Transformer Monitoring Breakthrough: Why Traditional Methods Miss Critical Failures

0:00
Welcome to the deep dive the show where we cut through the noise, get straight to the core insights from a whole stack of sources, and basically give you that shortcut to being genuinely well informed. Today, we’re tackling something that well, it probably keeps quite a few asset managers and operations execs awake at night. I’m talking about the unexpected failure of power transformers. Yep, you know how it goes. Right? Invest in maintenance. You’ve got your oil analysis data, you’re checking temperatures, but then, bang, a key transformer fails, seemingly out of nowhere. It’s costly, it’s frustrating, sometimes it’s even dangerous. So the big question is, why does this keep happening? The story we hear a lot isn’t it, you’re doing everything right? You think running your dissolved gas analysis, keeping an eye on temperatures. But our sources point to this really compelling case study from South Africa’s Northern Cape operating unit. They had this plant health indicator tool, the phi tool. And get this in some cases, it rated transformers as perfectly healthy category, a literally right before they failed catastrophically.

0:58
That’s not just a blind spot that feels like a, well, a complete failure of the monitoring approach, doesn’t it? It

1:04
really is a major disconnect. And the thing about that phi tool in that specific situation was its narrow focus. It was only looking at a few DGA parameters. What it completely missed were these really critical underlying factors, you know, things like how the cooling system was actually performing, the real health of the bushings, the tap changers, and not just basic temperatures, but the specific operating temperatures, the hot spots, and crucially, the actual rate of ageing, the loss of life happening in the Transformers insulation in real time and the well, the surprising thing the study revealed was that Most of those surprise failures, they were directly linked back to these exact things that weren’t being

1:45
monitored. Okay? So the monitoring was looking over here while the real problems were happening over

1:49
there, precisely, and the consequences, they’re really severe. We’re not talking small issues here. These failures lead to unplanned outages, which obviously hit the bottom line hard, significant costs from the main equipment, damage, sure, but also collateral damage to other G in the substation, then you need more spares, which are expensive and emergency transport. It all adds up. But honestly, beyond the money, there’s a serious risk to people. We’ve seen cases, unfortunately, of explosions, fires, things like porcelain insulators, heavy metal parts being thrown, you know, up to 60 metres, 60

2:24
metres, wow. Okay, that really changes the perspective. It’s not just about cost avoidance. Then it’s fundamental safety. Absolutely,

2:29
it’s a profound safety and reliability issue that these, well, traditional methods just aren’t equipped to catch early enough.

2:36
And you mentioned the South African grid specifically. What is it about a dynamic environment like that and maybe others like it that makes transformers particularly vulnerable?

2:45
That’s a great question, because it’s not just South Africa anymore, is it? Lots of grids are becoming more volatile, but in South Africa you have high demand, increasing failure rates. It’s kind of a perfect storm for transformer stress. The Source talks about grid volatility just not going away, and these constant changes, the fluctuations they create, this invisible electrical stress. Think about everyday power quality issues like voltage dips, unbalance, deviations from that nice, clean sine wave,

3:15
right? Things that might not seem like a huge deal individually, but they add up exactly. They

3:19
accumulate. Then you add in things like rooftop solar, decentralised generation popping up everywhere. Networks weren’t built for that. You know, they were designed for power flowing one way from the big plants downstream. Now energy is being pushed back up from the edges, which messes with its systems design parameters. It certainly challenges them and introduces, yeah, new power quality problems, and let’s not forget industrial loads, things like variable frequency drives, VFDs, they create harmonics, distortion. We call them dirty feeders sometimes. So all this stuff, this persistent, unseen electrical stress, it just builds and builds it hammers away at the transformers. Insulation accelerates the ageing that expected 35 year life. It can get drastically shortened. And the key point is your traditional tools, your DGA, your basic temp checks, they just don’t see this bigger picture of cumulative stress.

4:09
And that lack of visibility leads directly to those surprise failures. That’s the link. These hidden stresses aren’t just theoretical. Are they the numbers you mentioned from es com is Northern Cape unit really drive it home. Only 5% of the transformers, but nearly 2% 1.96% to be exact, of all the failures across the entire es com distribution network over 12 months. Yeah, it’s

4:31
disproportionately high. It screams that something specific is happening in that environment stressing those particular assets more. And it wasn’t just end of life failures, either. You said, No, not at all. The data was clear. 46% were winding failures. 36% involved bushings, and 9% were tap changer issues. So

4:48
specific components failing under stress, exactly, which

4:51
tells you these aren’t just old assets wearing out uniformly. They’re failing because specific parts are under duress, which actually points towards the point. Possibility of detecting these issues earlier if you’re looking at the right things. Okay,

5:04
so if the traditional data, the oil samples, the temperature checks, if they’re giving us an incomplete, maybe even misleading picture, what do we need? How do we get that early visibility into this, this invisible stress? How do we finally stop these surprises? Well,

5:20
the answer really lies in moving towards more sophisticated, fixed power quality monitoring systems. That unseen stress we’ve been talking about. It has a direct line to unplanned failures and that reactive emergency spending nobody wants but fixed PQ monitoring. It changes the whole equation. It’s not just more data. It gives you the why. It helps explain why DGA values might be creeping up, or why you’re seeing sudden temperature spikes, or why a relatively young transformer is failing. It provides that missing

5:48
context so it connects the dots between the electrical environment and the physical

5:52
condition precisely, and the business benefits of getting that early warning are pretty significant. You can fundamentally shift your maintenance move away from just time based schedules, you know, doing things because the calendar says so, towards true condition based maintenance CBM. You intervene when the equipment tells you it needs attention, which sounds much more efficient it is, and it directly enables deferred capex. By understanding the real stress and managing it, you can potentially extend asset life, push out those hugely expensive transformer replacements, right? Delay the big spends. Yeah, and think about reputation. Minimising outages keeps customers happier, avoids potential penalties. That has real value too.

