Substation Power Quality Analysis: What the Study Measures and What Networks Accept

A substation power quality analysis records harmonic voltages and currents, flicker, voltage unbalance and RMS variation at the connection point, using Class A instruments connected to the substation's CTs and VTs. Measurement typically runs for at least a week, and the statistics are compared with the emission limits the network service provider allocated to that connection.

That is the short answer. At a substation, the detail that decides whether a network accepts the result is mostly in the measurement setup: which CTs and VTs feed the analyser, whether they can be trusted across the harmonic range, and whether the instrument is set up to catch interharmonics and ripple control interference. This article focuses on that setup, for generators, industrial sites and large loads alike. If you need the full staged test sequence for a solar farm, from emission study to compliance report, see our guide to power quality assessment for solar farms.

When a Substation PQ Analysis Is Needed

Most substation power quality studies are triggered by one of four situations.

• New connection compliance. A solar farm, BESS, industrial plant or other large load has energised and the connection agreement requires measured evidence that it stays within its allocated limits. This is usually a hold point or a condition of moving to full output.
• Pre-connection baseline. Before the plant energises, the network's existing background distortion is recorded so that the plant's later contribution can be separated from what was already there.
• Complaints or unexplained equipment problems. Nuisance tripping, transformer or capacitor overheating, failed ripple control signals or customer flicker complaints often lead back to a substation.
• After plant changes. New inverter firmware, an added BESS, a filter removed for maintenance, a capacitor bank change or a new feeder can all move the result. Networks increasingly ask for re-measurement after material changes.

For generators, the measured result also has to line up with what the connection studies predicted. If you are still at the design stage, our guide to grid connection for solar farms in Australia covers where power quality sits in the overall process.

Measurement Setup

Class A Instruments

The instrument should be a Class A power quality analyser to IEC 61000-4-30. Class A defines how each parameter is measured and aggregated, how accurate it must be and how the instrument keeps time. Two Class A instruments connected to the same point should give the same answer, which is the whole reason networks ask for it.

In practice that means harmonics measured to IEC 61000-4-7, flicker measured to IEC 61000-4-15, values aggregated over standard intervals (10-minute values being the ones most limits are assessed against), and a time source such as GPS so that records from different sites, or from the network's own monitor, can be lined up.

CTs and VTs

The instrument is only as good as what feeds it. At most substations the analyser connects to existing secondary circuits, so the transducers need checking before anything is installed.

• Voltage transformers. Inductive VTs at distribution voltages generally pass the lower harmonic orders well. Capacitive voltage transformers, common at transmission voltages, can distort harmonic measurements significantly because of their internal tuning. Where only CVTs are available, a harmonic measurement correction, an RC divider or another measurement arrangement may be needed, and the report should state which.
• Current transformers. Metering-class CTs are preferred. Protection CTs are sized for fault current and can be inaccurate at normal load currents, which is where harmonic currents are small.
• Frequency response. For any transducer, the question is whether its ratio and phase error are known across the harmonic range being assessed. If they are not, the harmonic results carry an uncertainty that a reviewer is entitled to question.

The CT and VT Survey

A short site survey before any logging saves a wasted week. It should confirm:

• CT and VT types, ratios and accuracy classes, from nameplates and the substation drawings
• which secondary cores are free, and whether a metering core can be used rather than a protection core
• whether the VTs are inductive or capacitive, and what correction or alternative arrangement a CVT will need
• isolation points, test blocks and fusing, so the analyser can be connected and removed without an outage
• a GPS or other time source location, and power for the instrument for the full logging period

Correcting for Capacitive Voltage Transformers

A CVT is tuned to give an accurate ratio at 50 Hz. Away from 50 Hz its ratio and phase shift change, so harmonic voltages read through a CVT can be badly over- or under-stated depending on the order. The usual options are a frequency-response correction derived from the CVT's characteristics, a dedicated RC divider or harmonic measurement tap, or measuring on a lower voltage winding with inductive VTs and referring the results. Whichever is used, the report has to state it, because a reviewer will ask.

Interharmonics and Ripple Control

Interharmonics are components between integer harmonics. Inverter fleets, VSDs and cycloconverters can produce them, and an interharmonic close to a DNSP's ripple control frequency can disrupt hot water and load control across a feeder. Plant capacitance can also absorb the ripple signal and weaken it downstream. Where the network uses ripple control, set the analyser to record interharmonic groups or the specific ripple frequency, and log the signal level during scheduled ripple transmissions.

Industrial and Large-Load Sites

At an industrial substation or a large load such as a data centre, the emitting equipment is the load itself: VSDs, rectifiers, UPS systems, arc furnaces or welders. The setup is the same, but the logging period has to capture every production mode, including start-up, shift changes and maintenance shutdowns. Recording current on each major feeder, not just the incomer, lets the report attribute distortion to the equipment that causes it.

