Now live: a hybrid process digital twin for amine gas sweetening — it names the fault and the root cause, not just the alarm. Explore the platform →

QEM & OPS logo

Reference Model

Run

Live in the platform

Put a working model of the plant behind every reading.

Run is the grey-box process digital twin. You give it the boundary conditions — feed, solvent, steam, cooling — and it returns the value every monitored point should have if the train were healthy. The gap between that and reality is where every diagnosis on this platform starts.

See it in the platform How it works
13Modelled process blocks
26Equipment items covered
~100Live fields on the P&ID
0Meters required for soft sensors

What it does

Four things it does that a dashboard does not.

01

Physics first, data second

Kremser absorption factors for the absorber. Counter-current ε-NTU and LMTD for the exchangers. Performance curves for the pumps. Energy balances across the reboiler and the column profiles. The fitted parameters are health parameters — UA, cleanliness, stage efficiency — so when one drifts it means something physical, not something statistical.

02

An expected-healthy value for every tag

The model returns what each monitored response should read under the current boundary conditions. A fault then shows up as a residual — actual minus expected — instead of as a fixed alarm limit that has to be re-tuned every time the plant changes load.

03

Soft sensors where no meter exists

Rich-amine flow to the lean/rich exchanger has no meter anywhere on the train. Run computes it from the lean circulation, the acid gas absorbed between the feed and treated analysers, and the filter side-stream. Exchanger fouling is published the same way, as a live UA and cleanliness factor.

04

The P&ID is the interface

Every box that carries a live value is marked in the drawing itself. The engineer works on the plant diagram, not a spreadsheet, and the picture cannot drift out of step with the calculation.

How it works

The pipeline, end to end.

  1. 1

    Calibrate once, offline

    Each block is fitted against the training history and the equipment datasheets. The result is baked into a calibration file. This is the slow step, and it runs rarely.

  2. 2

    Take the boundary conditions

    Feed rate and pressure, inlet analysers, solvent strength from the laboratory, steam and cooling-water conditions. Everything the operator can actually set or observe.

  3. 3

    Solve block by block

    Absorber, stripper and reboiler, lean/rich exchanger, inlet conditioning, overhead condenser, filters, pumps, drums, system inventory. The absorber and regenerator loops close against each other.

  4. 4

    Return the healthy plant

    Every response tag, plus the soft sensors and the health parameters. No refitting happens at query time — the baked calibration is reused, so a single point runs in seconds on unseen data.

  5. 5

    Hand the residuals on

    The difference between expected and actual becomes the input to Detect and Prognose. One model, three tools.

Under the hood

The engineering, stated plainly.

Every figure on this page is countable in the delivered system. If you want to see where a number comes from, ask us and we will show you the record it is read from.

Method
Hybrid grey-box — physics core, data-fitted health parameters
Solvent system
Activated MDEA — 37.9 wt% MDEA / 7.6 wt% HEP
Absorber
Kremser absorption-factor model with fitted stage efficiency
Exchangers
Counter-current ε-NTU / LMTD, anchored to datasheet duty
Rotating equipment
Performance-curve health for P-401 to P-404
Health parameters
UA, cleanliness, fouling resistance, stage efficiency
Calibration
Baked to file; online queries never refit

What it does not do

The limits, published on the same page as the promise.

  • It is calibrated to one train. A new unit needs its own datasheets and a calibration pass — this is engineering work, not a configuration screen.
  • It predicts the healthy state. It does not optimise setpoints and it does not write to the control system.
  • Blocks were built one at a time and coverage is documented per block. We publish which blocks are complete rather than claiming the whole train is modelled.

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.

  • Gas sweeteningLive today
  • Post-combustion carbon captureLive today
  • Other acid gas removalMethod applies
  • Wider process unitsFuture

The rest of the platform

One model. Seven ways to use it.

DetectNames the fault, the root cause, and what it costs per hour.PrognoseProjects the slow-wear faults against the threshold that declares them.DiscoverThe whole train in KPIs, trends, profiles and correlations.qemIntelAsk the plant a question, and get an answer grounded in real data.OTSIn developmentTrain crews on faults they have never seen. In development.TuneIn developmentFind the valves and loops working against you. In development.

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