6:35
That financial case you mentioned earlier from the study was pretty stark, investing around 1.3 million Rand up front for the online monitoring system, yes, could save over 11 million Rand over 10 years. That’s more than 1.1 million Rand a year in savings compared to the nearly 1.3 million they were spending annually on the old way.

6:52
It’s a very compelling ROI. The numbers often speak for themselves once you analyse the total cost of failure versus the cost of monitoring.

7:00
So for an asset manager listening, thinking, okay, the ROI looks good. What’s the practical hurdle? Is it just the cost, or is there more to it?

7:08
That’s a fair question. The upfront cost is a factor, obviously, but like we said, the ROI often justifies it. I think often the maybe the bigger challenge is the shift in mindset. It’s integrating this richer, continuous data stream into existing workflows and decision making.

7:23
Ah, so it’s operational change, maybe cultural change, too. Exactly,

7:27
moving from being reactive, fixing things after they break, to being genuinely proactive, using data to predict and prevent but honestly, once you start seeing the insights this data provides, it tends to become indispensable pretty quickly. Okay, that

7:40
makes sense. It’s embedding it into the process. Into the process. So let’s talk specifics then, but maybe without getting lost in the engineering weeds. What exactly is this advanced PQ monitoring, seeing that the old methods Miss? What are these invisible stresses made visible?

7:56
Okay, good way to put it. So continuous, online monitoring is tracking several things constantly. For instance, in rushes and load volatility, think of these little electrical shocks repeatedly hitting the transformers. Insulation that paper. Insulation inside it undergoes something called depolymerization. Basically, it gets brittle over time due to heat and stress and fragile insulation can’t handle the mechanical forces during, say, a short circuit. There’s actually a rule of thumb, the paper’s life roughly halves for every six degrees Celsius. Its hot spot temperature goes above 98 degrees C. Monitoring tracks that heat stress accumulation. Got it so it sees that cumulative damage building up. What else it tracks voltage and current unbalance. When phases aren’t balanced, you get uneven heating inside the transformer that impacts performance, efficiency, stability. It’s a hidden stressor. Then you have harmonics and inter harmonics. These are basically distortions, frequencies on the power line that shouldn’t be there, and they can cause subtle problems like extra heating or even interference. There was a really fascinating finding In one study, it showed that sometimes uncured red oxide paint left inside transformer radiators from manufacturing could react and cause unnatural gas. Paint seriously, yeah, paint residue leading to higher levels of specific gases like ethane, methane, CO two in the oil. It’s a direct link between a tiny manufacturing detail and a DGA anomaly that only continuous monitoring, correlating gas levels with operating conditions could likely pinpoint,

9:27
that’s amazing. So it’s finding root causes that simple DGA alone would never reveal

9:31
exactly. We saw hints of this at that Kimberly distribution substation. KDs you mentioned, both transformers there showed odd ethane and methane levels, even though overall gases were low, pointing to that kind of subtle ongoing stress. Okay, what else does it capture? Spikes and transients? These are super fast events like lightning hitting nearby or switching big capacitor banks. They might seem brief, but they can jolt the system, excite natural frequencies and cause cumulative damage. This. System also actively checks the cooling system fans. It measures their current draw to make sure they’re actually running when they’re

10:06
supposed to, which sounds basic, but you’re saying it’s often overlooked,

10:10
incredibly overlooked. Often there’s no specific maintenance plan for the fans themselves. If they fail, especially in hot places like the Northern Cape, the transformer can overheat quickly, drastically shortening its life, even if it’s not overloaded,

10:24
right? The KDs monitoring actually confirmed the fans were working properly there, responding to temperature rises,

10:30
yes, which was good news, but it highlighted how crucial knowing that is, rather

10:33
than assuming makes sense. What about components like bushings? Ah, bushings,

10:38
critical components. Online monitoring gives you continuous data on their capacitance and dissipation factor, tan delta. We call it a change in capacitance, say, more than 10% can signal an impending breakdown, internal partial

10:51
discharge, and traditional testing misses this well, traditional offline

10:55
tests are maybe done once every six years, and you have to switch everything off take an outage. Online monitoring gives you real time data. 247, it spots tiny, gradual changes that those infrequent offline tests would completely miss. Plus it can monitor phase voltages on the high voltage side too, which isn’t typically

11:14
covered offline so much earlier warning without the disruption precisely at

11:19
KDs, the bushings were okay, but the demonstration of continuous monitoring, identifying subtle shifts was the key

11:26
takeaway. And tap changers, you mentioned, they were a big failure point.

