Power Quality Monitoring Guide

Power Quality Monitoring: A Practical Guide for Plant Engineers

Power quality monitoring is the systematic measurement of voltage and current characteristics at selected points in an electrical installation, with the goal of explaining equipment behaviour that the supply itself cannot account for. In industrial plants, most power quality work begins with a symptom — a drive tripping without an obvious cause, a contactor chattering, capacitors overheating, or a PLC rebooting during a shift change — and ends with a measurement campaign that either confirms or rules out a supply-side disturbance. This power quality guide sets out the disturbance categories you will meet, how to recognise their symptoms, how to decide between a portable survey and fixed instrumentation, and how to read the first report without over-interpreting it.

No single instrument answers every question. The choice between a portable analyzer carried from panel to panel and permanently installed monitors is a measurement-strategy decision, not primarily a purchasing one, and it should follow from the question being asked. The sections below treat that decision explicitly.

Disturbance Categories and What They Look Like in a Plant

Power quality disturbances are conventionally grouped by duration and by whether they affect voltage, current, or waveform shape. The taxonomy below is the one used in general engineering practice; acceptance limits depend on the applicable standard and on the sensitivity of the connected load, and should be taken from the standard that governs your installation rather than assumed.

Voltage events: sags, swells and interruptions

• Sags — a short-duration reduction in RMS voltage. Symptoms include contactors dropping out, VFDs tripping on DC-bus undervoltage, and IT equipment restarting. Sags are usually caused by faults elsewhere on the network or by the starting of large motors.
• Swells — a short-duration rise in RMS voltage, often following a sag or accompanying load rejection and capacitor switching. Sensitive electronics may trip on overvoltage protection.
• Interruptions — loss of supply voltage for a defined period. Distinguished from sags by magnitude rather than by cause; both may originate from the same upstream fault.

Waveform distortion: harmonics and interharmonics

Harmonics are sinusoidal components at integer multiples of the fundamental frequency, produced mainly by non-linear loads such as rectifiers, variable-frequency drives and switch-mode power supplies. In a plant they show up as overheating of transformers and neutral conductors, capacitor bank stress, resonance with power factor correction equipment, and unexplained breaker trips on circuits whose RMS current appears within rating. Interharmonics, at non-integer multiples, can produce flicker and sub-synchronous effects in some drive configurations.

Transients, flicker and unbalance

• Transients — very short impulsive or oscillatory deviations, typically from switching operations, capacitor energisation or lightning. They may not disturb RMS readings at all, yet can damage insulation or reset electronics.
• Flicker — repetitive voltage fluctuations in the frequency range perceptible as lamp flicker. In plants, the source is often an arc furnace, welder or a motor with a pulsating load.
• Unbalance — unequal phase voltages or currents. It causes negative-sequence current in induction motors, additional heating, and derating. Single-phase loads distributed unevenly across phases are a common cause.

Diagnosing Symptoms: From Complaint to Measurement Plan

A useful first step is to convert the complaint into a measurable quantity. "The line trips at random" is not measurable; "the drive trips on overcurrent roughly twice per shift, and only when the adjacent welding set is running" is. The table below maps common plant symptoms to the disturbance categories most worth investigating first. It is a starting point for a measurement plan, not a diagnosis.

Before you connect anything

• Define the question and the observation window. A disturbance that occurs twice per shift needs days of recording; a steady-state harmonic problem can be characterised in hours.
• Identify the point of common coupling and the boundaries of the installation, so that disturbances originating upstream can be separated from those generated on site.
• Check instrument rating and connection method — voltage measurement directly or via potential transformers, current via CTs or Rogowski coils — before energising.

Portable Analyzer or Fixed Monitoring?

The decision between a portable power quality analyzer and permanently installed monitors follows from three questions: how often the event occurs, whether the cause is known, and whether the point of interest is accessible during normal operation.

When a portable analyzer is the right instrument

• The disturbance is reproducible on demand, or occurs frequently enough to be captured during a scheduled survey.
• You need to compare several candidate locations — feeder panels, drive enclosures, the incoming switchgear — before committing to permanent instrumentation.
• The investigation is a one-off: commissioning checks, acceptance measurements after a modification, or verifying a contractor's work.
• You need waveform-level detail and event capture on a temporary basis. A portable unit with harmonics, event capture and a touchscreen interface supports on-site inspection of waveforms while the connection is still in place.

When fixed monitoring is the better choice

• The event is intermittent and rare. Waiting for it with a portable instrument is impractical; continuous recording is the only reliable way to capture it.
• You need a trend over weeks or months — for example, to establish whether harmonic distortion is rising as new loads are added.
• The measurement point is a critical feeder or the main incomer whose behaviour must be documented continuously.
• Multiple points must be correlated in time to establish the origin of a disturbance. A fixed monitoring line is positioned for this kind of permanent, multi-point installation, with the specific channel count and communication options of each model taken from its datasheet.

In practice the two approaches are complementary: a portable survey locates the problem, and fixed monitoring then documents it continuously or verifies that a remedy worked.

Where to Place a Fixed Monitor

Placement determines what the data can prove. A monitor at the wrong point will record disturbances that are real but irrelevant to the complaint, or miss the ones that matter.

