Transcript
Introduction to Ferroresonance
Good afternoon, everyone, and welcome today to our white paper webinar. Today, we’re here with Caleb, and we’re gonna talk about ferroresonance. This is a very interesting power quality topic. It doesn’t happen often, but when it does, it can be very damaging and also very hard to track down.
It’s also something that traditionally has been blamed on a lot of PQ issues because when you don’t know what it was and you’ve ruled out everything else, this is an easy one because it’s often hard to prove. You’re left with equipment damage and no real signs because it often happens abruptly, and unless you had a PQ recording going at the time, which you typically weren’t, you don’t have any evidence whatsoever. And so you’re just left with a damaged transformer or equipment and really no signs. It’s one of those mysterious ones that are hard to reproduce and tough to prove.
How Ferroresonance Works
Now, what happens in ferroresonance is that we have a non-linearity between the transformer inductance and the external capacitance. If you have a capacitance and inductance, there is typically a resonant frequency based on that LC network. But when a transformer’s lightly loaded or unloaded, that inductance is now variable. It’s a non-linear inductor, and this inductance really depends on the point and waveform.
You can end up, if conditions are right, with a positive feedback loop where you form an oscillator, where that inductance is changing versus time, and you get a positive feedback and now you’ve made an oscillator that’s powered by your own utility power. The transformer output is no longer a 60 hertz sine wave based on turns ratios. It’s an oscillator where it could be a square wave-looking thing, it could be all sorts of interesting shapes driven by that non-linearity between the inductance of the transformer core and that external capacitance.
That non-linearity in the inductance happens when you’re in that hysteresis curve, it’s lightly loaded, the magnetization current is most of the current in the transformer, or all of it, and when you have a single-phase event, typically.
Classic Ferroresonance Scenarios
The classic situation is you have a three-phase transformer that has no load or lightly loaded, and there’s a single-phase event, a single-phase interruption, a single-phase sag, some sort of disturbance that opens a phase on the transformer and it goes into ferroresonance. This can also happen when you’re energizing a transformer, if you’re doing it one phase at a time, if it’s a new installation and there’s no load on it.
It can also happen during solar witness testing where you’re intentionally dropping a phase to test the inverter to make sure it disconnects. Those are all classic cases of ferroresonance. Some utilities have guidelines, if it’s an underground cable feeding a large transformer, that underground cable capacitance is often a source of serious capacitance. And so they’ll be careful energizing a three-phase transformer that’s fed by an underground cable. You do it on all three phases together, or you put a load on it so that it’s not just purely unloaded.
Recognizing Ferroresonance in PQ Canvass
The example that we’re gonna be showing today here in PQ Canvass is not actually from a solar step-up transformer, but it is a really good way to see the characteristic results of what ferroresonance looks like if you happened to capture it in a PQ recording.
If you look at your standard recording, any portion of the recording, you can see there’s low imbalance. Everything looks fine, everything was healthy, floating around the nominal. And then this is an induced test where the utility in this case was actually trying to induce ferroresonance. You can see here, as the ferroresonance ramps up, you see this huge disparity between the max and the min voltages showing up here.
When you look at the triggered waveform capture, we can look and see what the individual waveform catch in the cell looked like during these. And you’ll notice that this is not even close to sinusoidal. Channel three here is nice and clean, looks great, but channel one is all over the place here. That’s not a 60 hertz sine wave anymore. You never wanna see that kind of waveform for voltage. You’ve made an oscillator, energy is going back and forth between inductance and capacitance as a positive feedback loop.
Voltage Readings During Ferroresonance
If you look here too, you can also see that our peak instantaneous voltage is still under 500 volts, whereas the scale here has changed up to almost four kilovolt. We’re not reaching a kilovolt, we’re probably in the 600, 700-volt range, but you can see here on the RMS, instead of the 290-ish volts, we’re running around 375, 380 or so.
In channel three there, it looks kind of funky with that sinusoidal pattern on the RMS value, because the recorder is phase-locking to channel one voltage. And it’s assuming it’s a 60 hertz sine wave, but it’s not. So it’s trying to follow the frequency of that weird waveform. And so that throws off the frequency measurements and RMS values for channel three, which is six years.