11:30
Yes, on load, tap changers, oltcs, they regulate voltage, constantly moving so there’s mechanical wear. They account for about 9% of failures. Online monitoring can detect mechanical sluggishness, contact, wear, abnormal heating, even things like coking, which is like a carbon build up that stops them working properly, allows for planned intervention before failure.

11:50
Okay. And finally, temperature and ageing.

11:52
Right beyond just top oil temperature, it tracks the calculated hot spot temperature. That’s the real limit for the insulation, and crucially, it correlates this with ambient temperature. We saw this vividly at KDs during a heat wave, 44 degrees C ambient. Wow. That high ambient heat alone pushed the calculated hot spot temperature above the insulations rating, accelerating ageing, risking failure, even though the transformer wasn’t

12:20
overloaded, so the environment itself was cooking the transformer, essentially, in a manner

12:25
of speaking, yes, it emphasises why you absolutely need to monitor ambient conditions alongside the Transformers internal state. Traditional methods often miss that link. It’s

12:35
incredibly comprehensive. And I think what really stands out, as you said about the KDs examples, is getting all this insight without needing an adage that minimises disruption, keeps the power flowing and just gives you this continuous stream of actionable intelligence. It feels like moving from guesswork to genuine foresight. That’s exactly the goal. Which brings us nicely to the wrap up. Why is this shift towards comprehensive power quality monitoring so critical right now. What’s the key takeaway for our listeners, the asset managers, the ops execs, the planners?

13:06
I think the core message is pretty straightforward. The grid isn’t getting any simpler. That volatility we discussed driven by renewables, changing loads, grid constraints, it’s here to stay. That’s the new reality. We can’t turn back the clock Exactly. But the good news is, with tools like fixed power quality monitoring, you can now see what’s happening. You can quantify the previously invisible stresses that are ageing your transformers prematurely. You can connect cause and effect.

13:32
You’re moving from the unknown to the known

13:34
precisely. And this leads me to, well, maybe a slightly provocative thought for everyone listening, if you wait for your oil data, your DGA results to tell you there’s a serious problem you’ve probably waited too long. Expand on that. Think about the timelines. DGA samples, maybe once a day, if you’re lucky, often less frequent. Lab analysis takes time. Bushing tests once every six years, maybe requiring an outage. It’s like looking at a single photograph taken last week versus watching a live high definition video feed of your assets, health, two, finer, four, seven.

14:07
That’s a powerful analogy. The snapshot versus the continuous movie

14:10
right online. PQ monitoring gives you that movie. It shows the trends, the correlations, the early warnings in real time. It gives you the chance to act before the failure, not just react afterwards.

14:21
So wrapping it all up, this comprehensive approach, it really empowers everyone involved in asset management, from operations to planning to finance. It lets you move beyond just fixing things when they break towards genuinely strategic, proactive management of these critical, expensive assets.

14:37
Absolutely, it’s about foresight, informed decisions and ultimately better outcomes,

14:42
leading to very real benefits and cost savings, definitely in risk reduction and enhancing the overall reliability of the power infrastructure you manage. So the final thought is, perhaps, if you’re not actively looking at the power quality stressing your transformers, those hidden factors are likely already impacting. Lifespan and setting you up for the next costly surprise. Maybe it’s time to start watching the movie, not just waiting for the snapshot.

What Power quality issues might be silently impacting the reliability, the lifespan, and perhaps even the safety of equipment in your complex plant, mine or data centre right now, especially if you’re not monitoring it continuously, right down to the waveform level?

0:00
Okay, let’s unpack this. Imagine you’re a power system engineer. You’re entrusted with managing, well, a whole fleet of incredibly expensive, absolutely mission critical assets, power transformers, and you’re doing everything by the book, regular dissolved gas analysis, DGA, as it’s known, routine oil sampling, continuous temperature checks, the works. You might even have one of those advanced plant health indicator tools, you know confidently, flashing a green light, declaring everything is category A perfectly healthy but here’s where the picture gets profoundly interesting, maybe even a bit unsettling. Despite all this vigilance, all this effort, you’re still getting hit with unexpected transformer failures. And these aren’t minor glitches. We’re talking catastrophic events, events that trigger massive outages, huge repair costs, even fires that pose a very real danger to people and damage surrounding gear for major utilities like es com, Northern Cape operating unit down here, this isn’t just some baffling technical puzzle, it’s a systemic problem costing millions and really hitting customer satisfaction hard. It’s a well a harsh reality. So this deep dive is really for you, the person who needs to understand the why behind these failures, what critical pieces of information might be missing from your current health assessment. We’re going to explore how focusing only on those traditional physical diagnostics, vital as they are, might be leaving, well, half the story untold, leaving your fleet vulnerable.

