How a power quality root cause analysis finds the real fault
    Power Quality

    How a power quality root cause analysis finds the real fault

    Raveenthan Thambirajah
    August 11, 2026

    Transformers kept failing and the newest equipment got the blame. How a power quality root cause analysis finds the real fault, and how long it takes.

    Quick answer

    A power quality root cause analysis uses high-sampling analysers to record voltage and current at several points in a system for weeks, then lines the recordings up against every failure and switching event. It replaces finger-pointing with measurement: either the suspected equipment is producing the disturbance, or the record shows what is. That keeps repair money on the real cause.

    Key takeaways

    • Repeat transformer failures have at least five candidate causes, from harmonic heating and overloading to a defect common to the units, and a walk-round inspection cannot tell them apart.
    • Intermittent faults usually need several weeks of recording, so the data covers the loading and switching conditions under which the failures happen.
    • The method works in both directions: the record can clear the suspected equipment or confirm it, and either result stops money going on the wrong fix.

    A power quality root cause analysis finds out why equipment keeps failing by measuring the system instead of arguing about it. High-sampling analysers record voltage and current at the suspect points for weeks, and the recordings are lined up against every failure and switching event. Either the suspected equipment is causing the disturbance, or the record shows what is. If you have a repeat failure now, send us the trip log and a line on the fault and we will tell you what we would measure first.

    When critical power equipment fails, the investigation usually starts with a verdict. Something went down, the outage cost money, and everyone on the call already suspects the newest or most complicated thing on the network. We have been on the receiving end of that. At one site, the transformers feeding a set of our DSTATCOMs kept failing, and the customer was adamant the DSTATCOMs were doing it: electrical noise, switching effects, something. This page is for asset owners and site engineers stuck in that kind of dispute, and the investigations it describes are run by our engineering consultancy team.

    Why the newest equipment gets the blame

    Blame in these situations follows a pattern. Power electronics are a black box to most site teams, so a device that switches thousands of times a second makes an easy suspect. The failures started after the new equipment arrived, so the new equipment must have done it. And every supplier involved would rather the fault were someone else's.

    It happens a lot with critical power equipment: there is an outage or an alarm, or another failure in the same substation, and the pushback lands on whichever product is newest. Meanwhile the cost of guessing wrong grows. Replace a transformer without finding the cause and the new one inherits the same conditions. The failure repeats. Blame the compensation equipment wrongly and you pay for modifications that change nothing while the real cause keeps working. Our guide to what poor power quality costs a site covers the wider bill.

    If that sounds like your site, send the failure history and trip log to sales@severnde.co.uk (UK) or sales@renewable-d.com (US), and an engineer will tell you what we would measure first and where.

    What high-sampling monitoring sees that a site visit cannot

    A high-sampling power quality analyser records voltage and current continuously, in enough detail to catch disturbances that are over in a few thousandths of a second. Most of the plausible culprits are invisible to monthly readings and control room averages. Harmonic distortion can be measured against published limits (IEEE 519 in the US, the IEC 61000 series in Europe) and traced to its source by comparing readings at different points. Transients, sags and swells get timestamps that can be lined up against equipment operations and failures, and loading and imbalance patterns show slow thermal stress that no single reading would.

    A repeat transformer failure is a good example, because the list of possible causes is long and each one leaves a different trace.

    Possible causeWhat to measureWhat it looks like in the record
    Harmonic heatingVoltage and current harmonics at the transformer and at the suspected sourceHigh distortion while the load is well below the nameplate rating, and it tracks when the source runs
    Sustained overloadingThe load profile against the rating, over weeksLong periods above rating, often at the same times of day
    Voltage transients from switchingHigh-speed capture at the transformer terminalsSpikes timed to capacitor or breaker switching elsewhere on the network
    Loose or overheating connectionsThermal imaging under load, alongside the analyserHot joints; the analyser may show little
    Insulation ageing or a defect common to the unitsOil and insulation tests, and failure records across the unitsNothing unusual in the electrical record, with failures clustered by unit type or age

    Analysers left in for weeks rather than hours show which of these is present, and which can be ruled out.

    What we measured at the site with the failing transformers

    At the site with the failing transformers, we did not argue the point. We put a set of high-sampling power quality analysers on the system and monitored what was happening on it.