Baseline Versus Post-Connection

Background distortion has to be measured before new plant energises, because afterwards there is no clean way to separate the network's existing harmonic voltages from the new contribution. Without a baseline, an exceedance measured later tends to be attributed to the newest connection. Our solar farm power quality assessment guide sets out how the baseline fits into the full test sequence.

What Is Analysed

Harmonic currents matter as much as voltages. The voltage at the connection point is what the network cares about, but the current is what your plant controls, and correlating the two is how you show cause and effect.

Assessing Against Allocated Limits

Planning Levels and Emission Allocation

Each network sets planning levels guided by AS/NZS 61000.3.6 and AS/NZS 61000.3.7, then shares the headroom above background levels between customers as emission allocations written into the connection agreement. Our guide to AS/NZS 61000.3.6 harmonic emission limits explains how those allocations are derived; the limit that matters for the study is the one in your agreement, not a generic figure from the standard.

Statistics

Power quality limits are statistical, not instantaneous. The most common assessment uses the 95th percentile of 10-minute values over the measurement period: the value exceeded for only 5% of the time. Some networks also look at very short (around 3-second) values at a higher percentile, or at maximum values, for specific parameters.

Two practical points follow:

• The plant must operate across its normal range during the measurement. A week of overcast weather on a solar farm, or a BESS sitting idle, gives statistics that do not represent the plant.
• Data has to be filtered properly. Intervals flagged under IEC 61000-4-30 (during dips, interruptions or swells) are normally excluded, and the report should say how much data was excluded and why.

The Modelling Side

Measurement shows what happened in one network configuration over one period. Networks also want confidence that compliance holds under other conditions, which is where modelling comes in.

The NSP usually provides harmonic impedance loci for the connection point: regions in the R-X plane, per harmonic order or frequency band, that cover the range of network impedances expected under system normal and contingency conditions. Your consultant combines these with the plant's harmonic model and frequency scans of the plant's own collector network, cables, transformers and filters to find the worst-case harmonic voltage at each order.

Measurement and modelling meet in three places:

• Validation. Measured harmonic currents should be consistent with the injection the model assumed. If they are not, the model assumptions need revisiting.
• Resonance. A measured amplification at a particular order often matches a resonance the frequency scan predicted, typically from cable capacitance in the collector system interacting with network inductance.
• Mitigation decisions. If the measured results are close to the limit and the model shows worse cases under contingency, a filter or inverter control change may still be needed even though the measurement passed.

This work typically runs in the same harmonic and power system tools used for connection studies. Our article on PSS/E and PSCAD modelling for renewables covers that toolset.

Deliverables Networks Expect

For a generator, the report also needs the plant operating record and baseline comparison covered in our solar farm power quality assessment guide. If you are assembling the wider connection submission, see what documents a DNSP needs.

The measurement also depends on design decisions made much earlier, including transformer vector groups, cable lengths, filter provisions and metering-class CT cores in the substation. Our guide to solar farm electrical design covers where those choices are made.

Commission a Substation PQ Study

ElectraGlobe plans and delivers substation power quality analysis for solar farms, BESS, industrial sites and large loads across Australia. That covers baseline and post-connection measurement, harmonic impedance modelling, filter design and reports formatted for the NSP that will review them.

See our power quality analysis service, or contact our team to discuss your connection point.

Frequently Asked Questions

How long should power quality be measured at a substation?

One week of continuous measurement is the common minimum, because it captures a full weekly load cycle and gives enough 10-minute intervals for meaningful 95th percentile statistics. Some networks ask for longer, or for measurements across specific operating conditions such as full output. Always confirm the duration against your connection agreement or the NSP's written requirement before you start.

What is a Class A power quality meter?

It is an analyser that meets the Class A measurement requirements of IEC 61000-4-30. The standard fixes measurement methods, aggregation intervals, accuracy and time synchronisation, so two compliant instruments at the same point give comparable results. Networks generally require Class A for compliance evidence.

Who sets the harmonic limits at my connection point?

The connecting network service provider sets them, guided by AS/NZS 61000.3.6 for harmonics and AS/NZS 61000.3.7 for flicker. It allocates part of the headroom between background distortion and its planning levels to your connection. The resulting limits are written into the connection agreement and, for registered generators, the performance standards.

Can existing protection CTs be used for power quality measurement?

Sometimes, but they are not the preferred option. Protection CTs are designed to be accurate at fault currents and can be inaccurate at normal load, where harmonic currents are small. Metering-class CTs are better, and whichever CTs are used, the report should state their class and any limitations so the reviewer can judge the results.