SystemsML

Run every plant at its optimum.

Gas turbine · power output
optimum
Target power300 MW Uncertainty5% of range ResultOn optimum

SystemsML is an AI layer above your DCS and APC that computes safe, uncertainty‑aware set‑points for energy, chemical and Waste‑to‑X plants.

Book a demo Try the live tool

Plants run inside
a safety cushion

AI could add about $13 trillion to the world economy by 2030. Almost none of it reaches safety‑critical plants.

Static set‑points. Targets are chosen conservatively and rarely revisited, so plants give up throughput, efficiency and CO₂ every shift.

The trust gap. Black‑box neural models carry no uncertainty and no physics guarantees, so they never clear functional‑safety review.

No measure of confidence. First‑principles APC works at design conditions, but cannot say how sure it is when feedstock, ambient or load drift.

One layer on top

Plant data

First, live measurements arrive from the plant you already run.

SystemsML

Then a hybrid mechanistic–ML model weighs its own uncertainty (the band) and computes the optimal, safe set‑point.

Operator sign‑off

Next, an operator approves it, as regulation requires.

DCS · APC

Finally your existing loops hold it. Nothing is replaced.

Deployable from lab, to pilot, to full industrial scale. Advanced process control typically pays back in 12–24 months.

Proven, not piloted

Peer‑reviewed on a 1,320 MW thermal power plant, a 660 MW coal power plant and a 395 MW gas turbine system.

Measured,
not promised

Three peer‑reviewed studies on operating plants, read against industry benchmarks for advanced process control.

13 $ million saved per year on a 1,320 MW thermal power plant.
2,900 t of COâ‚‚ cut per year on a 395 MW gas turbine.
660 MW coal plant optimised plant‑wide.
10–15 % production uplift, plus a 4–5% EBITA gain.
12–24 months to full payback on advanced process control.
Read the papers Energy, 2026 Ashraf et al. · 1,320 MW thermal plant Computers & Chemical Engineering, 2025 Ashraf et al. · 395 MW gas turbine arXiv, 2026 Ansar et al. · 660 MW coal plant

What it does

SystemsML optimises throughput with bi‑level models bound by KKT conditions, supervises the set‑points your DCS and APC already follow, and prescribes maintenance for rotating equipment before stress becomes failure.

Run the gas‑turbine optimiser yourself: set a target power and an uncertainty level. Open the live tool

The team

The researcher who wrote the papers, and the engineer who puts them on a live plant.

Bring one asset and thirty minutes.

Plant managers and process engineers: we’ll scope the margin your set‑points leave on the table. Try the live tool first

Prefer email? contact@systemsml.tech
London, United Kingdom · energy, chemical and Waste‑to‑X plants

Book a demo

Step 1 of 2