Switching over to the frequency component here, you can see that there are massive components, a 120-volt component in the second harmonic. These are not typical of your typical sine wave. So even still, that’s why we see this wild variation that you’re seeing. And we’ve got a series of those that are pretty characteristic throughout the remainder of this recording. That’s a really common characteristic.
Peak Voltage and Damage
The peak voltage that is achieved is far, far higher than it should be. And if this happens in the field, that’s usually gonna cause some sort of fault or a failure, and you’ll often have a protective device that then trips and once the transformer disconnected, it goes out of ferroresonance. And so you lose any clues that that’s what the root cause was.
Unless you happen to have a recording at the time, one of the few clues you might have is some voltage somewhere that is higher than normal. If you have an AMI meter or a relay or something like that that is logging voltage, its last voltage reading, if it’s only logging once a minute or once every 15 minutes, might be higher than normal on one or two phases and that might be your only clue, and it’s an RMS voltage likely, most likely. It’s unlikely to have a waveform.
But that might be the only clue you have, that the voltage went up to 150 volts instead of 120 on a 120 volt basis. That implies that probably the waveform is doing something crazy and that was perhaps ferroresonance. So those are some scraps of information you might have, is a very high peak voltage or RMS voltage from some reading nearby, when that happened.
Voltage Spread as an Indicator
As Chris had mentioned, that high peak voltage you can see here as we go from just outside the transition into the ferroresonance state, you can look at the average voltage is 284.9 volts. It’s the same with the min and the max. The spread there is essentially the same. When you come over here into the ferroresonance state, you’re looking at an average voltage goes from 284 to 340 with a peak voltage of 413 and a half volts. So as Chris mentioned, those peak voltages are pretty good dead giveaway there on the RMS.
Another red flag here is the spread between the min and max voltage. You’ve got 100 volt spread or a 50 volt spread. On a 277 volt, you should never see many, many tens of percent spread between the min and max voltage. And then towards the end it gets far worse. If you see a spread between the min and max voltage that’s more than plus or minus 10% or especially 20%, that’s a huge red flag right there and that often indicates ferroresonance. That voltage should never be bouncing around like that and if it is, it’s likely ferroresonance.
Here you see over 270 volts. That’s obviously horrific. But even before that, that’s way out of the ordinary. You should never see something bouncing between 275 and 380, that’s 100 volt spread on a 277 volt here, that’s clearly abnormal. That’s not caused by load switching or that kind of stuff.
Potential Transformers and Ferroresonance
Another scenario where you might see ferroresonance is in PTs, potential transformers. PTs by definition are lightly loaded. Their only load might be a piece of instrumentation at, say, the substation. And this has gotten worse with time because as PTs are powering electronic equipment, that equipment draws less and less power than it used to be. So you might have, like, a 1 KVA little transformer that’s stepping down primary to 120 volts and its only load is a one or two watt meter, much less power than they used to be, than maybe, say, an electromechanical relay.
Especially if those three phase PTs are ungrounded, more systems are grounded than they used to be, or if it’s an open delta where you only have two PTs, that can also be something more prone to ferroresonance. So watch out for PTs, especially in distributed generation systems where the PTs are failing.
CVTs, capacitive voltage transformers, are also prone to this because they have capacitors built in along with an inductive transformer. So those sort of things are also areas where you should be on the lookout for ferroresonance because they’re lightly loaded, because they’re not actually powering anything and they can easily be single phased. A fuse could blow on a PT, that sort of thing, and you might not really notice because maybe your relays aren’t using voltage for control, they’re only using the current, but now with one phase out it’s primed for ferroresonance.
Banked Transformers
You can have ferroresonance even when you have separate transformer banks. A lot of ferroresonance discussion talks about a three phase transformer and the windings are all in a common core, but you can also have this even if you have a banked transformer where you have three individual transformers wired together as a three phase transformer. You have windings on their own cores, those cores are still nonlinear things in the right situation and you can have ferroresonance even if you have three discrete transformers wired in a three phase system. So it doesn’t have to be a single transformer with a single core with windings all on the same core.
