing to RTO.
One unit, your historian, an agreed accuracy gate.
How accurate is a soft sensor compared to the lab?
Accuracy is benchmarked against your lab reference before go-live and reported in your units, with a confidence band published beside every estimate.
What data do we need to start?
Most engagements start with 6–24 months of historian, lab or document data. We audit coverage in the first week and tell you plainly if it is not enough — before you commit.
How long until we see results?
A proof of concept on your historical data typically lands in 4–8 weeks. Production deployment follows once the KPI target is met and your team signs off.
Where does it run, and who sees our data?
On your infrastructure — plant-side servers, your Azure or AWS tenancy, or an air-gapped network. Data never leaves your boundary, and nothing is used to train third-party models.
Who owns the result?
You do. Handover includes source code, documentation, training sessions and a retraining plan.