1:20
That’s a really crucial observation. It’s fascinating and honestly often frustrating, how our assumptions, our established ways of doing things, can create these, these significant blind spots. We get used to seeing the symptoms in our tools. You know, the gas levels rising in the DGA, the temperature creeping up, but we don’t always connect that back to the root causes that are actually driving the degradation. It’s a bit like like a doctor just treating a fever without digging deeper to see if it’s, you know, just a cold, or something more serious underneath, we react to the symptom, not the underlying issue that’s silently chipping away at the assets life. So today, we’re going to dive deep into what that missing half really is. We’ll explore how integrating high resolution, continuous monitoring of the electrical stresses, what we broadly call power quality, how that can complete the picture. This gives you the foresight, hopefully, to genuinely predict and prevent these catastrophic failures, moving beyond just reacting to symptoms towards a proactive root cause understanding. We’ll look specifically at the demanding realities of the grid here in southern Africa and how a more well holistic monitoring strategy can make a huge difference for your asset plan,

2:28
right? We sort of set the stage with the problem, but let’s really unpack the scale of it. Why are these unexpected failures such a huge deal, especially here in southern Africa, it feels like it carries a unique weight here. Think about the global picture. First, did you know that in many countries, something like over 70% of the installed power transformers have been running for more than 25 years? That’s a massive amount of critical infrastructure, well into its, let’s say, mature phase, definitely past its youth, exactly. And while the you know, the textbook estimated service life might be around 35 years under perfect rated conditions, the reality of grid operations with all the stresses can really shorten that lifespan.

3:10
Absolutely ideal conditions rarely exist in the real world. It’s

3:13
like an asset designed for a calm lab environment being forced to operate in a storm, day in, day out. That alone creates huge stress. Now let’s bring it home to South Africa. Es com distribution networks alone, they manage this enormous fleet, over 4100 power transformers, from one MVA up to big 160 MVA units.

3:34
That’s a massive installed base. It really is.

3:36
It’s a critical national asset base, so the stakes are incredibly high. When even one fails, sends ripples right, affecting reliability, supply stability, customer satisfaction. And the alarm bells really start ringing when you look at specific regions. Take the Northern Cape operating unit, the NCO you they operate about 5% of ES com distributions. Total transformers, roughly 200 units. But here’s the kicker, the really critical part from the CT report we looked at the NCUA, despite having only 5% of the fleet, contributes almost 2% 1.96% to be exact, of all es com distribution failures over a 12 month moving average. Wow, that’s disproportionately high, hugely disproportionate. It’s a clear red flag that something fundamental isn’t right there. It demands attention beyond just the usual checks, suggests a vulnerability and their existing system, the plant health indicator, the phi tool, it relies almost entirely on DGA to predict health. But here’s the problem, the CT report highlighted Transformers rated category A meaning healthy by this tool at the time of failure, still went on to fail catastrophically.

4:39
So the tool was saying green light right before disaster struck Exactly. It’s

4:43
not just a one off, it’s a pattern. It clearly points to a critical gap in their monitoring. It suggests the phi, while maybe good for DGA, isn’t capturing the full dynamic picture, leaving them wide open.

4:53
That really raises a huge question, doesn’t it? If our trusted tools say all clear, but failures keep. Happening. What are we fundamentally missing? It’s obvious the current snapshot isn’t capturing the real forces at play and beyond the immediate power outage, these failures, they have profound impacts, really far reaching. First, you’ve got the safety and environmental risks. They’re immense. Transformer failures aren’t quiet. They can be incredibly violent explosions throwing heavy porcelain bits maybe 60 metres. Think about the danger to people nearby.

5:24
Yeah, terrifying, especially for crews working there.

5:27
Absolutely and fires are a huge concern. The report mentioned 37% of major bushing failures end in fire. It’s not just about equipment. It’s about human safety, plus the environmental side, huge oil spills, contamination, really complex and costly cleanups. Then there’s the economic devastation. It’s multifaceted and frankly, crippling. You have the obvious direct cost repair replacement. The NCLU spends heavily replacing failed units, often way before their expected end of life. Turns an asset into a recurring expense, instead of a long term investment, precisely, but it goes beyond that. Explosions often cause massive collateral damage, destroying neighbouring gear, switch gear, cables, CTS, other vital stuff that jacks up the repair costs and the downtime. Prolonged outages and a high failure rate puts immense supply chain strain on utilities, more spares needed tying up capital. When a unit fails in the NCUA, it can even deplete spares from other areas, leading to higher transport costs, longer restoration times across the whole network, a real domino effect, right? Cascading problems, exactly. And finally, utilities face big penalties and reputational damage for long outages, impacts KPIs, public trust, customer satisfaction, preventing failures directly saves money on claims and environmental issues too. And this brings us to those silent stressors in the South African grid. While a C u t report focused a lot on physical issues, other guides like afscc and NCP CSA really highlight the underlying electrical problems. The Grid environment here is challenging, harsh, often, ideally, Transformers want stable, smooth power, which they rarely get, pretty much the grid, is dynamic, frequent voltage sags, swells, lots of transients, those fast, short, undesirable events the AFSC guide mentions, plus power frequency variations from load and generation imbalances, very noticeable during instability and the increasing amount of embedded generation solar farms, rooftop PV, plus all The VFD driven loads, they pump significant harmonic distortions and unbalances into the

7:23
system. So it’s not just load, it’s the quality of the power itself.

