Process operator competency simulators
Your operators,
trained on
your own unit.
Procedures tell an operator what should happen. They cannot give the experience of working through an upset: a flooding column, a lost pumparound, alarms coming faster than they can be cleared. That is where the wrong move costs real production. A simulator of your own unit lets your people build that experience safely, before it matters.
We build a working model of your actual unit, your configuration, your tag names, your DCS graphics, in eight weeks. Then your operators can run it, break it, and recover it as many times as they need.
Inside the simulator
A full operator console.
A live, working model of your plant, the exact board your operators watch every shift, so the console is familiar before the training even starts. Not a recording, not a generic lookalike.
Instructor & grading
Drop in a failure. Score the response.
An instructor introduces a malfunction into a live session without warning: a pump trip, a fouling exchanger, an instrument failure. The simulator records exactly what each operator did, when, and in what order, and scores how they handled it. The upset and the grade are one loop: because you set the test, you prove competency instead of just logging attendance.
| Operator | Alarms raised | Avg response | Score |
|---|---|---|---|
| OP-01 | 3 | 3.2 s | 92 |
| OP-02 | 6 | 6.8 s | 78 |
| OP-03 | 4 | 4.1 s | 85 |
| OP-04 | 9 | 11.4 s | 61 |
Trends
Scan back and ask why
A trend historian your operators can scroll back through to see exactly how a situation developed, tag by tag, in the order it really happened.
And every screen you just saw runs on one live model, not disconnected screens. Cut a pump and the column feels it minutes later; lose a pumparound and the whole balance shifts, and you watch the consequences move through the plant in real time, exactly as you would on the real board.
Crew training
One plant. The whole crew.
The units aren't separate simulators sitting side by side. They're one authoritative plant, so a crew trains the way they actually run a shift: each operator at their own console, coordinating across units the way a real crew has to.
Coupled, not mirrored
Trip a steam header and it doesn't land on one board. The column loses reboil, the compressor loses its driver, and the treater's regenerator sags, all at once. Because it's one authoritative plant, an upset ripples the way it really would, and the crew has to catch it together. That's the thing a solo simulator can't teach.
One instructor, the whole board
The same instructor station that arms malfunctions and scores the response now spans a whole crew rather than a single seat. One person watches every console at once and grades how the operators respond together: who caught it, who compensated, and whether the handoff between units held.
An upset doesn't wait for one operator to catch up.
Neither should the training.
The interactive demo runs a single operator. Crew mode puts a full crew on one coupled plant, and it ships with your deployed system.
Schedule
Eight weeks, not a procurement cycle.
A custom simulator is scoped like a capital project: RFPs, committees, approvals. We deliver a working model of your first unit in eight weeks, at a fixed scope.
Specification
What actually gets delivered
- 01Your unit, not a template
- Built to your P&ID, your equipment sizing, your tag names and your DCS graphic conventions. An operator should recognise the screen before they recognise the software.
- 02Upset and trip scenarios
- Built to whatever your training group needs, standard upsets like pumparound loss, compressor surge and instrument failure, plus the specific incidents your own unit has lived through. Each one runs end to end, in real time, with real consequences.
- 03Built by a working panel operator
- Not by a modelling team that has read about a console. These are built by someone who has spent over a decade sitting at one, and every unit is checked against a second operator's judgement before it reaches you.
- 04Runs where you need it
- Browser-based for classroom use, or deployed on-premise on an air-gapped training network. No connection to the live control system, by design.
- 05Fixed scope, built to grow
- Start with one unit under a fixed scope, and expand across the plant from there. When you revamp, the model gets revised rather than rebuilt, because it was built to be revised.
- 06Real physics, where it counts
- Distillation columns run on EmpiraCore, our physics engine: a first-principles vapour–liquid-equilibrium model gives real tray temperatures and product qualities, not dialled-in numbers, while the rest of the plant is dynamically modelled and tuned to your unit. Realistic for training, without pretending to be an engineering property tool.
The engine
What is EmpiraCore?
EmpiraCore is our physics engine, a deliberate blend: first-principles physics where it matters, and dynamic models tuned to your unit everywhere else.
Distillation columns
Real equilibrium
For every column, EmpiraCore solves a first-principles vapour–liquid-equilibrium model, what each tray is actually boiling at, from its composition and pressure. Column temperatures and product qualities come out of real phase equilibrium, not dialled-in numbers. It's an idealised equilibrium model, not a full equation-of-state package: realistic enough that things move for genuine physical reasons, not a tool for predicting exact stream data.
The rest of the plant
Tuned dynamics
For everything else, from reactor behaviour to the way units push and pull on each other, we use dynamic models with real time constants tuned so the plant responds the way your unit actually does. A rigorous property calculation wouldn't change what the operator sees at the board here, so we capture what matters for training instead: the rates, the lags and the direction of every response, checked against a working operator's judgement.
For training operators rather than sizing equipment, that's the right trade.
Scope
Not just refineries
A crude column, an amine contactor and a recovery boiler are the same class of problem, and the same engine builds every one of them.
- Refining
- Gas processing
- Renewable diesel
- Petrochemicals
- LNG
- Ammonia & hydrogen
- Upgraders
- Pulp & paper
- Cement
- Desalination
Refineries are the hardest environment to model correctly, which is why we start there. Everything else is a simpler problem, and we build it the same way, just as fast. Whichever process is yours, we can build it.
Send us one unit.
Tell us which unit gives your training group the most trouble. We will come back with what a simulator of it would involve, and roughly what eight weeks looks like against your schedule. No deck, no discovery call series.