Transcript
Introduction to Voltage Notching
Good afternoon, everyone, and welcome to today’s white paper webinar. Today we’re talking about voltage notching. This is an interesting power quality issue that comes up more often for utility customers rather than utilities.
A voltage notch is a periodic, almost like a very tiny voltage sag in the waveform itself. You can see here in figure two an example of a voltage notch where in the voltage waveform, you have a very small change in the voltage at certain points in the waveform itself. So we have that notch, that notch, that notch. These are all effectively really, really short voltage sags that last just tens or hundreds of microseconds at a time, and it’s repetitive throughout each cycle.
What Causes Voltage Notching
This is typically caused by what are known as commutation problems, where a customer has a brushed motor or some sort of rectification system that’s rectifying the AC voltage, and it’s supposed to hand off the conduction path from one set of diodes to another at certain points in the sixty-hertz cycle. Sometimes there’s a very brief amount of overlap when you have multiple conduction paths at the same time, and that causes a surge in current during that commutation overlap.
On a brushed motor, this is where you have multiple brushes and multiple motor windings energized in parallel very momentarily. That’s not supposed to happen. It should be a break before make connection, where one set of windings is energized and then de-energized before the next set of windings happens. And it’s not always the case that that happens.
So during those brief conduction periods, you have an abnormally high amount of current being drawn, and that causes what’s effectively a very tiny voltage sag on a repetitive basis.
Notching Is Generally Customer-Caused
This is generally not a power quality issue that’s generated by a utility. Utilities don’t have these kind of loads—the utilities are supplying the power, not drawing the power. So when you have voltage notching, the cause is generally from the customer, not from the utility.
Another property of voltage notching is that the notches are generally very short in duration, and they don’t propagate very well through the distribution network. So if you have a customer that has voltage notching issues, the root cause is almost always within that facility itself and is likely not making it to the primary affecting other customers.
That’s not always the case, but generally, if it’s a notching power quality problem, the cause is going to be on the customer side of the meter, and it’s likely not affecting other customers unless they’re on that same transformer secondary. I haven’t seen that, but it’s possible to have, say, a residential situation where you have multiple customers on one transformer, and one can affect the voltage for the other.
Harmonic Effects of Notching
Here in figure two, we have an example of notching. This is repetitive throughout the cycle. Every cycle has this, and this will cause steady state harmonic issues on voltage. So here we have very high voltage on fifth harmonic, seventh harmonic. You can see even the seventh harmonic is higher than the fifth.
This is gonna be a distortion problem for other customers. That high voltage distortion at the seventh harmonic is going to cause trouble for other customers. It’s gonna increase transformer heating. And so that obviously is a problem if others are being affected by this.
IEEE 519 Notching Limits Are in the Time Domain
Notching is interesting in that even though it’s a steady state voltage distortion and has a harmonic breakdown, the limits for notching as defined by five nineteen aren’t in the frequency domain. IEEE 519 does have a section on notching, and the limits are again in the time domain, not the frequency domain.
If you look at the overall spectrum of a notch, for example, here we have a specific notch with a notch depth and a notch width. You can see that there are many, many harmonics that are high. We’re going up to the 20th harmonic here, and you can see there’s a lot of harmonics that are high. If the notch depth or width changes even slightly, this pattern drastically changes. So instead of a null here at the 40-some harmonic, this shape might be very different. So harmonics don’t really simplify the analysis when we’re talking about notches.
A customer may try to filter certain harmonics to try to get rid of the notch. In general, that filtering isn’t gonna be consistent because the notch depth and width is not going to be necessarily the same.
Notch Area and Depth Limits
Consequently, the notch limits are in the time domain. IEEE 519 specifies a limit on the depth of the notch and the notch area. This area is depth X the duration, and duration in microseconds. The limits in 519 for the notch area are in volt microseconds, assuming a 40-volt base. You can scale that proportionally if it’s a 120- volt or a 208 type system.
This volt area is related to the amount of voltage that’s missing in the waveform if it’s charging up capacitors. If you have this feeding into an AC/DC rectifier, the amount of missing volt microseconds, the amount of area determines how much charge that capacitor is not getting at every cycle. The bigger that is, to a point, the worse it is in terms of load performance. So that’s why this is in an area. And then the absolute voltage depth is also a limit. So this is not a frequency domain type of limit and not based on the harmonic pattern. It’s based on the time domain.
Mitigation Strategies
Utilities would generally not be trying to mitigate this themselves. The root cause of these notches is almost always going to be on the customer side. On the customer side, really the fix is to fix the commutation error in the first place. Why do they have multiple conduction paths? Is that a problem with an AC/DC converter? Is that a problem with a particular motor? The best way to mitigate this is by preventing the current spikes that are caused in the notching in the first place.
If that’s unavoidable, then if this needs to be filtered on the customer side, rather than having a tuned filter that’s trying to address certain harmonics, in general, you would need a low-pass filter, like a series reactor or something that would filter out any sort of notch just as a function of a low-pass filtering process. That’s really the best approach on the customer side if they literally can’t address this on the equipment side itself.
Excitation Current Example
One exception to what I just said was the example that we see here in the white paper where we’re looking at what is actually excitation current. We can go into PQ Canvass to look at this in more detail. This is an example where we have voltage in red, current in blue, and we see these notching, not from conduction current or commutation problems, but from a transformer excitation.
This is a special type of load, and we see voltage notches here at the step changes in current. On a three-phase basis, we can see how this is happening. When I switch to a delta mode, we can see that we have voltage notches at each of those current transition points. So this is a situation where it’s not due to commutation problems. You would need line reactors if this is a problem for the customer. That’s a nice example here in the white paper.
Summary
The bottom line is voltage notching is a steady state distortion problem that is kind of unique in that even though it’s periodic and repetitive, you don’t use harmonics to measure the limits for this. It’s done in the time domain per 519, and utilities generally aren’t gonna see this on a transformer secondary unless a customer is causing this for themselves. This is more often seen inside a customer facility.
A customer may be complaining about voltage quality issues, and they’d be making measurements at the terminals of a piece of equipment and showing the utility an ugly waveform like this. But it may not look like this even at the meter base because it’s so high in frequency. But the utility still needs to investigate it. And if they do see a notch like this, more often than not, it’s caused by the customer themselves.
If they’re violating the 519 standard, then they would have to do the mitigation on their side, ideally by fixing the equipment that’s producing this notch. Or if that’s not possible, then a series reactor that’s effectively a low-pass filter or even a one-to-one transformer, isolation transformer that would operate as a low-pass filter, filtering some of these high-frequency components from making it to other pieces of equipment.
The closer to the offending source of equipment they can do that, the lower cost that’s gonna be. Rather than filtering their entire building, if they can find that offending load, filtering right there will be the most cost effective. Thanks for attending, everyone. Have a great day.
