Testing a 3 phase 12 Kilowatt Solar Inverter Using a PV Simulator and IntegraVision PA

Hi, I’m Bryan Boswell with Keysight Technologies.
I am a design engineer, and I will be describing a PV inverter test set using the Keysight
Technologies N8937 APV photovoltaic array simulator and the PA2203A IntegraVision Power
Analyzer. The device under test that we will be talking about today is a 12 KVA transformerless
480 volt 3-phase inverter. The input is normally supplied by a solar array. We will be simulating
that with the N8937 APV, which is sourced by a 208 volt 3-phase input which then generates
the DC input to the inverter. The output is connected to 480 volt 3-phase, and I will
be measuring the output of the inverter using channels 1, 2, and 3 on the three phases connected
to neutral. On the input I will be using channel 4 to measure the voltage and current of the
DC input to the inverter. These are wired into the power analyzer. That’s our configuration.
I will now talk about how it is actually wired. Our 208 3-phase input comes into our power
supply, which does the 3-phase AC to DC conversion. The output of the power supply then supplies
the input to the inverter mounted here. The inverter output is then connected to 480 volt
3-phase, which then goes out to the 3-phase power of our building. Power is being transferred
from the 208 through back out to the 480 volt 3-phase. One thing I might note here on my
wiring conventions is – because I am using a standard breakout box here – I’ve defined
the power as being delivered from the wall. So, positive power would be coming out of
the wall. Because the power flow through our test system is from the 208 to the 480, I
should expect to see negative power as measured on our input group. We will see that once
we power up. With that, I will go ahead and get set up.
For consistency, I like to store all of my setups on my USB stick and then just recall
them when I want to use them. We’ll put that in and recall our setup. There’s my setup.
I’ve now got the power analyzer configured. I’ve preconfigured the PV simulator and what
I’ve got it doing is tracking to a solar simulation curve which we program with the short circuit
current and the open circuit voltage and then the maximum power point voltage and current.
This is the output characteristic which corresponds to the following power curve, and you will
notice what we expect the inverter to do is start the open circuit voltage, increase power
until it reaches the maximum, and that is where it is going to track to that maximum
power. Once we power on the system, we will actually see that works.
With that, let me go ahead and apply the 480 volt input to the system. We now see I’ve
got input voltage that we’re measuring that is coming into the inverter. I haven’t actually
turned on the array simulator, so nothing is actually being converted yet. But, we can
go into the system and actually look at the voltages being measured; 280 volts lined to
neutral, which of course is 480 volt 3-phase. Our voltage is there, and if you look at the
power, very little power being developed at this point. To kick things off, what I’ll
do – we can see that the array simulator is in the curve mode. It is set up for 200
volts at 20 amps, which would be 12 KVA. It is exactly the maximum power of our inverter.
So, we’ll go ahead and start that. If you notice, the power supply went to 750 volts
just like it is supposed to. We are now sitting right at this point waiting for the inverter
to actually turn on. In a few seconds you will hear the inverter start to turn on and
it will start converting power. Give it just a second. Now we can hear the inverter starting
to kick on and as you watch the power supply you will actually see when it starts to sense
the power and do some conversion. There you are starting to see some current flow. I actually
like to click this over and watch the power buildup. This is the output voltage and this
is the output DC power being delivered by the N8937 APV. We are up to 5 KW, and
if you look at the front of the power analyzer you see we’ve now got current being built
up as well as the power waveforms being displayed. More and more power is being converted by
the inverter from the DC out to the 480 volt. Voltage waveforms are at the top; the current
waveforms are in the center, and I am displaying power waveforms here on the bottom. We’ve
pretty much reached our maximum power point of 12 KVA. Things are pretty stable right
now. If we’re looking at the grouping here on the output power, again, the voltage is
still 280 volts. I now have 13.8 amps flowing, which converts to 11.65 kilowatts. Notice
the sine – negative power being driven out back onto our 480 volt grid. So, the system
is now fully up and running. The next thing we might want to look at is: What is the efficiency
of the system. If we look at our CWA analysis, one of our options here is efficiency. I have
predefined a couple of different efficiencies that we will be looking at, efficiency 1 being
