Transcript
What Is an Impact Load?
Good afternoon, everyone, and welcome to today’s white paper webinar. Today we’re gonna talk about impact load. An impact load is a load on a transformer that exceeds its nameplate rating, a repetitive load.
An impact load is a type of load on a transformer where it sees a repetitive load current that is higher than its nameplate rating. This looks like a fault or a short circuit on the transformer winding, and that causes a lot of mechanical stress on the core and the winding itself.
So the damage done by an impact load to a transformer isn’t thermal, it’s mechanical. Those repeated inrush currents or very high currents apply mechanical stress to the magnetic forces involved that cause degradation and mechanical damage to the transformer itself, resulting in early transformer failure. So again, this is not a thermal issue, it’s a mechanical issue on the transformer.
Where Impact Loading Is Most Commonly Seen
This is most commonly seen with what are known as unit transformers, transformers that are sized for one to a specific load, typically a motor. This isn’t as commonly seen in the utility world because most utility transformers feed an entire building where you have a diverse set of loads. No one load in the building is typically sized that’s comparable to the transformer itself. But we’ll show you some exceptions where that’s not the case.
Traditionally, this is more common with a load in a transformer inside a building where a transformer’s sized, say, just for the steady state running current of a particular motor, and in which case the motor’s inrush current is over the nameplate rating of the transformer. The bottom line is if a transformer is subjected to impact loads, it may need to be derated.
The Derating Curve
The standard way of doing this is with the derating curve that was developed back in the 1970s from an important paper written by Frank McCann and Robert Aristo. This is still the definitive reference on how to derate a transformer if it’s seeing an impact load.
The way to do this is to compute the pulse swing that the transformer sees. That pulse swing is a ratio of the kVA of the pulse divided by the kVA of the transformer itself. The kVA of the pulse is simply the current times the voltage during that inrush pulse, and then the bottom number is the kVA rating of the transformer. So if this is over one, that means that the pulse current is higher than the transformer’s steady state rated current.
In this type of approximation, you can assume roughly that one horsepower is about one kVA. If the pulse swing is over 1.0, that does mean you’ve exceeded this nameplate rating from a steady state basis. And if it’s under one, that means the transformer should be able to take that indefinitely.
We have a graph here where on the x-axis we have the pulse swing, and the vertical axis is how many times the transformer sees this per hour. If you’re under one in this range, you should be able to do this an infinite number of times. There’s no limit to the number of times the transformer can see a current that’s under its nameplate rating.
Where we get into limits is when we’re exceeding that. For example, if we have an inrush current that’s three times the nameplate rating of the transformer, we can do that about six times per hour, every 10 minutes, before we exceed the safe operating area of the transformer. If we are in this region, that transformer is likely to fail early. If we’re in the region under the curve, that should be okay.
When you exceed four, that’s pretty much a hard cutoff. You should not do that at all on a transformer. If it’s repetitively seeing inrush currents that are more than four times its nameplate rating, that could cause damage quickly and you should not subject that to a transformer.
Example: Poultry Plant
Here is a case from the white paper. This is a poultry plant that was fed by a one MVA transformer, and this transformer died after six months of service. The peak thermal demand is about 1.1 MVA when you just look at the 15-minute average current. So this transformer was sized pretty well, roughly the same as the thermal demand. This transformer didn’t fail from steady state current.
Here we have the graph of the current, and we see the characteristics of an impact load. The min, max, and average current are drawn, and the min and average current are very small relative to the max. This purple is the max current. This is a two-minute strip chart interval, so every two minutes we get that one cycle minimum, one cycle maximum, and the two-minute average. You can see that the purple trace is in some areas consistently high, over 1,500 amps above the average, and this is happening roughly every two minutes. The blue trace governs the amount of thermal temperature rise on the transformer, but the purple trace, the max current, is related to the impact load.
Ruling Out Harmonics
One of the first suspicions was maybe this transformer was overheated because of harmonics. Maybe this poultry plant uses a lot of VFDs and that has a lot of harmonic distortion, so that 1.1 MVA load may be causing much higher than normal heating because of harmonics.
But if we look at the waveforms from this PQ recording, that’s not the case. The current distortion is very low, under 5%. The waveforms for current—voltage on the top, current on the bottom—are nice clean sine waves. These are not VFDs. There’s no real harmonic distortion to speak of at this location.
But what we do see are some hard start motors. The current just abruptly goes from one value to another value. These are hard start motor starts. Another characteristic of an impact load is high amounts of flicker. We can see that we’re going above the PSD of one line. Because of those large current fluctuations, they generally cause large voltage fluctuations. So another clue to a possible impact load is a high amount of flicker at that location.
