Transcript
Introduction to PV Monitoring
Welcome to today’s webinar. Today we’re going to be talking about PV monitoring. Some recommendations and best practices for making PQ recordings, specifically situations.
This is a very common situation for utilities as the rise of DG gets more and more prevalent, as the penetration gets higher and higher. That causes a lot of power quality issues. And starting from the very basics of RMS voltage, voltage regulation. There’s lots of concerns when it comes to DG with waveform distortion, harmonics, flicker, even frequency, and those are all concerns.
But even just basic voltage regulation can be a real challenge when it comes to dealing with DG. Where is the end of line? Does that change day versus night? Do you have a high voltage or overvoltage on your system because of too much generation?
And especially reactive power control, using cap banks for voltage control, is more difficult because you may have real power supplied downstream of the substation, but reactive power is only going to the substation, so that makes regulation even trickier.
Distributed Generation Monitoring Concerns
These are some of the concerns when we’re talking about distributed generation monitoring. Again, voltage regulation is the most basic PQ issue, and it’s often the most common that you’ll have trouble with when it comes to DG.
The behavior of these inverters under SAG conditions or fault conditions is difficult to characterize. That can make protection coordination very difficult. And again, harmonics, high frequency distortion—there’s a lot of concern about that, and that can be a problem. More often than not, the real problems are voltage regulation and light flicker with these inverters as they fluctuate.
Recommended Strip Chart Data Types
Voltage regulation is the most basic, and here we’re looking at correlations between RMS voltage and generation. And to record that, as the paper talks about, you want to record strip chart data, of course, for RMS voltage and current. Always want to record that.
But you also want a minimum real power and reactive power THD for voltage and current frequency. Frequency is not as often a problem with the inverters as the others, but you want to have that enabled. And also, you want to turn on flicker so you can get PST and IFL limits.
A PQ recorder like the Seeker, the Evolution, the Guardian, those can all record these basic PQ types, and you want strip charts for all of these. And that’s important for determining that correlation for looking at graphs like we see here in PQ Canvas.
Analyzing DG Data in PQ Canvass
Here I’ve got PQ Canvass loaded. These are some example DG files, and we’re really looking at the relationship between voltage and current and voltage and current power. We can see this DG site where every day the sun rises. Some days we have good generation, where it’s pretty much cloud-free day and we get the maximum output. And other days it’s cloudy, and that cloudy time really gives us that chance to see before and after.
With a waveform like this, well, if it’s only smooth continuing changes, there’s no real change in voltage during this time. We see the regulator kick the voltage down here as the sun’s rising. But another view is if we have cloud cover, what does that do on a voltage basis during that transition?
So this is really the important time to look, or one of the important times. If there’s fluctuations in generation output, what fluctuations are there in the voltage output? We can see here with the power transitioning very quickly on the order of tens of minutes based on clouds rolling in, rolling back out. With wind turbines, this could be abrupt changes in wind speed.
What does the voltage do on that sort of short timescale? Do you have regulation issues that are on tens of minutes, not tens of hours timescale? And that’s where you want to zoom in on the periods where the inverter is changing its output relatively rapidly, on the order of tens of minutes.
RMS Current Versus Power for Mixed Load Sites
You want to really look at both situations. You can look at the longer timescale. We’re looking really at the peak. What is the voltage at the peak of the generator output, or is there a point where the voltage is higher at the intermediate output? Really the peak voltage is not at peak. This should be generation output, but because our regulator has dropped this down a bit, but right before that.
In the paper we also show some other types of graphs where we’re looking. Here I’m looking at RMS current. And really you can look at current or power if the only device on that service is generation, because all the current is generation current. But if it’s a mixed situation where it’s some load and generation, then you want to look at the total power because you can’t separate that with just RMS current.
Going back to the presentation, if you have a situation like this where you’ve got a few hundred kilowatts of solar on the roof and a few hundred kilowatts of actual load, you’re only seeing the net. And so that’s where you want to look at power, not just RMS current when you’re looking at these correlations.
Recording Real and Reactive Power
So to do these types of analysis, you want to record RMS current, RMS voltage. Real power, reactive power can be helpful, especially if the inverter is in any sort of VAR control mode. 1547 compliant inverters can supply or absorb reactive power in addition to real power.
If you have communications control to the inverter, you may be putting it in certain modes. If you have a DERMS system or ADMS system behind your SCADA system, then you want to monitor that reactive power flow, and you may put other monitors at other points in your system to determine where is that reactive power flowing? Is the inverter doing what it’s supposed to do?
And because these inverters can put out arbitrary amounts of real or reactive power, the relationship between the two is not as simple as it is with normal inductive loads. And so you’re really doing a record both of those independently.
Waveform Capture for High-Frequency Noise
Now, the other important data type for characterizing these systems is waveform capture. This can be helpful to look at the waveform itself, especially if you’re looking for high-frequency noise or resonances.
In this example, we have an inverter system, and we have a large amount of high-frequency noise. And if you look at the RMS strip charts, you don’t really see that, of course, in the RMS data. You might notice there’s a slightly higher spread between the min and max voltage, but it’s really not a noticeable change. If you look at the THD on the other hand, fairly elevated THD. But it’s not severely elevated. It’s still just a few percent.