the power being delivered from channel 4; that’s the DC input to the inverter, and the
output being the group A, or the output of the inverter. If you look at that setup, it
is showing an input power of 11.9 kilowatts, an output power of -11.65, for a total efficiency
of -97.4 percent. This particular inverter, the manufacturer specifies efficiency of greater
than 97 percent, so it appears to be doing what the manufacturer said it would do. I
also set up a second efficiency set to look at. In this case, I am looking at the input
being the group A total power, and the output being one of the phases of the power. We will
see how well the power is being converted on a phase-by-phase basis. If you look at
this case, -11.6, -3.9 for an efficiency of 33.45. You can see it is being split fairly
equally between the three phases. You can look at each of the phases if you would like
by scrolling through them. Again, just about exactly 33 percent on each one of those. Another
test that you might want to look at would be the harmonics of some of the signals. In
this case I am looking at the harmonics of one of the phase currents. In this case phase
current 1 which is displayed, has a total harmonic distortion of about 4 percent, and
each one of the harmonics of that are shown here. You can also display that in a table
format and see the various sizes of each of the harmonics there. You can look at the same
thing for maybe one of the voltage channels. In that case, a significantly lower THD, higher
signal integrity. The voltage is actually measured at about 1 percent in this case.
Another type of measurement you might want to be making on inverters would be the phasor
diagram. In this case we are going to look at it for group A, that being the output.
The phasor diagram in this case is very close to ideal, with each one of the voltage waveforms
being 120 degrees out of phase and the current being nearly identical to 180 degrees out
of phase. We can verify that. If you come back to the power quality panel and look at
the power factor you see the power factor on group A at .993; very, very efficient inverter
at a very high power factor. As I was putting together this little demonstration, one of
the things that I noticed about this setup that surprised me was when I looked at this
current waveform I noticed that there was a small signal on the input power – that
being channel 4 – signal. I was interested because the input is actually DC but I can
see that there is a waveform on there. The PA2203 has the ability to zoom in on that
signal and investigate it. So, one of the things that I did was take a look at that
current signal. The best way to do that would be to be AC coupled on it. We are now AC
coupled. We will center the waveform. We will turn on auto-ranging
and we will expand that waveform such that
we can actually look at it. We will do the same thing with the voltage waveform, increase
the resolution on that waveform. We’ll do the same thing on the power waveform. We’ve
now got them AC coupled, autoranged, and zoomed in to the point where we should be
able to see some signal. I notice that it is a high frequency signal. I increase my
sweep speed and I can now see the signal come over to my power analysis, switch over to
channel 4, and I can see that the
frequency of the current waveform is switching as an AC signal at 32 KHz, which I suspect
is actually the switching frequency of the inverter itself. My AC coupling on the waveform
– we’re able to remove the DC offset, zoom in on the signal, and actually see the switching
waveform. The other thing that you may notice as you are watching the waveform here is occasionally
I am seeing an increase in the power waveform. That might be a second thing we could zoom
in and look for. One of the things that the PA2203A has is the ability not only to see
the power waveform, but you can trigger on the power waveform. Let’s change our triggering
from voltage 1 – which was the nice stable channel – to the power 4 waveform. Once
we do that we can adjust our trigger level to then see the intermittent power levels
that are occurring there. We know it is an event that is happening fairly seldom, so
I decrease my sweep speed to the point where I can actually see this intermittent event
and trigger on it. What we see is occasionally there is an adjustment to the voltage waveform.
By triggering on this power signal we can zoom in and see that. This is just an example
of some of the measurements that you could make with the PA2203 and find information
in those signals that can add insight into the design that you are trying to analyze.
With that, I am going to call the video to a conclusion. Hopefully you have enjoyed the
information presented. Thank you.

1 thought on “Testing a 3 phase 12 Kilowatt Solar Inverter Using a PV Simulator and IntegraVision PA

  1. Thats what I call power electronics! Would like to see in those things.
    Great power analyser, lots of useful information you can get with it!

Leave a Reply

Your email address will not be published. Required fields are marked *