Transcript
Introduction to Rapid Voltage Changes
Good afternoon, everyone, and welcome to today’s white paper webinar. Today, we’re talking about rapid voltage changes, or RVCs. A rapid voltage change is more of a nuisance power quality issue than an equipment damage issue. A rapid voltage change is, as the name implies, a kind of a step change in RMS voltage that is within a certain timeframe.
If the voltage transitions very gradually from one level to another, no one’s really gonna notice that. But if it makes a very quick step, that often manifests itself as visible light flicker, and so RVCs are really strongly associated with light flicker.
IEEE 1159 Definition of RVCs
An RVC is defined by IEEE 1159 typically as a steady state change from one level to another in a voltage that is within a certain time period, a few cycles, but less than a sag. RVCs technically could be increases in voltage, but they’re almost always drops in voltage, voltage dips that are too small to be considered a true sag.
Per 1159, a 10% or 90% retained voltage is usually considered the cutoff between an RVC and a voltage sag. If you have a drop in voltage that is very quick that goes more than 10% of the nominal, that’s considered a voltage sag, and a sag is in the region of where you might cause equipment misoperation or something else.
If it’s smaller than that, say in the two, three, 4% range, that’s considered a rapid voltage change, and 3% is a very common rapid voltage change threshold. If it changes less than 3%, that’s not considered really anything. That’s just normal variation. But if it’s 3% or higher all the way up to that 10% drop, that’s in the range of a rapid voltage change. And if the sag was even deeper, that’s a true IEEE voltage sag.
RVCs and Flicker
3% is where most utilities define the border of what is a true RVC versus just an imperceptible change in voltage. And again, rapid voltage changes manifest themselves as flicker if they happen often enough. A single isolated 3 or 4% voltage change usually isn’t going to generate noticeable flicker. So it’s a combination of the depth of the voltage change and the frequency, how often they happen.
You can go back to the traditional GE flicker curves, which defines so many dips per minute or per second or per hour based on magnitude, and consider whether that rapid voltage change pattern is gonna cause flicker or not. Or you can use the more modern ST and PLT flicker metrics from 1453, but rapid voltage changes still contribute to flicker regardless of the measurement method.
Spotting RVCs in PQ Canvass
Rapid voltage changes can be a little bit difficult to spot in a waveform capture because they’re so small. Visually, it’s hard to see a rapid voltage change when you’re looking at sine waves. Here in the white paper, we have the waveform capture in PQ Canvass, and I’m gonna switch to PQ Canvass just to kind of show this.
Here we have an example where we have a file, and we can look at the raw data with PQ Canvass, and we can start with the voltage or current strip chart. You’re not gonna see rapid voltage changes very easily here because they’re so small. You really need to be looking at waveform capture to see this.
These waveforms have already been classified by Merlin, our AI system, and you can see the classification here on the right. This is cheating in some sense because it’s telling you where the rapid voltage changes are: RVC, RVC, RVC. But if you didn’t have Merlin, the AI analysis on this, you wouldn’t have this column already filled out.
Using the RMS View
If you’re just looking through these by hand, here’s the waveform capture. It’s hard to tell that the voltage here in red has dropped a few percent from one cycle to the next. When you’re looking at instantaneous sine waves and you’re graphing from -200 volts to +200 volts on a 120-volt RMS basis, a few volts of change, even 10 volts of change, is just a pixel or two tall. It’s really hard to see by eye that the RMS voltage is changing down here.
To see this, you wanna click RMS, and now we have a continuous RMS graph where PQ Canvass has taken the raw sine wave data and computed a point-by-point RMS value. That RMS window slides across the dataset, and that’s what we graphed with the RMS view of this.
Here it’s much clearer that we have a rapid voltage change. We’re going from about 121 volts down to 120.8 volts. This is just a couple-of-volt change, but that’s enough to be classified as a rapid voltage change. What you see here is a drop on two legs, actual increase on phase B here. But again, this is almost invisible if you’re looking at sine waves, and not really easy to spot if you’re looking at waveforms. You have to cycle through these by hand.
Merlin AI Classification
As the paper shows, Merlin is a much easier way to do this. We could click on the Merlin output, and here is a much easier way to get a handle on rapid voltage change. Here’s the waveform report, and we can look at any one of these waveforms that have already been classified.
Here is the analysis of that waveform, of why it’s a rapid voltage change per IEEE 1159. It’s also mentioning that the THD is elevated, but we see that the voltage drops about eight to 9% from the baseline, so that’s not enough of a drop to be considered a sag. It would have to drop at least 10% to be considered a true voltage sag.
So this is not a voltage sag, but it is a rapid voltage change. It’s beyond the 3% threshold that Merlin was looking for, so this is in that kind of Goldilocks sweet spot of a small dip, but not too serious. It’s not likely to cause equipment misoperation, but if this happens often enough, it’s likely to produce flicker.