Computing the Pulse Swing
Given that we have the maximum current and the minimum current and the average, we can compute the kVA of the inrush pulse. Here we’re looking at a total kVA or channel one, two, and three kVA, and we can see they are peaking at about 840 kVA per phase. We add those together, that’s two and a half MVA of the inrush pulse with a one MVA transformer.
So that gives us a pulse swing of 2.5 over one, which is two and a half, 30 times per hour or once every two minutes. If we look at the derating curve, two and a half is happening about 30 times per hour. That puts us right here for that transformer. It’s outside the safe operating area, and it shouldn’t be surprising that that transformer failed. We were exceeding the impact load that’s allowable with that transformer.
The utility replaced this with a one and a half MVA transformer, which brought us inside the curve. At one and a half MVA, we’re still pulsing at 30 times per hour, but the inrush pulse is now two and a half over 1.5 or 1.67. So that is within the safe operating area of this transformer.
Why the Poultry Plant Had Impact Loading
You may wonder how this poultry plant was generating these inrush pulses in the first place. The answer is that this place was using a lot of synchronized motors. It had a conveyor system and the chickens in the poultry plant were moving down the conveyor system and this conveyor stopped and started repeatedly. It was fed by a lot of hard start motors that were starting and stopping at exactly the same time. So the large number of smaller motors looked like one large motor with one large inrush pulse.
Here the assumption that a transformer that feeds an entire building has a lot of load diversity wasn’t true because most of the loads in this building were synchronized together. So it acted as one large load. This is an exception to the rule.
The rule of thumb is where the service transformer sees a lot of load diversity and doesn’t normally see an impact load. That’s why it can be surprising when you do see it, because utilities are often not looking for impact loads because it’s not as common at the service entrance. This can also be the case when you have a special load, like an irrigation pump or a utility service that’s only feeding one or two motors, a pump or some other type of load where you need to consider the impact load in addition to the actual steady state thermal demand.
Example: Shipyard
We have another example here. This is a shipyard. It’s a primary metered customer, and they have 12 giant cranes that move those huge cargo containers off and on ships from the dock to the ship. The only load for this customer is one of these 12 cranes.
Every time the crane starts, that generates a 150 to 200 amp inrush pulse at 34 kV. So that’s a seven MVA inrush pulse, and there are 12 of them. Each one starts about every 90 seconds. That means every seven seconds, one of them is starting up. So that’s a seven MVA inrush pulse every seven seconds.
You can see the characteristic in this recording where we see a large spread between the max current, the average, and the min. Here, the average is really just a mathematical average between no cranes moving and more than one crane moving. It depends on how many cranes are moving at any one time to determine the average. This is definitely another location where the impact loading of the transformer is just as important as the thermal load demand on this transformer.
Using Tools to Identify Impact Loading
Back to the paper and using the tools to determine this. The first clue is looking at the min, max, and average current. You want to watch out for that spread between the max current and the average current. That’s really the hook that you can use to generate a report to pull this out of the data for you.
Whenever you see that max current being consistently high, that’s the warning sign. The report is geared to pull that out and give you some insight as to how often that’s happening. That’s the current out of limits report.
We can see the examples here in the white paper. We’re running the out of limits report. We can put in the thresholds. You’d run this on the min and max values, and the upper limit for the max value is the one that you’d want to put in. That causes the report to generate a log every time that max current’s exceeded, and you can easily see from the times how often this is happening.
ProVision and PQ Canvass
You can do this in ProVision or PQ Canvass. In PQ Canvass, our web-based system, it’s simply a matter of going to the recording and picking the out of limits current report and entering these parameters.
On PQ Canvass is an analytics report, an impact loading analytics report that would actually go through the data and analyze this data. If you can give it the MVA of the transformer, the transformer size, it will show you where you fall on that chart. You can use the out of limits report as shown in the paper to take a look at this on a manual basis.
Key Takeaways
The bottom line is, especially if you have a transformer fail unexpectedly, suspect harmonics, suspect for a resonance, but also suspect impact loading. If you have a recording where you have a large spread between the max current and the average current, that’s the warning sign for impact loading.
You need to be looking at the data with a PQ recorder or some other recorder that can give you at least one cycle resolution, because these are typically very fast inrush pulses. These aren’t steady state values we’re looking at. These are cycle or just a few cycle long inrushes. So you do need a PQ recorder or some other device that can give you that fine time resolution.
If anyone has any questions about impact loading, give us a call anytime at 1-800-296-4120 or send an email to support@powermonitors.com. Everyone have a great afternoon.