Harmonic Analysis in Power Systems: Beyond Equipment Tests

A supplier’s harmonic-distortion figure is useful evidence about a piece of equipment. It cannot, by itself, establish harmonic performance across an electrical facility.

Harmonic analysis in power systems evaluates how distortion-producing equipment interacts with the network that connects it. The answer depends on the source characteristics, electrical impedances, background distortion, and operating conditions. Individually acceptable components can still form a system that needs mitigation or a different operating arrangement.

For an owner reviewing UPS equipment, variable-frequency drives, or battery inverters, the practical decision is whether the assembled installation will satisfy its agreed acceptance criteria. That decision should be made while the electrical design can still change.

Why this question belongs in current project planning

MISO’s September 8, 2026 announcement placed 15 proposed projects, totaling approximately 7.3 GW, into its fifth ERAS cycle. The mix includes solar, wind, and battery storage. These projects are entering evaluation rather than representing operating additions. MISO ERAS announcement.

For project designers, that continuing mix of converter-based equipment makes network interaction a relevant design question. A harmonic study is one way to examine a specific part of that interaction. Its scope should follow the project’s equipment and connection conditions, with clear separation from dynamic stability and ride-through assessments.

What an equipment submittal actually establishes

Start by reading the conditions behind the published number. Was distortion measured at rated load? What supply impedance and voltage waveform were used? Which operating mode was tested? Does the information describe current distortion at the equipment input, voltage distortion at its output, or something else?

A single total harmonic distortion value compresses a spectrum into one number. Two devices can have similar totals while injecting different amounts at individual harmonic orders. That difference becomes relevant if the network has a pronounced response near one of those frequencies.

Equipment operation can change the spectrum as well. Partial loading, control settings, switching arrangements, and bypass operation may require separate data. For UPS systems, the electrical path changes with the operating mode, so define which internal and downstream loads the submitted data actually represent.

Ask for applicable test results or manufacturer-supported harmonic spectra. Record their operating conditions and limitations. An unqualified brochure value leaves too much interpretation to the person building the system model.

Set the acceptance boundary before running the model

Agree on the point of common coupling, or PCC, with the utility and identify any separate internal acceptance locations. IEEE 519-2022 addresses steady-state voltage and current waveform distortion and defines the PCC as the interface between sources and loads. Its scope should not be treated as a universal equipment-terminal test. IEEE 519-2022 standard description.

The study basis should identify the applicable standard edition, utility criteria, contractual obligations, measurement locations, and evaluation methods. A project that also exports power through inverters may have additional interconnection requirements. Establish their applicability instead of carrying a load-only checklist into every installation.

Keep voltage and current metrics distinct. Current THD uses the fundamental current as its reference. Total demand distortion uses a specified demand-current reference. Confusing those denominators can lead to an incorrect comparison, especially at light load.

Meeting a utility criterion at the PCC also leaves internal equipment questions to answer. Check the buses and branches where distortion or additional harmonic loading could affect the design.

Harmonic analysis in power systems needs two views

The first view describes the sources of harmonic current or voltage. The second describes how the network responds at the relevant frequencies.

PNNL’s January 2026 data-center modeling report separates harmonic-equivalent development from calculation of resulting grid voltage distortion. It also explains why harmonic load-flow tools can be more practical than detailed EMT simulation for this purpose. Model detail is useful only when the necessary input data can support it. PNNL report, Section 3.1.3.

Build the network representation around the actual paths. Transformers, cables, capacitors, filters, rotating equipment, and the upstream system influence the response. Their frequency-dependent characteristics may matter. The utility’s fundamental-frequency fault level provides useful context, but it cannot fully describe impedance across the harmonic spectrum.

A frequency scan identifies how bus impedance changes with frequency and where resonance may occur. Harmonic load flow then evaluates distortion for the modeled sources and operating cases. ETAP’s frequency-scan documentation describes this impedance-based approach and its use in evaluating filters. ETAP harmonic frequency scan.

Where source phase information is available and reliable, use it appropriately. Avoid assuming perfect cancellation among converters. Equally, treating every source as permanently aligned can produce an overly conservative estimate. Document the aggregation assumptions and test their effect on the decision.

A small example shows why the network matters

Consider a deliberately simplified, hypothetical single-frequency example. Ignore background distortion and hold the injected harmonic current constant at 40 A. Represent the network seen by that source with a single equivalent impedance at that frequency.

The values follow the magnitude relationship between current, impedance, and voltage for this simplified equivalent. They are not measurements, compliance results, or typical design values.

The equipment emission did not change. The voltage response increased fourfold because the impedance at that frequency increased. In a real system, switching a capacitor bank, changing a supply path, or removing a transformer can alter the frequency response. The direction and size of that change require analysis.

A detailed study also includes other harmonic sources and background distortion. Their magnitudes and phase relationships affect the resulting waveform. The example isolates one relationship so the reason for studying the network is clear.

Test the configurations the facility will use

Begin with the normal arrangement, then select additional cases based on the operating plan. Useful cases can include partial loading, different capacitor-bank steps, a transformer unavailable, an alternate utility supply, UPS bypass, or generator operation where relevant.

Tie every case to an actual permitted condition. An exhaustive list of impossible combinations obscures the decisions. Conversely, studying only the final buildout can miss an early phase in which fewer loads or a different network arrangement produces the more restrictive condition.

Existing sites offer a useful starting point for measurements. Capture enough operating context to interpret the data: equipment status, load level, capacitor switching, and measurement location. For a new site, identify the utility information available and the uncertainty that commissioning measurements will need to resolve.

Select mitigation after identifying the mechanism

A harmonic filter should address a demonstrated problem. Possible measures include changing equipment specifications, adding suitable impedance, revising capacitor arrangements, or applying passive or active filtering. Their suitability depends on the harmonic spectrum, network response, operating range, and installation constraints.

Passive filters interact with the network and require evaluation across relevant configurations and component tolerances. Active filters have current and control limitations that must be checked against the intended duty. A solution sized for one operating point may provide insufficient margin elsewhere.

Include equipment loading in the evaluation. Harmonic currents can affect transformer, cable, and capacitor duties even when attention is concentrated on bus voltage distortion. ETAP’s harmonic-analysis documentation includes source modeling, distortion calculations, equipment loading indicators, and filter evaluation. ETAP harmonic analysis.

Carry the study into acceptance

The deliverable should connect each concern to a location, operating case, acceptance criterion, and proposed action. Preserve the equipment spectra, network assumptions, uncertainty cases, and mitigation parameters so later substitutions can be assessed.

Specify how commissioning measurements will be compared with the study. If the facility opens in stages, explain which conclusions can be verified at initial loading and which need follow-up as capacity increases. A low-load measurement cannot demonstrate every future operating condition.

PowerTek supports harmonic analysis and power-quality studies for data centers, industrial facilities, and energy projects. Before accepting equipment submittals or selecting a filter, establish the network conditions and acceptance basis. Harmonic analysis in power systems should leave the owner with a defensible design decision and a practical way to verify it.