7:27
Precisely, these are constant, aggressive electrical stressors, silently shortening transformer life, accelerating ageing. But as the CT report implies, they often aren’t monitored with the same detail as oil or temperature a huge oversight. We’re only seeing the tip of the

7:43
iceberg. So what does this really mean for an asset manager if these critical, expensive transformers are failing despite healthy reports? It really suggests our standard monitoring isn’t giving us the whole truth. It’s like like having a doctor who only checks temperature and blood pressure but ignores diet, stress, genetics, the things silently impacting long term health. Let’s break down what those traditional tools do measure and more importantly, what crucial info they consistently miss. First dissolved gas analysis DGO look. It’s crucial, absolutely. The CT report calls it the most important and most effective method. It detects potential internal faults by identifying gases from an insulation breakdown, saline, methane, ethylene, hydrogen, CO, CO, two essential early warnings. Can’t argue with its importance, no, but, and this is the key distinction, DGA mostly show symptoms. The gases are a result of a problem, right?

8:33
It doesn’t always directly reveal the electrical stresses that are often the root cause. For example, high gas levels might indicate overheating, okay, but DGA alone doesn’t tell you why it’s overheating. Is it simple overload, a cooling fan failure, or something more subtle, like continuous harmonic currents or severe unbalance, pushing it thermally, even at normal loads,

8:56
right? It identifies that there’s an issue, but not the complete why it’s the smoke, not necessarily the source of the fire.

9:03
And the CT report even noted a challenge traditional interpretation methods like Duval’s triangle or IAC guides sometimes couldn’t be fully used because gas levels were too low, even when problems were brewing. So it can be too subtle. Sometimes exactly, it hints at the limits of relying only on DGA. It gives a partial view, sometimes so faint, the alarms don’t trigger until it’s really progressed. That’s a crucial point, and it really begs the question, how do we move beyond just seeing the smoke to understanding and maybe preventing the fire, getting to that root cause? Let’s look deeper at what else gets missed. Take temperature sensors, oti. WTI, crucial for detecting heat. We know heat is the main cause of transformer ageing, as the report says, it causes that cellulose insulation, the paper to depolymerize,

9:49
basically gets brittle, breaks

9:50
down, exactly loses its strength. But these sensors, like DGA tell us that temperature is high, not always. Why is it just ambient heat, or are electrical and. Anomalies like extra losses from harmonics or core saturation from distorted voltage generating that internal heat without electrical context, we’re guessing. And a big gap the CT report found in the NCO U system was not monitoring cooling fan operation. Cooling is vital, especially in hot places. If fans fail undetected, the transformer age is much faster. Huge impact. Okay. What about relay data that seems important. It is absolutely critical for protection. Relays react fast defaults, trip circuits essential for safety, preventing wider damage. But relay data is primarily event based. It reacts to trips, big deviations. It doesn’t usually give you the continuous high res condition based data you need to spot gradual decline or subtle electrical issues before they cause a trip. It’s like waiting for the engine to seize before checking the oil. Good analogy, yes, and traditional bushing tests, tan delta capacitance often done offline every six years, like es com distribution does. That means their condition is unknown until the maintenance period arrives. Requires outages, leaves you blind for years. It’s reactive, not predictive, and the cost of this incomplete monitoring is clear. The CT report states that relying on limited monitoring, even with good maintenance practices, means utilities still face high costs from failures and unnecessary interruptions impacting KPIs that 35 year expected lifespan highly dependent on operational factors, often overlooked by symptom focused monitoring, utilities are leaving value on the table. Shorten asset lives, incurring avoidable costs. Okay?

11:26
So if the traditional tools are missing key pieces, only showing effects, not causes, what is the missing lens? We need to complete the picture. This is exactly where comprehensive power quality monitoring comes in. Let’s unpack what that really means, what hidden electrical stressors it reveals, and how these forces impact your assets. First defining power quality. Fundamentally, it’s about how much the actual AC voltage and current wave forms deviate from that perfect ideal sine wave, the textbook ideal exactly, ideally smooth, consistent magnitude, no extra frequencies. But the kicker is that ideal waveform doesn’t exist in the real grid. Our systems have loads of nonlinear gear switching loads, core saturation and transformers and motors, power electronics everywhere, industrial drives, solar inverters, all distorting

12:15
the waveforms

12:15
right, pushing them away from that perfect sine wave. And these deviations aren’t harmless. High resolution power quality monitoring systems capture continuous, detailed data on a whole range of these electrical phenomena that directly hit transformer health, often invisibly to physical sensors. Let’s look at some key ones. First harmonics. These are extra frequencies, multiples of the fundamental, 150 hertz, 250 hertz, 350. Hertz for a 50 hertz system, total harmonic distortion. THD measures the overall distortion,

12:47
and sources like VFDs and embedded generation, like rooftop PV, pump these out

12:51
exactly creating unseen losses, localised heating and transformers not designed for it, then voltage and current unbalance when the three phases aren’t perfectly equal in magnitude or angle, often due to uneven single phase loads or grid issues,

13:04
creates negative sequence components, bad news for performance and insulation leads to uneven heating, differential stress.

13:10
Next, transients, the afscc guy calls them undesirable, fast and short duration events can be impulsive, like lightning surges or oscillatory like capacitor bank switching.

