Synchronous Generator Harmonic Distortion

[mls1]

We monitored a small (~1MW) synchronous, 1800RPM, 13.8kV generator with a three phase measurement unit and found that that there was significant harmonic content in the current and voltage waveforms. Before synchronizing, the no-load voltage waveform (measured on the secondary of an open delta PT) had obvious distortion. An FFT of the waveform indicates that the distortion is significant 5th and 7th harmonic with also noticeable 23 and 25th harmonic.

After synchronizing, the current waveform was heavily distorted with ~30% THD. The FFT indicates that it is pretty much all 5th harmonic.

The unit has a brushless exciter with rotating rectifier assembly on the rotor. My question, is the presence of significant fifth harmonic in the current waveform a strong indication of a specific problem such as a bad exciter rectifier fuse? Has anyone seen this kind of distortion before and what was the resolution?

At this point, we do not have access to measure the exciter circuit but I think that is the next step but thoughts about potential sources would be appreciated.

 

[mls1]

Please note, in the above post I meant to say:
"is the presence of significant fifth harmonic in the current waveform a strong indication of a specific problem such as a bad exciter rectifier DIODE?" Not FUSE.

 

[ScottyUK]

Does this machine have a bus-fed exciter? Prior to sync the only load of any significance will be the excitation rectfier, which is a strong source of 5th harmonic current when operating normally. What do you mean by 'significant'? After synchronising, what is the load on the machine comprised of? If it is largely power electronics (rectfiers, converters, phase controllers, etc) then that would likely explain the highly distorted current.

It's unlikely that a bad exciter diode would influence the output voltage waveform because the inductance of the rotor is too large to allow the necessary large and rapid variation of current required to modulate the output voltage at a harmonic frequency.

Have you ruled out instrument error? Can you post a screen dump from your scope or power analyser?


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[mls1]

"Does this machine have a bus-fed exciter?" - No, it has a shaft mounted PMG as the exciter source

"What do you mean by 'significant'?" - 30% THD at a load current of ~6.6A which is ~16% of full load.

"After synchronising, what is the load on the machine comprised of?" - This unit is in parallel with a larger (5.5 MW) unit. Both are tied to a utility distribution system via a step up transformer. There are no significant harmonic load (a small battery charger, etc.) in the power house and most energy is exported.

"It's unlikely that a bad exciter diode would influence the output voltage waveform because the inductance of the rotor is too large" - that is a good point but wouldn't that only be effective when there is load current on the machine? the voltage distortion was measured with the unit breaker open and no load. Is it also possible the distortion is caused by slot harmonics?

Have you ruled out instrument error? - No, but the voltage distortion was measured right of the generator PTs and the currents were consistent between two sets of differential CTs (these were not properly sized and are only 75/5 with a C20 rating, there is a plan to replace them).

I've uploaded a shot of the current distortion.

 

[mls1]

Here is a shot of the voltage waveform before and after the generator breaker was closed.

 

[ScottyUK]

"that is a good point but wouldn't that only be effective when there is load current on the machine"

No, you still need a fair bit of excitation current just to raise terminal volts to nominal so I think it is valid regardless of whether the machine is synchronised.

The voltage waveform is a mess, I am going to guess that there is some adverse interaction between the two machines because of their winding design. I hope Wolf sees this thread because I think he has rather more machine design experience than I do.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[mls1]

"No, you still need a fair bit of excitation current just to raise terminal volts to nominal so I think it is valid regardless of whether the machine is synchronised." - Yes, very good point.

One more data point, the voltage waveform is line-to-line AB measured from open delta PTs. Overlaying the BC waveform and the point by point difference (CA) they are all exactly symmetric. I think this eliminates stator and external influences and limits the cause to the rotor or exciter.

This waveform is the a little over one cycle of the voltage wafeform with the fundamental filtered out (by zeroing the fundamental component of the FFT and taking the inverse FFT). This pattern repeats every cycle.