Exploring the Benefits of 6 Phase Inverters: A Scientist's Perspective

[tim9000]

Simple question, What's the advantage of having more phases in an inverter? Like why have 6 or twelve phases? Is it something to do with harmonics? I know that the higher the switching frequency the further off in the freequency spectrum you can push the harmonics, but I can't see that being relevant here.
Cheers

 

[anorlunda]

Yes it is possible and it has been thoroughly researched, but it never found widespread application.

Google "high phase order" or "polyphase"

https://en.m.wikipedia.org/wiki/Polyphase_system

 

[tim9000]

Yes it is possible and it has been thoroughly researched, but it never found widespread application.

Google "high phase order" or "polyphase"

https://en.m.wikipedia.org/wiki/Polyphase_system

thanks
Ok, so 'Higher phase order' does reduce harmonic generation, but that didn't say how.

 

[anorlunda]

thanks
Ok, so 'Higher phase order' does reduce harmonic generation, but that didn't say how.

Any inverter approximates a sin wave. The more phases, the better the approximation.

 

[tim9000]

Any inverter approximates a sin wave. The more phases, the better the approximation.

Ah yeah, yeah so the inverter sine wave is like little discrete steps.

 

[sophiecentaur]

I reckon this boils down to a specific application of basic sampling theory. The more samples per cycle (i.e higher the sampling rate) the higher the harmonic artefact frequencies and the lower the distortion in the waveform. In most cases, the distortion is referred to as Quantising Noise but, as the output frequency is locked to the sample frequency, it's simple harmonic distortion.

 

[Hesch]

If the load of the inverter is a 6 phase motor or transformer, the inverter must have 6 phase output.

So maybe the question could be: Why do we have 6 phase motors/transformers ?

 

[tim9000]

If the load of the inverter is a 6 phase motor or transformer, the inverter must have 6 phase output.

So maybe the question could be: Why do we have 6 phase motors/transformers ?

If I asked that question would you have any thoughts to throw in about an answer?

 

[tim9000]

I reckon this boils down to a specific application of basic sampling theory. The more samples per cycle (i.e higher the sampling rate) the higher the harmonic artefact frequencies and the lower the distortion in the waveform. In most cases, the distortion is referred to as Quantising Noise but, as the output frequency is locked to the sample frequency, it's simple harmonic distortion.

Well that's sort of what I thought, I mean I know that the MMF of each phase in a motor super impose to form one sine wave, so I can imagine 6 phases adding to form a cleaner MMF wace in a motor. But Yeah the switching frequency used in the inverter is what I thought would determine how close the inverter sin wave would be to an actual sin wave...not the number of phases?
the only real thing I remember about inverters is that hexagon where each point is represened by three binary digits to make the vector you want.

 

[anorlunda]

If the load of the inverter is a 6 phase motor or transformer, the inverter must have 6 phase output.

So maybe the question could be: Why do we have 6 phase motors/transformers ?

Careful, it is easy to get confused. We could invert to 6 phase power. Motors designed for that are better than 3 phase.

6 phase transmission has been explored. It is superior to 3 phase. That is better if the whole grid was 6 phase. But if 6 is only a small part of the grid, every interface to the 3 phase grid would need an expensive 6 to 3 transformer. Therefore 6 phase never got used in practice to my knowledge. Below are experimental 6-phase towers. The one on the right is double circuit 6P, not 12P.

But what I believe we are discussing in this thread is 6 phase inverters to supply 3 phase power. If you design the switching logic properly to select from 6 sources instead of 3, you can do a better job of approximating the 3 sinusoids. Any number of inverter phases >=3 can be used to generate 3 phase power. I hope I explained that clearly. I wish I had a graphical depiction of the waveforms, but I couldn't find it.

 

[Hesch]

If I asked that question would you have any thoughts to throw in about an answer?

Well, say you have a transformer with 6 phases, the output could be rectified more smoothly by means of a 12-diode bridge.

As I calculate it, the ripple ( peak-to-peak without LC-filter ) in theory will be:

1 phase: 100%
3 phase: 13.4%
6 phase: 3.4%

 

[tim9000]

Well, say you have a transformer with 6 phases, the output could be rectified more smoothly by means of a 12-diode bridge.

As I calculate it, the ripple ( peak-to-peak without LC-filter ) in theory will be:

1 phase: 100%
3 phase: 13.4%
6 phase: 3.4%

out of curiosity could you show that rough calculation?

 

[tim9000]

Careful, it is easy to get confused. We could invert to 6 phase power. Motors designed for that are better than 3 phase.

6 phase transmission has been explored. It is superior to 3 phase. That is better if the whole grid was 6 phase. But if 6 is only a small part of the grid, every interface to the 3 phase grid would need an expensive 6 to 3 transformer. Therefore 6 phase never got used in practice to my knowledge. Below are experimental 6-phase towers. The one on the right is double circuit 6P, not 12P.