    The test cuts both ways, and that is why it works. If the recordings had shown the DSTATCOMs injecting damaging distortion, the next step would be simple: the customer is right, here is the fix, and we do it. If the recordings show nothing coming from the equipment that could explain the failures, the cause is elsewhere and the budget goes there instead of into modifying equipment that was doing its job. On a job like this, the only way to deal with the dispute is to do some troubleshooting and some root cause analysis.

    How a root cause investigation runs

    1. Frame the failure. Gather the history: what failed, how often, what changed on the network beforehand, and what the trip records show. That gives the list of candidate causes.
    2. Instrument the system. Install high-sampling analysers at the points that can tell the candidates apart: the suspect equipment's terminals, the failing equipment's supply, and reference points elsewhere on the network.
    3. Record long enough to catch the pattern. Intermittent problems need weeks, sometimes longer, because the recording has to cover the operating conditions under which failures happen.
    4. Analyse and assign the cause. Compare the measurements against the standards, line events up with operations and failures, and follow each disturbance back to its source. The output is a finding you can act on. Where a dispute has hardened, it is also evidence solid enough for warranty and liability conversations.

    Persistent harmonic problems often end in a filter, covered in our guide to harmonic filters, and voltage sag problems often end in compensation, compared fix by fix in our guide to why voltage sags trip drives and how a DSTATCOM stops them.

    What drives the cost of an investigation

    The cost depends mainly on how many points need an analyser, how many weeks they stay in, whether fitting them needs an outage, and how far the analysis has to go. Set that against what one more failure of medium voltage equipment would cost.

    The same method for inverters, UPS and other power electronics

    A power system root cause analysis follows the same four steps whether the equipment failing is a transformer, an inverter, a UPS or the wider network. Transformer root cause analysis leans on the thermal and harmonic record. Root cause analysis for power electronics, inverters included, leans on switching events and transients captured at the terminals, and often on taking the failed part apart.

    That last step can settle it on its own. A vented capacitor was the root cause in a recent power conversion module failure we investigated. Burned-out equipment gets the same evidence-first treatment before anyone decides between repair and replacement, as in our on-site rebuild of a DSTATCOM inverter section.

    What to do next

    1. Write down the failure history. What failed, the dates, and anything that changed on the network before it started.
    2. Pull the trip and event logs for the failing equipment and for anything you suspect is causing it.
    3. Send both to us with a line on what you think is happening, and we will tell you where we would put the analysers and for how long.

    Equipment failing and nobody can say why? We run power quality root cause investigations that record the system for as long as it takes to catch the fault, then tell you plainly what the record shows, including when it points to our own equipment. Send the failure history and trip log, and we will tell you what we would measure first. Send an enquiry, email sales@severnde.co.uk (UK) or sales@renewable-d.com (US), or call +44 116 279 6131 (UK) or +1 414 562 6559 (US).

    Frequently asked questions

    What causes transformers to keep failing?

    Repeat transformer failures usually trace to a persistent condition rather than bad luck: harmonic currents causing extra heating, sustained overloading, loose connections, voltage transients from switching, insulation ageing, or a defect common to the units. Because the list of candidates is long, measurement over several weeks is the reliable way to find which condition is present.

    How does power quality monitoring find a root cause?

    High-sampling analysers record voltage and current waveforms continuously at several points in the system. Engineers compare the records with standards such as IEEE 519, line up each disturbance with equipment operations and failure times, and trace the disturbance back to where it starts. The result either points to a source or rules it out with data.

    How long does power quality monitoring need to run?

    Long enough to capture the conditions under which the problem happens. For frequent disturbances a few days can be enough. Intermittent failures usually need several weeks, so the record covers the full range of loading and switching, and seasonal problems may need longer still. The monitoring period is set after the failure history has been reviewed.

    Tags

    power quality root cause analysispower quality analysistransformer failure causesharmonics transformer overheating

    Working on a project like this?

    Talk to our engineering team about your site, your grid connection or your power quality problem.

    Get in touch
    GET IN TOUCH

    Let's Solve Your Power Challenge.

    Send an enquiry to our engineering team. We'll help you identify risks and uncover efficiency opportunities.

    North America

    Renewable Energy & Drives

    5325 S Moorland RdNew Berlin, WI 53151United States

    UK & Europe

    Severn Drives & Energy

    Units 2 & 3, Kingsley Business ParkNew Road, Kibworth BeauchampLeicestershire, LE8 0LEUnited Kingdom

    Send an enquiry

    Fill out the form below and we'll be in touch.