...
Perhaps, it is easier to prove inverter-power generation is isolated from building loads, and harmonics, if grid-tied or energy-storage inverters never feed loads directly. Rather than prove triplen harmonics are canceled within inverters, or explain inverters that feed 120v / 277v loads directly without utilizing any neutral current.
Well, right. See below.
Thanks for the references.
So, jgatzka I think the summary is 'it depends on the specifics of the inverter'.
1) If the inverter output topology is essentially delta, then the inverter _cannot_ place current on the neutral, no matter what the harmonic loading or fundamental imbalance. In which case not only would you not need to take the 0.8 derate, you could plausibly reduce the size of the neutral. Good luck getting suitable documentation!
'Reducing the size of the neutral' would mean not running a neutral to the combiner at all in that case.
But we know that the OP's inverter output is wye.
2) If the inverter output topology is essentially wye, then the inverter can supply current on the neutral. The neutral will need to be at least full size.
Again, not necessarily. See article 705 references up thread. But apparently getting SMA to stipulate is in question.
3) If the inverter output topology is wye, _and_ the inverter output current includes harmonics drawn by the load, then additive harmonics could show up on the neutral and you would need to take the 0.8 derate if the load meets the requirements of making the neutral a CCC.
Somewhere in the 480/277V system there must be a component which carries any unbalance neutral current and permits harmonic currents to flow. However this component is not necessarily the inverter; the inverter can be feeding the system in a pure delta fashion, and the utility transformer could supply the neutral currents (current on the wye neutral would show up as current circulating in the transformer delta coils, with _power_ coming from the inverter but no neutral current coming from the inverter).
Certainly the utility transformer can supply unbalanced current. I wonder first of all if the 'seasoned EE' has even considered this. There is no indication in the OP that the EE has even brought up harmonics. But the harmonics question is certainly relevant and interesting as well.
The inverter in question has a "Harmonics THD" specification on the data sheet of ">3%". I'm not sure if this specification means that the inverter will shut down if it cannot maintain that specification on it's output, or if it means something slightly different. For what it's worth, my understanding is that this inverter has three 277V outputs internally but that it is configured as a unit to always output equal (or very, very close to equal) currents on all three phases and will shut down if it cannot. I appreciate tallgirl's point about how the inverter references voltage (ramsy's solectria reference also speaks to this). However, to my understanding, the effect of system harmonics on inverter output current would be limited to the magnitude of changes in the fundamental voltage waveform. (i.e. the inverter current does
not follow the non-linear current waveforms of the loads). My intuition is that these voltage changes would be small enough to justify not-applying 310.15(B)(5)(c). But admittedly I really lack the skills to do the math on that. So please correct me if I'm getting something wrong.
Others here with actual field experience will describe which of the above are most plausible, but I think you have to go to the inverter documentation to provide an answer for the EE.
-Jon
Agreed. I mean, I remain interested in this from a personal educational standpoint, but my advice to the OP at this juncture might be that it will be less of a headache to upsize conductors. Still, I wonder if the 'seasoned EE' that is the subject of this thread has actually considered the behavior of a grid-tied interactive inverter or is just making an unwarranted assumption that the inverter somehow follows the load. (It doesn't.)