But what I believe we are discussing in this thread is 6 phase inverters to supply 3 phase power. If you design the switching logic properly to select from 6 sources instead of 3, you can do a better job of approximating the 3 sinusoids. Any number of inverter phases >=3 can be used to generate 3 phase power. I hope I explained that clearly. I wish I had a graphical depiction of the waveforms, but I couldn't find it.

Out of interest, did you have any info on the physical construction of a 6 to 3 or 3 to 6 phase TX?

 

[Hesch]

out of curiosity could you show that rough calculation?

A 3-phase full wave rectifier will have 6 DC pulses ( green curve ) per 360 electrical degrees. So 1 pulse will have a width = 360°/6 = 60° = 2*30°.
The upper peak value of a pulse = V_peak* sin(90°) = V_peak*1.
The lower peak value of a pulse = V_peak* sin(90°-30°) = V_peak* sin(60°) = V_peak*0.866.
Thus a peak-to-peak ripple = V_peak*(1 - 0.866) = 13.4%.

A 6-phase full wave rectifier will have 12 DC pulses per 360 electrical degrees . . . . . . . . . . = 3.4%

 

[anorlunda]

Out of interest, did you have any info on the physical construction of a 6 to 3 or 3 to 6 phase TX?

If you just images.google.com "transformer core" together with words like three, six, single, polyphase. You'll see dozens of variations. Here is just one. You should do your own googling before posting questions here.

 

[tim9000]

A 3-phase full wave rectifier will have 6 DC pulses ( green curve ) per 360 electrical degrees. So 1 pulse will have a width = 360°/6 = 60° = 2*30°.
The upper peak value of a pulse = V_peak* sin(90°) = V_peak*1.
The lower peak value of a pulse = V_peak* sin(90°-30°) = V_peak* sin(60°) = V_peak*0.866.
Thus a peak-to-peak ripple = V_peak*(1 - 0.866) = 13.4%.

A 6-phase full wave rectifier will have 12 DC pulses per 360 electrical degrees . . . . . . . . . . = 3.4%

Aaah, cool.
Thanks

 

[tim9000]

If you just images.google.com "transformer core" together with words like three, six, single, polyphase. You'll see dozens of variations. Here is just one. You should do your own googling before posting questions here.

Thanks, I ddi a quick search but it was hard to get a schematic and picture together, that I liked.

 

[Windadct]

Actually 3 to 6 Phase transformers are very common in large DC applications and large VFDs, and other power converters for exactly the reasons above. ( Look up 12 pulse rectifier) - typically these are going for 30 and 60 Degree phase shifts to manage harmonics due to the non-linear ( switching ) load. - Generally created by connecting various combinations of Wye and Delta windings. -- Pure 6 phase would just be a Y-Y and a Y-(inverted)Y - so Phase A of the 1st trans is 180 Deg off set from the Second Transformer - but this would not help the harmonics issues as much as the 30 and 60 degree phase shifts. For direct feed to a motor - there are technical befits - but the economics (cost to benefit) of these systems is almost never practical. (As like to say "Just because you CAN does not mean you SHOULD").
Above there are some comments abut the output of the inverter - really a different set of issues - there are number of ways to get a cleaner - or more pure sine ( Higher Switching Freq as mentioned above, but also Multilevel , and cascade topologies are used particularly in IGBT based systems where the losses associated with Switching, make the High Fsw impractical.

 

[tim9000]

I enjoied reading your post, however you'll have to excuse me I'm a bit slow, and out of practice (a couple of years since I did a relevant subject at uni):

Pure 6 phase would just be a Y-Y and a Y-(inverted)Y - so Phase A of the 1st trans is 180 Deg off set from the Second Transformer

So what does the 180deg phase shift of Phase A accomplish?
All I pretty much remember from TXs is that a Delta Wye, traps the zero sequence current. I wouldn't mind a referesher on any other benefits of transformer combintations people cared to mention.

but this would not help the harmonics issues as much as the 30 and 60 degree phase shifts.

Sorry, could you elaborate on what those specifics are, please.

For direct feed to a motor - there are technical befits

So using a transformer after an inverter, between it and the motor would or 'could' add a lot of life to the motor, yet it would cost a lot? Is that what you're getting at?

really a different set of issues - there are number of ways to get a cleaner - or more pure sine ( Higher Switching Freq as mentioned above, but also Multilevel , and cascade topologies are used particularly in IGBT based systems where the losses associated with Switching, make the High Fsw impractical.

I may be out of line asking this, on this thread: I take it from you post that higher the frequency switching, the higher the losses (that might ring a bell but I can't remember why...if I had to guess I'd say somethign to do with high current from stored charge in on the junction during near instanious discharge on switching?)
But how does using 6 phases then converting to 3 phases, equal less losses than just 3 phase inverting at a higher switching rate? (as far as I can remember here all I've heard is that the AC ripple on the DC on rectification is smaller, not to say I didn't gloss over a realisation I should have made)
Did you have any links you'd recommend regarding "Multilevel or cascade topologies"?

