QMerse · AI workflows for power systems
Physics-aware AI agents, with an engineer in the loop
Agents run your study through PSCAD, PowerFactory and Python. An engineer signs off every result.
Input
Brief & scope
The requirement, the applicable standard and the operating envelope, captured before anything runs.
What it does
- Parses the brief into an explicit study scope
- Maps grid-code clauses to concrete checks
Recorded for audit
- Brief and scope versioned
- Applicable standards logged
Works withPSCAD/EMTDCDIgSILENT PowerFactoryPSS®EPythonExcel & CSVGrid codes & PDFs
Co-pilot for PSCAD
Ask for a study in plain English and QMerse runs it in PSCAD.
QMerse co-pilot
YouRun FRT on all 4 operating points at 0.15 pu for 140 ms and check it against the grid code.
- Open plant_v7.pscx · 4 operating points
PSCAD - Apply fault: 0.15 pu retained, 140 ms
- Run FRT cases36
- Check Vpcc and Iq against the FRT envelope
- Case 23 near the limit → engineer
EngineerApproveRe-run case 23Approved · logged to the audit trail
Ask QMerse to run, check or explain a case…↵
SchematicGraphsOutput
EMTDC · readyEMTDC · solvingEMTDC · 36 / 36 completeIllustrative
Requests become PSCAD runsEach action is loggedResults checked against the grid codeAn engineer approves the result
Use cases
Some of the workflows it can run.
Developers · OEMs
Grid-code compliance evidence
- Grid code and plant model
- Clauses mapped to checks
- FRT, FFCI and SSO cases
PSCAD - Checked against criteria
- Engineer review
- Submission-ready evidence pack
TSOs · OEMs
EMT / RMS model validation
- Vendor EMT and RMS models
- Benchmark cases in both
PSCAD · PF - Behaviour reconciled
- Assumptions documented
- Engineer review
- Validation report
DNOs · Developers
Connection study packs
- Network model and scenarios
- Load flow and fault level
PowerFactory - Voltage and protection screening
- Checked against limits
- Engineer review
- Connection study pack
All teams
Data orchestration across tools
- PowerFactory, PSCAD and PSS®E data
- Assembled and validated
- Every value versioned
- Engineer review
- One checked dataset
TSOs · Planners
Scenario screening
- Operating and outage scenarios
- Every case swept
Python - Ranked by margin
- Engineer review
- Shortlist of governing cases
Examples
Governance
What gets recorded, and who signs off.
Run 0427 · audit trailSealed
- Brief v3 versioned · grid-code clauses mapped
- Network model validated · 4 operating points
- PSCAD 36 cases · PowerFactory 124 · tool versions logged
- 154 / 160 within limits · 6 flagged
- Signed off by the engineer · 2 re-runs requested
- Evidence pack sealed · a3f9·61c2
An engineer decides
QMerse prepares and checks the work. The engineer approves it.
Access by role
You set who can run, review and release a study.
Where it runs
Hosted by us, in your cloud or on your own servers.
Your data
Used only for the workflows you agree. You can use your own model keys.
The platform
Build your own workflows on qmerse.com.
Your team can build workflows on the same canvas. It is free to start.
Open qmerse.comDemonstrator · Multi-agent
Smart Agent for Grid Reinforcement
Agents decide what to study, run PowerFactory and Python, interpret the results and draft the report.
Demonstrator · Distributed stability
Stability Potential
Qualifies stability services from networks below 132 kV, with a procurement-ready scorecard for the transmission operator.
FAQ
What agentic AI means here.
What does agentic AI mean in power systems?
Software agents that plan and carry out a multi-step study: set it up, run the tools, check the results against criteria and flag exceptions. The engineer defines the intent and reviews every consequential decision.
How is this different from asking an LLM?
A chat model writes text. QMerse agents operate the established tools (PSCAD, PowerFactory, PSS®E and Python), respect model and grid-code requirements, and check outputs against physical and regulatory criteria. The aim is to automate the repetitive, auditable parts of the work, not to replace engineering judgement.
How do the workflows stay auditable?
Each step is recorded: inputs, tool versions, assumptions, scenarios and results, so a reviewer can reproduce and challenge the work. Because the agents drive accepted tools, the evidence is the kind a network operator or regulator already expects.
Where does QMerse run?
Velon-hosted, in your cloud or on-premises, with role-based access. Your data is used only for the workflows you agree.
Next step
See it running on one of your studies.
01A demo→02A pilot on one study→03We run it with you→04Licensed to your team