THD and Harmonic Issues
Merlin is also pointing out a bigger issue here: your THD is bad. If we click on the waveform, there’s a flat topping here and the voltage is indicative typically of third harmonic. We can click on the Harmonics here, and indeed there is a very large amount of third harmonic. That’s pretty common with single-phase harmonic loads, and the flat top is characteristic of that third harmonic. So we have a couple of different problems in this waveform, and both of them were flagged by Merlin.
Waveform Report and Patterns
You can click on the Waveform Report and it will then summarize what happened. The patterns are rapid voltage changes, rapid voltage change motif here. You have this pattern throughout this, which tells you that it is gonna become a flicker issue. These are in the 3% to 5% range at raw. They’re associated with customer current, and so this is telling you that attribution. There was this customer-caused or load-caused rapid voltage change.
Rapid voltage changes, again, are small changes in voltage that are difficult to spot by eye, but fairly straightforward for Merlin or other types of systems to flag. They’re not gonna cause equipment misoperation, typically. Rather, they’re gonna cause irritating levels of flicker if they happen often enough.
Quantifying Flicker from RVCs
A single isolated rapid voltage change, or if they’re small or infrequent, still aren’t gonna contribute to flicker. It’s only if they are happening fast enough. If you need to quantify that, the PSD strip chart or the IEEE 1453 analysis is the best way to analyze that.
We don’t have PSD turned on in this recording. We can always fall back on the old GE flicker curve, and we did have some exceedances here. We did have enough rapid voltage change, and we saw some of those were 7% or 8%. That’s almost a sag, and at that level, you can’t really tolerate too many of them per hour before you start to exceed the GE flicker curve. So it’s not a real surprise in this recording that we are exceeding that. Here are the exceedances per IEEE 141, the GE flicker curve.
Mitigation Strategies
As we described in the paper, ways to address rapid voltage changes are usually on the customer side. Generally, the transformer’s not heavily loaded from these rapid voltage changes, so if the transformer’s already sized correctly for thermal reasons, you usually would not upsize the transformer just for rapid voltage changes.
Usually, the customer side is where these mitigation things happen. This would be like:
- Soft starts on motors
- Converting a motor to a VFD
- Making sure that the soft start or VFD hasn’t been bypassed
That’s also a common issue where a motor used to have a soft start, but it stopped working so they bypass it, or it’s not getting set correctly. But generally, customer-caused rapid voltage changes are on the customer to address, assuming the transformer’s sized correctly for loading or thermal reasons.
Distribution-Side Solutions
If there’s a persistent problem on a feeder, and if you have a customer that is disrupting things for the entire feeder or it’s coming from upstream, there are more advanced methods on the distribution side to address these. A steady bar compensator, STATCOMs, distribution-style STATCOMs, or even active power quality mitigation devices that attach directly to the primary that can put energy into the system during the RVC itself. These are able to compensate for that in real time. Those are fairly expensive, but can be a solution for intractable problems that are kind of permeating an entire feeder.
In most cases, the solution for rapid voltage changes, if it’s contributing to flicker, is on the customer side. Or if customers are sharing a transformer, a shared secondary, then the customers causing the RVCs need to mitigate that so they’re not disrupting voltage for everyone else in that secondary. A solution there also would be to break up a transformer into multiple transformers so that the customer causing the RVCs was isolated from the others.
Summary
That really covers the topic here in the paper. Many PQ engineers aren’t that familiar with the RVC term. You can think of it as a really small voltage sag. It’s essentially a sag that is too small to be considered a true sag in that it’s not likely to cause equipment misoperation. Rather, it’s at most going to be a nuisance with light flicker.
If you have questions about the whitepaper, feel free to type them here in the questions box. Or if you have questions later, give us a call anytime, 1-800-296-4120.
Q&A: RVC Report in ProVision
Good question, “Is there an RVC report in ProVision?” There actually is not. The closest you would have is really the flicker report, the GE flicker report, which effectively is classifying RVCs in the sense of did they exceed the GE flicker curve. If you did not exceed the curve, it doesn’t log it at all. But if you had enough RVCs at a certain level, it will log that.
The ProVision report looks very similar to the one here in PQ Canvass where we have a tolerance. This is a percent change, so this would be considered like a 1.2% RVC or a 0.9%, and so forth, and you hear the tolerances under allowed limits all the way up to 7%, and how often do you think you can have these happen in a certain time period. Kinda indirectly, this is the closest to a true RVC report in ProVision, where you get RVCs logged basically if they were frequent enough or deep enough to exceed the GE flicker curve.
If you want the Merlin classification, that has to be done in PQ Canvass. You can transfer a file from ProVision to PQ Canvass, if you have an account, and run the Merlin AI system, and have it then classify all the waveforms and sort which ones are RVCs and which ones are other types of disturbances, and get that RVC report. Merlin will say if there are any patterns to the RVCs or the customer versus utility contribution to those.
One last thing about Merlin™: if you wanna try Merlin™ out, go to pqcanvas.powermonitors.com/demo, and you can try it out free if you wanna see Merlin already run on some example files.
Well, that’s all the questions we have. Thanks for attending, everyone, and everyone have a great afternoon.