13:20
Both cause sudden high current or voltage spikes like hammer blows, yeah,

13:24
stressing insulation, windings, causing fatigue, micro fractures. Over time. We also track voltage dips and spikes, sags and swells, momentary drops or rises in voltage, very common motor starts, fault clearing

13:37
seem fleeting, but repeated stress adds up, degrades insulation, accelerates ageing.

13:42
And finally, power frequency variation and phase angle drift. The afscc Guide mentions power frequency variations, deviations from the nominal frequency, often from generation load imbalances, like during load shedding.

13:54
Maintaining phase angle coherence is critical for stability and asset health. Significant drift puts mechanical and electrical stress on equipment. This is precisely where that high res PQ data becomes so valuable it moves us beyond just seeing a symptom like gas or heat, and explains the fundamental why behind what traditional tests show, it’s the direct, often invisible link between the grid conditions outside and the asset degradation inside. Each of these PQ issues contributes directly to that silent degradation, often without setting off alarms until it’s way too late. Let’s look at thermal stress and insulation breakdown. Harmonics, as you said, cause extra IRR losses, meaning even at nominal load, distorted currents cause additional often localised overheating that bulk temperature sensors miss the afscc guide notes. This causes additional temperature rise. It applies directly to transformers. This excess heat accelerates that depolymerization of the cellulose insulation the paper breakdown exactly, shortening transformer life dramatically. Remember paper lite halves for roughly every six degrees C, the hot spot goes over 98 degrees C, harmonics constantly push that limit. Burning it out silently. Voltage and Current unbalance also cause increased losses and uneven heating. The AFSC guide says unbalance hamper system performance and causes insulation problems. Imagine one phase consistently running hotter due to unbalance, even if overall load seems fine, that uneven stress causes premature breakdown localised weak spots, treats vulnerabilities precisely, transients and dip spikes, though short, have severe cumulative impact. Those sudden changes are like mechanical blows to windings and insulation cause fatigue, micro cracks over time. They can even excite natural frequencies, creating damaging resonance can lead to rapid failure. And secondly, these stressors lead to accelerated chemical ageing, increased thermal stress from harmonics, unbalanced transient speeds up chemical reactions in the oil and paper. Faster degradation of the whole insulation system manifests as more moisture buildup, which kills dielectric strength and faster gas production, even if the cause isn’t a traditional fault, but continuous grid stress. The CT report nailed it. Main cause of transformer. Ageing is isolation temperature. Cellulose undergoes depolymerization, caused by temperature and time. Anything driving up temperature, including electrical stressors, accelerates ageing, makes that 35 year life optimistic, high risk. EQ monitoring closes this gap, reveals these root causes, allowing proactive mitigation.

16:17
Okay, this is where the rubber meets the road. Theory is one thing, but do we have the data, the real world proof from actual operations showing power quality issues directly correlate with transformer degradation and failure? Yes, we absolutely do. And those are detailed studies from the Northern Cape the NCLU give us some compelling real life examples. Remember the NCO use phi system reporting transformers as healthy right before they failed. A deep dive into actual DGA results from transformers at the Kimberley distribution substation, KDs, the focus of that CT report study, showed interesting, initially baffling patterns. For instance, both transformer one and three at KDs consistently showed not normal ethane, c2 h6 and methane, CH four, slightly higher than normal production levels, not a huge spike, just persistently elevated, exactly, not enough to trigger big alarms. May be dismissed as minor in traditional checks. So what explains the subtle, ongoing gassing in otherwise healthy units? The report revealed a surprising root cause from lab analysis uncured red oxide paint inside the radiators, manufacturing or refurbishment issue, yeah, when this uncured paint hits hot transformer oil, it causes unnatural gassing, specifically ethane, CO two, methane, hydrogen. It’s a fascinating direct example, a hidden physical condition interacting with operational heat creates specific DGA symptoms.

17:39
Wow. Something continuous online DGA could track, especially if contextualised with PQ data showing localised thermal stress, maybe accelerating that reaction

17:49
precisely. It’s about why the gases are there, not just what they are. Impossible insight from periodic samples alone. Another practical example from the report, shows the direct impact of PQ on monitoring itself during data collection, the online monitoring server got overloaded with data, 25,000 events in six months, over 420 100 events a month. What caused that flood relatively large fluctuations in voltage on the transformers, high voltage HV side, not a fault, just grid instability. The system correctly logged deviations, but the sheer volume overwhelmed the database highlights how an unstable grid directly creates data challenges. Underscores the need for robust PQ measurement and intelligent processing. That

18:27
really gets to the heart of why combining these datasets, physical diagnostics and electrical environment is so powerful. It’s about seeing the full causal chain, not just isolated snapshots, transforming data points into the assets life story, let’s look deeper at ambient temperature and hot spot degradation, moving beyond just overload causing high temps. Data from KDs t1 and t3 during a January 2019, heat wave. Remember, Northern Cape gets extreme heat. 44 degrees C recorded showed something crucial, even when load current stayed constant, no electrical overload, both top oil and winding hot spot temps rose significantly just from the ambient heat. Exactly It shows ambient temp alone can push the hot spot above the papers rating accelerating degradation without overloading the transformer. The implication for hot climates is profound, often overlooked. This may result in premature failure due to extreme increase in ambient temperature. End quote, crucial for design and operation, online monitoring captures this dynamically allows accurate lifetime estimates, real time adjustments, periodic checks would miss this prolonged stress. Now consider why identical Transformers Age differently. Seems perplexing, right, right. Imagine two identical units, same design, same nominal load. One runs in a pristine electrical environment, stable voltage, no harmonics, perfect balance. The other operates in a harsher reality, frequent voltage swings, harmonics, unbalanced transients common in southern Africa due to load shedding in rush switching, embedded Gen VFDs,

