Effective grounding requirements & applicability to inverter based DG

[ReRoy]

There are several utility codes in NA that require ANY distributed generator (DG) whether a rotating shaft generator or a grid-tied inverter based generator (e.g,,PV, Wind) to have a "solidly" or "effectively" grounded neutral (EGN), per IEEE definition. Most medium sized (up to 500kW) inverter based DGs for PV systems have a 3-PH/delta connection(usually 208V or 480V), and therefore no accessible neutral point which to ground.

The utilities maintain that even these devices need an EGN and must have a grounded neutral conductor. To accomplish this requires an additional delta/wye isolation transformer or possibly a zig zag transformer - or changing the internal transformer output winding (requires safety agency recertification)in order to comply.

However, inverters are not voltage based machines - they are current sources that monitor the connected grid voltage and inject current to the grid based on the power available from the PV array. The terminal voltage developed by these devices will always follow that of the connected grid. If there are any disturbances in the grid voltage outside of the requirements of UL1741 the inverter is automatically disconnected. The units are smart in that they measure line to line voltage both in magnitude and phase, and via calculations determine the equivalent neutral voltage, and switching at frequecies up to 18kHz have more than sufficient capability to guarantee no variation from the appropriate center point.

The symmetrical analyisis model of this class of generator has no zero or negative sequence components - it is capable only of positive sequence operation. The EGN requirement is based on rotating machinery and the definitions of "effective grounding" are based on zero and negative sequence currents flowing - and they do not in this situation.

It is my contention that the extra transformer to accommodate an EGN is not necessary for this kind of DG. Does anyone have any comments or experience with this topic - from your experience does the argument given above hold water?

 

[davidbeach]

I'm sure there is an impedance in the negative and zero sequences, just like every thing else connected to the electrical system.

The whole problem is time to disconnect. One unit may disconnect very quickly, multiple units, of multiple manufacturers may take a bit longer. Well balanced against remaining load in an island it can take even longer. Should that island be ungrounded because it is separated from the transformer at the substation and the DG isn't effectively grounded, no current flows into the ground fault, and the unfaulted phase voltages approach line-line voltage to ground. Nothing connected line-ground likes that 173% voltage.

That's way it must be effectively grounded.