So would I be right in inferring that a six phase to three phase transformer would do wonders in alleviating harmonics from a six phase supply? (from an isolation perspective)

Sorry to be a bother, I am interested in cleaner switching, I'm just pushed for time atm with a final year project..

 

[Windadct]

First I am curious about the 6 phase inverter you are referring to - and what is the application.

As for the 180 Phase shift - that is how a 6 phase system would look, relative to a ground point. The only real way to discuss is to look at the vector diagrams. ( http://electrical-engineering-portal.com/understanding-vector-group-transformer-1 )
For the Rectifier - a standard 3 phase system will be called a 6 pulse, meaning as each diode begins to conduct you see a current spike, with 60 Hz for each phase you see 2nd harmonic - on the DC side you see this effect from all three phases to there is 360hz ripple - etc. Both the AC and DC issues need to be dealt with. - Sorry to not write the whole thing up there are lots of good resources covering this. ( Example) .
As for feeding a motor I mentioned direct feed ( no inverter) - also depends on the motor and application if there is a benefit. In general the main reason I see a transformer put between an inverter and motor is to use Low Voltage inverter to drive a Medium Voltage motor. - Still not 6 phase. Most of the 6 phase inverter / motor combos are I am finding are academic - again they can be done, and have some technical benefit - but in the end they are not practical - typically due to complexity or cost.
Last point - as I said a different set of issues... 12 pulse rectifier is to reduce harmonics seen by the source. (consider the rectifier as being completely independet of the inverter)--- The inverter (feeding a motor, or a computer ( like a UPS) - or other devices like the Grid in a Solar Inverter - want to generate a clean signal for what ever the load is ( motors are actually about the easiest case - but they do not like the high dV/dt "noise" from the switching - when VFDs first came on the market they often needed ruggedized ( expensive) motors that were "Inverter Grade" to deal with this. MOSFETs are very close to an ideal switch and generat little losses on each turn on/ off - so many designers get used to running very high switching frequensies - IGBTs have much greater switching losses but can handle higher voltages and currens ( they parallel a little easier as well) - so for larger power the Switching frequency choice a major design consideration.

 

[tim9000]

First I am curious about the 6 phase inverter you are referring to - and what is the application.

As for the 180 Phase shift - that is how a 6 phase system would look, relative to a ground point. The only real way to discuss is to look at the vector diagrams. ( http://electrical-engineering-portal.com/understanding-vector-group-transformer-1 )
For the Rectifier - a standard 3 phase system will be called a 6 pulse, meaning as each diode begins to conduct you see a current spike, with 60 Hz for each phase you see 2nd harmonic - on the DC side you see this effect from all three phases to there is 360hz ripple - etc. Both the AC and DC issues need to be dealt with. - Sorry to not write the whole thing up there are lots of good resources covering this. ( Example) .
As for feeding a motor I mentioned direct feed ( no inverter) - also depends on the motor and application if there is a benefit. In general the main reason I see a transformer put between an inverter and motor is to use Low Voltage inverter to drive a Medium Voltage motor. - Still not 6 phase. Most of the 6 phase inverter / motor combos are I am finding are academic - again they can be done, and have some technical benefit - but in the end they are not practical - typically due to complexity or cost.
Last point - as I said a different set of issues... 12 pulse rectifier is to reduce harmonics seen by the source. (consider the rectifier as being completely independet of the inverter)--- The inverter (feeding a motor, or a computer ( like a UPS) - or other devices like the Grid in a Solar Inverter - want to generate a clean signal for what ever the load is ( motors are actually about the easiest case - but they do not like the high dV/dt "noise" from the switching - when VFDs first came on the market they often needed ruggedized ( expensive) motors that were "Inverter Grade" to deal with this. MOSFETs are very close to an ideal switch and generat little losses on each turn on/ off - so many designers get used to running very high switching frequensies - IGBTs have much greater switching losses but can handle higher voltages and currens ( they parallel a little easier as well) - so for larger power the Switching frequency choice a major design consideration.

Sorry about the delay, I actually didn't have an application in mind, I've only ever used a 3 phase inverter to control an induction motor...once.

Admittedly I only had a very brief look at those liks, I downloaded the second and had to bookmark the first, because I've got a project due in a month an too much else on my plate. But I think you're right, I need to study the vector diagrams, and I will be sure to in December. I think I did gleem the picture of the twelve pulse rectifier, though I'm still not certain exactly how it is an improvement over the 6 pulse rectifier I'm familiar with.

Ok, so you'd be happy to use a direct inverter to motor if the inverter was giving the correct voltage? (The harmonic arn't a major issue in your opinion?) And is this true for 6 phase too? There's no real benefite to use a 6 phase inverter over a 3 phase inverter, for a harmonics perspective?

P.S.
I found what you said about IGBTs and MOSFETs very interesting too. Cheers