19:53
the reality

19:54
for many assets here precisely, if you only look at traditional DGA or temp readings, they might look. Similar at one point in time, you’d think they’re ageing the same, but trend high. Res waveform data, alongside physical diagnostics and the true to merges, you see the continuous electrical stresses battering the second transformer. This granular data reveals it will experience accelerated insulation degradation, increased mechanical fatigue, even if its load looks similar. This explains why it might fail sooner or show worse. DGA changes over time. The ncpcsa Guide nails this stressing. The need for 12 cycle or better. RMS trend before, during and after the event. For root cause analysis, understand the direction the source this waveform data is exactly what differentiates how identical Transformers Age key to predictive management, tailor strategies based on actual stress, not just age or load profile.

20:43
Theory is great. Data is compelling, but how does this play out on the ground, managing real fleets in tough environments? Let’s look at how comprehensive monitoring is, proving its worth in actual South African settings, bringing clarity and enabling proactive strategies. The CT report started with that stark reality for the NCLU, their PHI tool rated transformers, category, a healthy despite failures. The report blamed this on the phi monitoring only certain DGA parameters and critically missing other vital aspects, cooling status, bushing health, tap, changer performance, detailed temps, ageing

21:16
calculations, the very things that were causing the failures.

21:18
Exactly, most ncou failures were linked to these unmonitored parameters. The KDs study was a direct response aiming to introduce comprehensive online monitoring to cut these failures. The study at KDs used an advanced integrated system, C tram, Kelman, DGA, bushing, monitoring devices, first time in Africa for this research context, continuous, 24 hour data on parameters traditional methods miss or only check periodically, often needing outages. Let’s talk direct fault detection. The report has these jarring images of a failed KDs transformer, severe damage, windings, yoke, pressure rings, carbon, molten copper everywhere. Really shows the violence of the failure. It does, and the report states explicitly, the DGA data from the online condition monitoring system could have assisted an engineer to remove the transformer before the

22:07
failure could happen. Wow, that’s a powerful statement, preventable with the right data.

22:11
Undeniable testament to predictive power, revealing what was missed, what could been prevented. The system also nailed, identifying subtle issues like that uncured paint the persistent, slightly high ethane and methane from online DGA and kdst one and p3 anomalies easily missed in periodic samples were linked to that paint.

22:30
Incredible detail shows how continuous monitoring plus deeper knowledge explains subtle anomalies, pinpoints root causes, quietly degrading the asset impossible with snapshots,

22:41
plus, it gave crucial insights into operational health of key components, bushings, for example, online monitoring at KDs showed consistent capacitance and dissipation factor for t1 and t3 bushings well within limits, ongoing assurance. More importantly, it monitored phase voltages continuously on the HV side, something es com strategy didn’t cover without outages. Online visibility spots potential issues early via subtle deviations, like a 5% change between phases,

23:05
huge advantage over offline tests every six years, absolutely

23:09
same for cooling fans. Continuous monitoring of fan currents confirmed they operated correctly responding to temperature even during that heat wave, vital since the report said failed fans significantly impact lifespan and loading, and shockingly, no maintenance strategy covering the cooling fans existed. Then online monitoring filled that huge gap, ensuring critical thermal management work right. And finally, real time temperature and ageing insights monitoring clearly showed how high ambient temps directly hit winding hot spots, even without overload, this dynamic data allows much more accurate lifetime estimation, informed operational adjustments, optimising use while extending life strategically.

23:48
What’s really striking here is how these real world examples prove traditional methods give snapshots at best, but continuous multi parameter online monitoring provides a dynamic movie of the Transformers health, including its electrical environment, the full connected story. But we need to be realistic. Implementing this in a challenging, unstable grid like South Africa has complexities. The KDs study itself hit snags which actually reinforced the need for this monitoring. The Online Server got overloaded with data. Remember 25,000 events in six months, mostly from large fluctuations in voltage on the HV side, continuous grid instability, not faults. The grid itself creating noise Exactly. They had to adjust alarm thresholds just to cope. Shows the grid as a major stressor, and monitoring needs robust handling and intelligent filtering, plus they noted issues correlating data due to sensors having separate timestamp and setting data unchanged, not fixed intervals, maybe precise correlation, like matching a gas change to a specific electrical event. Harder synchronisation challenges, right underscores the technical complexities in real world unstable grids, but these challenges highlight the necessity of overcoming them for true sync. Actionable insight without it that missing half stays hidden, leaving utilities blind. Okay,

