When did this occur, recently?
Maybe depends on what you call recently. It was about seven or eight years ago.
What do you suspect were the large non-linear loads causing the condition?
Primarily controlled lighting. Dimmers. About 3MW worth of the little blighters. None a problem itself but in aggregate......
Here's one of the harmonic spectrum analyses I did for a single channel. It was based on actual measurements with a digital scope and the data dropped into a spreadsheet. The harmonic spectrum was calculated from that.
It illustrates the problem quite well. The fundamental component is about 4.2A and the third harmonic (Hn
3) about 2.2A or maybe a shade more.
The system is three phase. With same load on each of the three phases, the fundamental components cancel. But the Hn
3components add arithmetically.
So the neutral current from this component alone is 6.6A. Or over one and a half times the current in each phase.
I generally don't size neutrals larger than the phase conductor anymore, because many offenders have become more sinusoidal to be sold in Europe and they use the same design here.
The installation I mentioned above was in the Middle East.
As you are probably aware, my main area of activity is in the variable speed drives field.
The usual customer specification requires compliance with the Electricity Association's G5/4 which is, to a significant extent, concerned with supply voltage distortion at the point of common coupling. (Sounds a bit rude but I'm sure you know what I mean.)
There is a "harmonised" IEC standard (IEC61000) for harmonics but, from memory, it is applicable to equipment that draws between 16A and 75A which would exclude your average dimmer, CFL. wall warts etc. In fact anything that draws less than about 3.7kW at nominal voltage.
It may have been updated since, but I'm aware that it was current (no pun intended) in 2005.