Consultancy · study

Control interaction studies

As converters cluster on the grid, their fast controls can interact with each other and with the network in ways no single project sees. We look for those poorly damped or unstable modes and find what drives them.

Discuss a project Read the case study

Interaction mode · complex planeIllustrative

Converters online nearby

Dampinginteraction mode–%at least 5%

Frequencyof the mode– Hzfor reference

Gain marginvoltage loop– dBat least 6 dB

When it's needed

When you need it.

Several converter plants or HVDC links connect close together electrically.
The connection sits on a weak part of the grid, where control margins are thin.
Equipment from different vendors has to work together, and nobody has checked the mix.
Unexplained oscillations appeared during commissioning or operation.

Results

The power swing, before and after.

Choose a case to see the result.

Power swing · dampingIllustrative

Damping −0.4 %

Damping 7.9 %

Before retuning, the swing never dies away after a small disturbance: it slowly grows.After retuning, the swing dies away within two seconds.

Method

How we run it.

01
Map
Map the converters, their controls and the network around them.
02
Screen
Use impedance methods to find where damping is marginal.
03
Reproduce
Recreate suspect modes in EMT to confirm cause and severity.
04
Stress
Test grid strength, units in service, settings and dispatch.
05
Fix
Recommend setting changes or added damping, and verify them.

Tools and scenarios

Every scenario through every study.

PSCADTime domain
Impedance scan toolsFrequency domain
Vendor control modelsEach converter

EigenvaluesImpedance scanEMT time domainSet-points

One plant online

Several plants online

Weak grid

Outages (N-1)

HVDC or STATCOM nearby

Different vendors

Several plants online · EigenvaluesWith six converters online the interaction mode dropped below 5% damping. Lower voltage-loop gains fix it.HVDC or STATCOM nearby · Set-pointsThe STATCOM and the wind farm chased each other’s voltage. A shared droop settles it.Different vendors · Impedance scanOne vendor model hid a resonance at 400 Hz. The updated model shows it, and a filter change removes it.

22 checks · 3 led to a change · every case scripted and rerunnable

Deliverables

What you get.

ReportInteraction assessment with the modes found and what drives them
EvidenceImpedance and damping results with confirming EMT waveforms
FixesRecommended changes and the results that verify them
MethodAssumptions and model sources, written down

How we check it

Findings are checked with more than one method, impedance screening and time-domain runs, and the conditions that decide the result are stated so the operator can review them.

Selected work

Related case studies.

Case 03 · SSO and compliance40 MW battery: SSO screening and NESO complianceImpedance scan, Nyquist, voltage injection, FRT and FFCI from one model.Case 02 · EMT modellingGenerator EMT model checked against site measurementsA full PSCAD model with excitation, stabiliser and protection, validated on site.

FAQ

Questions engineers ask.

How is this different from a single-plant compliance study?

Compliance studies assess one plant against the code. Control-interaction studies look at how multiple controls and the network behave together, which can reveal problems that each project passes individually but that emerge in combination.

Do you need every vendor's model?

To be conclusive, the relevant converter control models for the region are needed. Where some are unavailable, we are explicit about what can and cannot be concluded.

What mitigations are typical?

Often control-setting adjustments, retuning of inner loops, or supplementary damping. The right answer depends on the mode; we verify any recommendation in the model before proposing it.

Next step

Have a study like this coming up?

Discuss a project [email protected]