25:06
so we’ve clearly seen the limits of just traditional internal diagnostics and the incredible insights from comprehensive power quality monitoring. The obvious conclusion we need both. They’re not competing, they’re complementary, creating a synergy for a truly predictive, proactive strategy. So what does this combined complete monitoring strategy actually look like? At its core, physical diagnostics are the foundation still essential for internal condition. DGA is fundamental, acetylene, methane, ethylene, etc, but the evolution is online. DGA units like the Kelman transfix, continuous surveillance catches issues in infancy, enables condition based maintenance, avoids missing, rapid changes between real time visibility, exactly temperature sensors, oti, WTI, still vital for thermal performance. Hotspot tracking, direct link to ageing. Bushing monitoring systems, continuous insight via capacitance, tan delta and crucially, phase voltages, massive leap from periodic offline tests, needing outages, monitoring cooling systems, fan currents, pump operation ensures thermal management works critical for life and loading, especially in hot climates and on load, tap changer, OLTC monitoring, looking at tap positions, operations, differential temps, helps predict mechanical issues, contact, wear, major failure contributors, these give the what is happening inside

26:18
exactly, and connecting This to the bigger picture. It’s not just adding sensors. It’s synthesising diverse data streams into a holistic, actionable, intelligent view of asset health, a dynamic, digital twin. This is where power quality and waveform diagnostics become the missing half unveiled, the critical layer, revealing the external, electrical environments, relentless, often invisible, impact systems, capturing high res waveforms like those, retaining 12 cycle RMS trends, full waveforms mentioned by ncpcsa provide the granular detail needed. This includes continuously monitoring harmonics, THD, individual harmonics causing extra heating stress, eroding insulation. Life silently voltage and current unbalance, persistent unbalance, causing uneven temps, losses, localised stress, premature failure points, voltage dips, swells, transients, capturing these rapid events that physically stress components, degrade insulation, excite resonance, load factor and profiles, not just peak load, but the dynamic nature, rapid changes interaction with ambient temp, affecting hot spots and ageing, complete thermal stress, Picture phase voltage stability fluctuations, indicating grid instability, affecting assets like the KDs server overload from HV instability context of the environment the so what of this waveform data is huge. It provides the indispensable root cause analysis. It helps answer why DGA gases are appearing, why temps are high despite normal loads, why a component ages prematurely? Lets you differentiate between an internal fault, seen in DDA and external grid stress, accelerating deterioration seen in PQ, transforms reactive maintenance to proactive asset management, intervention before symptom becomes catastrophe.

27:54
So what does this all mean for you, the utility or asset manager, when you combine these two halves, internal, physical diagnostics and external electrical environment diagnostics, the outcome is, well, it’s game changing for managing critical assets. First a truly predictive and defendable transformer strategy. The sase report explicitly says online monitoring announces change in advance of failure. Effectively reduces risk of unexpected catastrophic failures, shifts maintenance from reactive or time based to truly condition based. You operate with intelligence foresight leads directly to optimised maintenance instead of costly periodic offline tests, needing outages like those six year bushing tests schedule maintenance precisely when needed based on real time data, minimises downtime, cuts costs, improves efficiency. The Sales Report projected big savings for ES com, R, 11 point 4 million over 10 years. R1, point 1 million per year, just from online monitoring versus existing strategies,

28:49
powerful ROI, significant financial benefit,

28:51
definitely plus by understanding and mitigating stressors, both internal and external, you gain the ability to extend asset life, maximise design life, maybe push beyond 35 years, defers costly capital replacements, strategic use of funds. The impact on enhanced reliability and availability is immediate, proactive intervention, drastically cuts unplanned outages, improves reliability metrics like sai fi and Saad, crucial for performance and compliance. Fewer interruptions mean higher availability, better customer satisfaction and trust and trust. Finally, with precise, continuous knowledge of health and remaining life. Managers make data driven capital planning decisions. Invest where impact is greatest, target vulnerable assets. First, not just age based cycles, smart strategic resource allocation for a resilient grid. Here’s where it all comes together. We’ve seen that the standard approach, relying heavily on internal diagnostics like oil analysis, just gives us half the story, the other half the external, electrical environment. The grid itself, is a constant, dynamic, often aggressive stressor, if you’re not watching it with high resolution power quality monitoring tools capturing subtle wave form shifts your effect. Flying blind to a huge invisible chunk of your assets, degradation, missing the forces, shortening the life of critical infrastructure. This deep dive should leave you really thinking about those hidden forces. Are you truly prepared for what the grid throws at your assets every second? What stands out most to you about combining these two worlds, internal, physical, external, electrical. For that complete picture. The tools exist. The data is available. The benefits are clear. It’s not just about symptoms anymore. It’s understanding the entire ecosystem of transformer health, moving from reactive fire fighting to proactive foresight.

30:32
This really raises that final important question. Most transformer degradation, especially from continuous grid stressors, is fundamentally invisible to traditional tools until it’s too late, often irreversible. The grid might be unstable, harmonics, unbalances, transients hitting your assets constantly. But your asset strategy doesn’t have to be blind to these powerful forces. Consider what proactive steps you can take now to complete your monitoring picture. What crucial insights are you missing that could prevent the next unexpected failure and truly transform your asset management from reactive to predictive. The power, quite literally, lies in seeing the whole picture.