Control technique for a 3-phase inverter

[festivus]

Hi

Is there a way of performing a voltage control on a 3-phase voltage source inverter without making use of the dqo transformation? Should the voltage reference signals for such a case be sinusoidal or constant rms values? Any sort of help would be great! Thanks.

 

[FvM]

You should clarify the inverter topology and application. Is it a kind of UPS with output inductors or a motor inverter? Do you have additional output filters?

Generally I would expect a sinusoidal reference for a sine UPS. And most likely a voltage control loop manipulating the reference magnitude using a multiplier. If you are able to control the instantaneous output voltage directly, depends on PWM frequency and output filter dimensioning.

 

[festivus]

It is the standard 3-arm inverter topology with a LC filter at the output side. For now, I was simulating such a technique.. So theoretically, any PWM frequency should be possible.
The multiplier that you are talking about, for one particular multiplier value, it seems to control for a certain range of input voltages only. How do you decide this multiplier value actually? I decided more on a hit-and-trial basis only.

Thanks a lot for the help.

 

[FvM]

You didn't yet tell about the exact application. I assumed an UPS inverter where you try to keep the output voltage constant independant of load variations. I also assume, that the filter group delay effectively doesn't allow a direct closed loop control of instantaneous output voltage. A feedforward control based on measured load current may be possible however. So you can only implement a slow control loop reacting on rms or average rectified output voltage.

Obviously the voltage manipulation range will be limited.

 

[festivus]

Yeah, it is to keep output voltage constant independent of load variations. Actually, I had earlier simulated a half bridge 1-phase voltage source inverter using a direct closed loop control of instantaneous output voltage. It wasn't controlling too well for very different load values and/or reference voltage values.. So using a feed forward control should improve that, is it?
What if the control loop is reacting on instantaneous voltage instead of rms.. assuming balanced loads and that the output frequency remains constant?

 

[FvM]

What if the control loop is reacting on instantaneous voltage instead of rms.. assuming balanced loads and that the output frequency remains constant?

I would expect loop stability problems with reasonable gain values. The behaviour can be checked in a simulation using actual filter parameters.

 

[festivus]

All right. I'll try both ways. Thank you!
Just one more thing, can you also help out a little regarding how to decide the parameters of the output LC low pass filter? Say, I want a 220V rms at 50 Hz output in a 1-phase inverter..

 

[FvM]

The filter parameters will also depend on switching frequency, nominal output power and acceptable THD.