I found an excellent pdf:
http://ieeexplore.ieee.org/iel5/9671/30555/01409326.pdf In installations where there are a large number of single-phase, electronic non-linear loads connected to the same neutral, a high load factor can be found in that neutral. In these installations, the neutral current may exceed the phase current and a special attention must be given to sizing the neutral.
The diversified power absorbed by such a group of loads is generally limited, and even if the neutral current exceeds the line current, then the neutral conductor capacity is only exceeded in extreme circumstances if its size is equal to the line conductor's.
A common practice in these conditions is to use a 200% neutral conductor. This possibility is admitted by the National Electrical Code (USA), but does not form part of the electrical/ building regulations.
In high power installations (order of magnitude: P>lOOkVA or I>lSOA), various factors contribute to reduce
the neutral conductor load factor:
- More and more high quality IT equipment (work stations, servers, routers, PC, UPS, etc) include Power Factor Correction circuits, reducing considerably the generation of 3d harmonic currents
- Heating, Ventilation and Air-conditioning (HVAC) equipment in large buildings are supplied by a threephase network, and as such do not produce 3d
harmonic currents
- Fluorescent lighting equipment (with magnetic or electronic ballast) are generating 3" harmonic currents which are phase shifted with harmonic currents generated by PCs, giving a partial vector cancellation.
This attenuation effect is all the more as the installation power is high.
Except in exceptional circumstances, the 3d harmonic level in these installations does not exceed 33%, so the neutral current does not exceed the line currents. It is not therefore necessary to use an oversized neutral conductor.