Capturing Ferroresonance in Power Quality Data
If you are collecting power quality data, you don’t really have to do much to catch this. Usually it’s going to be caught very readily because the signs are so bad. But again, you have to be recording in the first place to catch this, so be on the lookout for ferroresonance. Again, it’s something that is uncommon, but when it happens, it doesn’t leave a lot of evidence except broken equipment, damaged equipment and open protected devices, unless you have some data available.
Merlin AI Analysis in PQ Canvass
One of the other things that we can show you here today is the PQ Canvass. If you’re not really sure, this is a pretty textbook signature of what ferroresonance looks like. PQ Canvas offers another AI driven analysis that will look at the whole recording end to end. It’ll look at the waveforms, the strip charts and the other individual records like your event change records, significant change records and flicker records. And this is the GE flickered records versus the IFL and PLT flicker and PST flicker standards.
What Merlin, our AI analysis suite, will do is actually look through all the records in your recording and come back with a gradation. We have a series of compliance reports: the NT C84 for voltage regulation, IEE 1668, 1453, NT C84 one again and IEE 519 for harmonics. It’s gonna give you a score of zero to 10 severity score telling you you’ve got really nothing to worry about here — sags, flicker and harmonics, voltage regulation — maybe some issues which is likely going to be that high imbalance at the end of the recording, and then here’s your imbalance right here, which is a 10. So it’s about as bad as it gets.
Waveform Severity Rankings
When we go into the waveform itself, it’s going to rank every single waveform by severity. It’s going to give you a brief one line overview of what’s happening in this waveform. So this is capture number 35, we have a brief interruption with over voltage and heavy harmonic distortion. As we looked earlier, all of our waveform captures started during the ferroresonance period, so it’s pretty ugly, and it’s gonna give you a little bit of information about that.
Now, Merlin™ itself has not been trained to identify ferroresonance in and of itself. We don’t have anything specifically specified for that just yet. But we were able to clue it in by leaving some notes in the header report, and when you start a Merlin™ analysis, you can give it a little bit of information about what’s being captured.
So it says, “The memo indicating that the ferroresonance test on the 1500 kV at pad-mount transformer suggests it’s a deliberate transformer —” It’s actually backing off a little bit on its assessment because it knows that this is a ferroresonance test we’re recording. And so that’s why the severity is probably not any higher than it is. But it does mention the wide disparity, the extremely high total harmonic distortion, because remember, there’s a ratio to the fundamental which it’s having a really hard time even picking up what that fundamental is on these waveform captures.
Each one of these are ranked one by one, each one, and then it gives you a severity index, a severity overview. We have 94% of them are here, and then we have the 6% remaining in the higher end.
Chatting with Merlin
And then, of course, you get a waveform report, which is an overview of all of these combined, so it’s a more holistic report telling you what’s going on. And if you have more questions, you can go and talk to Merlin. You can ask it questions, you can chat with it, and it can give you an overview of what’s gone on.
We had prompted it earlier, “Hey, what’s going on here?” And it’s given us some key points across the whole recording. So Merlin gives you a summary of the whole recording, not just one part — flicker, harmonics, sag, swells, all your compliance waveforms strip charts. You get this large, very detailed report that tells you what’s going on in the recording.
An issue matrix shows voltage imbalance, voltage regulation pretty high on the list, some sags, swells. And most of those were probably triggered based on the fact that the RMS version of those waveform captures were all over the place during the ferroresonant period.
There’s a lot going on in this report, and so you can chat with Merlin and have it give you a much quicker, a much shorter, real quick analysis of what’s happening in the recording. It can give you time of day correlations — from 8:05 to 8:07 AM, voltage one and two swinging in opposite directions. So again, going back to what Chris was talking about earlier, this is a very big indicator of ferroresonance in the system, where they have that huge spread between the min and max volts. Mirrored behaviors, classic of an open, very high impedance neutral. Again, all the criteria, it’s finding those, and you can chat with it here.
Contact Information
If you have any questions about the white paper or about ferroresonance, give us a call anytime at 1-800-296-4120 or send an email, support@powermonitors.com. Thanks for tuning in, everyone. Have a great day.