• At the main incomer — establishes the supply-side baseline and separates upstream disturbances from those generated inside the plant.
• At the point of common coupling with other users — needed when the question is who is causing a disturbance, and useful where a dedicated transformer serves the site.
• On feeders serving disturbing loads — drives, welders, arc equipment, large motor groups. This is where harmonic and flicker sources are usually found.
• On feeders serving sensitive loads — PLCs, instrumentation, IT and control supplies. This is where the consequences appear.
• Adjacent to capacitor banks and filters — to check resonance conditions and switching transients.

Two further rules apply. First, place the monitor where it can be read and maintained without a shutdown, and where the CT and voltage connections are already available or can be added safely. Second, record the placement and the connection details (CT ratio, VT ratio, phase orientation) as part of the record; a report without this metadata cannot be interpreted later.

What to Record, and for How Long

Recording duration should be matched to the recurrence of the event, not chosen by habit. As a general guide:

• Steady-state characterisation (harmonics, unbalance, slow voltage variation): one full production cycle, including the heaviest and lightest load periods, is the minimum useful window.
• Intermittent events: record until the event has been captured several times, so that the pattern — time of day, associated load, preceding switching — becomes visible. Days to weeks is common.
• Trending: weeks to months, with aggregated values stored at intervals rather than full waveforms, to keep data volume manageable.

Parameters and settings to agree before the campaign starts

• Which quantities are logged: RMS voltage and current per phase, harmonic magnitudes and phase angles, unbalance, flicker, frequency, and event records with pre- and post-trigger waveform capture.
• Aggregation intervals — short intervals for event analysis, longer intervals for trending.
• Trigger thresholds for event capture. Set these deliberately and document them; thresholds chosen to match a standard's compatibility levels are more defensible than arbitrary ones.
• Time synchronisation across multiple instruments, so that records from different points can be aligned.

A common failure mode is recording too much waveform data at one point and too little context everywhere else. Distributing a few instruments across the relevant points, each with a clear logging plan, generally produces a more interpretable result than one instrument recording everything.

Reading a First Report

A first power quality report should be read in a fixed order, so that conclusions are not drawn from a single striking waveform.

• Check the metadata first. Confirm the measurement point, connection method, CT/VT ratios, recording period and trigger settings. Errors here invalidate everything downstream.
• Establish the baseline. Look at the steady-state statistics over the whole period before examining events. This tells you what "normal" was for that point.
• Correlate events with plant activity. Compare event timestamps with production logs, shift changes, motor starts and switching operations. A sag that coincides with a large motor start on the same feeder has a plausible local cause.
• Distinguish cause from consequence. A voltage dip recorded at a sensitive-load panel may have originated upstream; a harmonic current recorded at a drive feeder may be flowing toward the supply. Direction and location matter.
• Resist single-event conclusions. One captured transient is an observation, not a diagnosis. Look for repetition and for the conditions that accompany it.

Finally, record what was measured and what was concluded separately, and keep the raw data. A later investigation will often need to re-examine the same record with a different question in mind.

Selection Checklist

The following checklist summarises the decisions covered in this power quality guide and can be used to scope an investigation before any instrument is ordered or connected.

• Complaint converted into a measurable quantity, with a defined observation window.
• Disturbance categories shortlisted from symptoms, not assumed.
• Measurement points identified, including the point of common coupling and the boundaries of the installation.
• Portable survey versus fixed monitoring decided on the basis of event frequency and required duration.
• Instrument ratings, connection method and safety arrangements verified for each point.
• Logging parameters, aggregation intervals and trigger thresholds agreed and documented.
• Time synchronisation planned if more than one instrument is used.
• Production and switching logs available for correlation.
• Raw data retention and report structure agreed with the people who will act on the findings.

Instruments for Power Quality Work

LOGGON TECH supplies instrumentation for industrial metering and monitoring. For survey work, the APM556 analyzer provides harmonics measurement, event capture and a touchscreen interface for on-site inspection of recorded waveforms. For permanent installations, the LOGGON-310 fixed monitoring line is positioned for continuous measurement at selected points in the installation; channel count, communication options and other specifications for each model are given in its datasheet.

Further technical information is available on the power quality monitoring page, and analyser details are listed under power quality analyzers. Selection of the appropriate instrument should follow from the measurement plan described above rather than precede it.

FAQ

How long should a power quality monitoring campaign run?

Long enough to capture the event or the load variation being investigated, and long enough to establish a baseline. Steady-state characterisation typically needs one full production cycle including heavy and light load periods; intermittent events may require days or weeks of continuous recording until several occurrences have been captured. The duration should be set from the recurrence of the symptom, not from a fixed default.

Can a portable analyzer replace permanent monitors?

For reproducible or frequent disturbances, and for one-off surveys and commissioning checks, a portable analyzer is usually sufficient and more flexible in where it can be connected. For rare intermittent events, long-term trending and multi-point correlation, fixed monitoring is the appropriate approach because it records continuously without an operator present. Many investigations use both: a portable survey to locate the problem, fixed monitoring to document it over time.

What should be checked first when reading a power quality report?

The measurement metadata: point of connection, CT and VT ratios, phase orientation, recording period, aggregation intervals and event trigger thresholds. If any of these are wrong or undocumented, the recorded values cannot be interpreted reliably. Only after the metadata is confirmed should the steady-state baseline and then the individual events be examined and correlated with plant activity.