One goal, everything else a constraint
The objective is reboiler steam, the unit’s main energy input. Treated-gas quality, amine losses and pump power must stay no worse than today, and no new fault or early warning may appear.
Optimise
Find the best settings, without tripping a fault.
Optimise searches the ranges you approve for the lowest reboiler steam flow. Every candidate is screened by the same diagnostic engine that watches the plant, so a setting that looks cheaper on paper but edges toward foaming or flooding never makes the list.
11.1%Median steam cut, within recorded range
3Set-points moved
1Goal: minimise steam
0Constraints allowed to worsen
MVPEvery figure above is counted in the minimum viable product: the amine train modelled for gas sweetening and post-combustion carbon capture.
What it does
Four things it does that a dashboard does not.
A genetic search, checked by the diagnosis
A genetic algorithm converges on the fast reference model, then puts its winner through the full diagnostic engine. If the winner fails, the search walks back toward today’s settings for the largest move that still passes.
The same answer every time
Candidates sit on a fixed grid and the search is seeded from the moment alone, so the same request always returns the same answer. Editing a range moves the result by a step, not somewhere random.
Priced in money
The saving is shown in tonnes of steam per hour, thermal megawatts, dollars per hour and dollars per year, next to the constraints it respected.
How it works
The pipeline, end to end.
- 1
Pick a moment
Any point in the plant history. Feed and other loops stay at their real measured values.
- 2
Approve the ranges
Set how far lean circulation, reboiler steam and antifoam may move.
- 3
Search on the model
The genetic algorithm scores candidates on the reference model in milliseconds each.
- 4
Verify on the diagnosis
The best candidate is run through the full fault diagnosis before it is offered.
- 5
Show the saving
Optimised against current settings, with every constraint and its margin.
Where it applies
Built for amine systems today. Not limited to them by design.
The platform is focused on acid gas removal — gas sweetening and post-combustion carbon capture — because that is where the model is calibrated and where the fault library is real. The method underneath is not amine-specific: a physics reference model, residuals against expected-healthy, fuzzy diagnosis over graded indicators, and costed consequences. Extending it to another unit means new datasheets and a new calibration, not a new platform.
The fastest way to judge this is to open it.
The platform runs on a real amine train with thirteen days of one-minute history behind it. Nothing on these pages is a mock-up.
Open the platform