But if we look at the waveforms, we see something completely different. Here we have a resonance that is activated as soon as the inverter switches on. So we have the inverter off. In a nice situation, we have a nice clean 60 Hertz sine wave. But when the inverter switches on, we have a huge amount of high-frequency noise, and this is on the primary side of this circuit.
Cable Resonance at Inverter PWM Frequency
In this situation, there was a long, long underground cable going to this inverter that had a resonance. The capacitance of that cable resonated with the rest of the system at 3,600 hertz, which happens to be the inverter PWM frequency. So what would normally be a very small amount of inverter noise was amplified into a very large voltage distortion at that frequency because of that resonance that happened to be at the inverter frequency.
We can verify that if we look at the harmonic content for this waveform. I’ll turn on the harmonic graph, move this rectangle over to the noisy section. And to really see this, we need to turn off the fundamental and also turn on the harmonics past the 50th. We need to go up to the 100th harmonic to see this.
And if we do that, we can see that we have significant energy at the 59th harmonic and the 61st harmonic. These are side bands of 3,600 Hertz. You’ve got 20 volts at the 59th harmonic. In reality, this really isn’t synchronous distortion. We’re talking about harmonics here, but in reality, this is non-synchronous distortion that’s at the inverter frequency, PWM frequency. It’s not really driven by 60 Hertz. These are super harmonics, and there’s no real limits for these right now.
But this was certainly causing a large amount of trouble on that circuit. This was enough distortion to be heard in telephones and other types of equipment. And so you want waveform captures to be able to see exactly what’s going on.
How the Resonance Problem Was Fixed
A good question, how is this problem fixed in that situation? The manufacturer’s able to pair inverters together such that the PWM frequencies are 180 degrees out of phase with each other. So each pair of inverters canceled their PWM residual noise out. And so overall, that was able to reduce the distortion enough to where the resonance didn’t really matter.
Periodic Versus Triggered Waveform Capture
If you’re doing witness testing, a periodic waveform capture can be useful so that we can get waveforms at various points in the cycle that are normal steady state background. And you really need that because even at this distortion, it’s not that high a distortion relatively speaking, but it causes a lot of problems because it’s one specific frequency.
Going back to the paper, the recommendation for waveform capture is periodic capture, and this really depends on how long you’re going to record. Here in the white paper, we recommend at least one hour. So if you’re making a 24-hour recording, that gives you shots from the inverter output every hour, so you get different amounts of generation. Those waveforms are often uglier at lower outputs. So this can be very useful to have on record in case there are resonances or other types of waveform distortion. So you not only know which harmonics are present, but the shape of the waveform itself.
Now for the triggered thresholds, often these inverters don’t have abrupt changes in voltage or current, so the triggers for waveform capture aren’t necessarily as useful as the periodic capture.
DG Settings in PQ Canvass
We also have a special distributed generation setting in PQ Canvass for this application. If you hit the DG canned button, you can configure the recorder through PQ Canvass automatically for these recommended settings. But you can also set these manually either with PQ Canvass or with ProVision, and that turns on the periodic capture on one-hour increments and turns on the appropriate strip charts together, the RMS current and voltage, real power, reactive power, and frequency.
Common DG Problems and Voltage Regulation
Again, these are the most common DG concerns, and voltage regulation is the most common problem. A high voltage, even just on the secondary, where the inverter, the transformer will be fairly small, and so you have the opposite of voltage drop across the transformer. You have a voltage increase across the transformer. So the secondary voltage is higher than what the primary voltage is, taking into account the transformer ratio.
So the inverter raises the secondary voltage because there’s a high impedance across the transformer, and then it trips off because of high voltage. And so now you have a customer that can’t generate the way they want to.
One solution to that is to upsize the transformer, lower that impedance. That more tightly couples the inverter to the primary side. Or increase the size of the connectors on the service drop. Or sometimes the weak link is the cable from the meter base to the inverter itself, if that’s long. So you may have the voltage increase just across that cable. So this is also a very common complaint, and this is a pure secondary side issue in many cases.
But again, to capture this, of course, you need RMS voltage, and you want that customer current and real power.
Summary of Recommended DG Monitoring Settings
That really covers the material here in the paper. These are the recommended settings for a DG monitoring system. The keys are enabling strip charts for real reactive and apparent power in addition to voltage and current, and also flicker and frequency.
You can record individual harmonics, but another way of doing this is setting the periodic waveform capture to give you these waveform captures all along the day, so you can look at the waveforms themselves and use the ProVision or PQ Canvass tools to look at high-frequency content.
We also talked about the thresholds for waveform capture, but again, unless you are energizing or de-energizing the system, likely the triggers won’t be as useful because the DG system itself doesn’t usually have abrupt changes in current.
Contact Information
If anyone has any questions, give us a call anytime at 1-800-296-4120 or send an email to support@powermonitors.com. Thanks everyone, and everyone have a great day.