I've been doing some research on ways to drive the half-bridges in a 3-phase, grid-tied, pure sine wave inverter. I knew about SPWM but not about SVPWM. Based on what I've read, it's superior to SPWM in multiple ways, but what exactly does an SVPWM waveform look like compared to SPWM? Is it even used in grid-tied inverters? Everywhere I've seen it so far, it was in the context of a motor speed controller, and the output waveform didn't look exactly like a sine wave (like on the image in Shameer N's response to this question). Lastly, are there any other alternatives to SPWM and SVPWM?
-
\$\begingroup\$ Link to a document that describes or mentions SVPWM please. \$\endgroup\$– Andy akaCommented Jun 25, 2020 at 21:04
-
1\$\begingroup\$ electronics.stackexchange.com/questions/501397/… \$\endgroup\$– user16222Commented Jun 25, 2020 at 21:42
-
\$\begingroup\$ Oh I thought it was a question about PSWVTF ... which is pseudo smoothed wigner ville time frequency distribution, an advanced form of FFT. \$\endgroup\$– bandybabboonCommented Jun 26, 2020 at 6:10
1 Answer
It is a space-vector PWM modulation. It is superior to sine-wave modulation in a way that better utilizies the DC link voltage by adding triangular waveform at frequency three times the fundamental frequency. This makes phase voltages to look ugly, but line voltages are sine waves as third harmonics cancel out.
If you want to generate three-phase voltage, then by using the sine-wave modulation the bipolar duty cycles are calculated as:
ma = va / (vdc/2)
mb = vb / (vdc/2)
mc = vc / (vdc/2)
where ma,mb,mc
are bipolar duty cycles in range -1..1
, which translates to 0..100%
on-time for high-side switch.
The space-vector modulation would just add the median in ma,mb,mc
to all three duty cycles:
ma += (mmed/2)
mb += (mmed/2)
mc += (mmed/2)
This is also known as “zero sequence voltage”.
As for the alternatives, there are actually different variants of space-vector modulation. For example there is a variant that reduces number of switching cycles by 30%, but at a cost of increased THD in current. You could also add sinusoidal waveform instead of triangular at frequency three times the fundamental frequency, also known as “third harmonic injection”. This would better utilize the DC link voltage than sine-wave modulation, but still not as good as space-vector modulation.
Note that all these techniques are not appropriate if you connect neutral, because common mode voltage (3rd, 9th etc. harmonics) will close through the neutral which will cause a large current in that conductor. As a side note, imagine you have star-delta transformer and you connect the star directly to the grid, including the neutral conductor. You will see a fairly large current in the neutral conductor due to the third harmonic in the grid voltage!
I focused here mostly to directly answer your questions. Of course, there is much more to this topic than explained here. You can find some useful illustrations in the following post:
Difference between SVM and sinusoidal PWM in BLDC
Here is also a very useful tool to illustrate different modulations:
https://microchipdeveloper.com/mct5001:start
If you want to dig deeper into the theory of this, I recommend
“Pulse Width Modulation for Power Converters” by Holmes and Lipo
But be aware, this is not an easy read.
-
2\$\begingroup\$ Great answer that would be even better with some plots :) \$\endgroup\$ Commented Jun 25, 2020 at 21:39
-
\$\begingroup\$ Thanks for your answer, I’ve got a few new search terms to work with. You mentioned SVPWM makes the phase voltages look ugly, so does that mean this would not be suitable for a 3-phase wye grid, where pretty much all loads will be connected to neutral and a phase? As for the book recommendation, nearly 700 pages seems like quite a bit for someone who doesn’t read a lot but I’ll definitely consider it. \$\endgroup\$– CedricCommented Jun 25, 2020 at 23:56
-
1\$\begingroup\$ @pericynthion I will rather focus on directly answering the question. There are other useful places where to find illustrations how space-vector works. \$\endgroup\$ Commented Jun 26, 2020 at 7:20
-
1\$\begingroup\$ @Cecemel You are right, it should not be used with neutral, as the common mode voltage (3rd, 9th etc harmonics you add) will close through the neutral. I modified my original answer to also include this information, and I also included some other useful resources. \$\endgroup\$ Commented Jun 26, 2020 at 7:23