In the 2012 NFPA 70E, there are provisions for both AC and DC voltage. What is the case if you have both AC and DC components to the voltage. For Instance, in converting DC from solar cells to AC you might have an AC voltage riding on a DC potential. So you can have ungrounded 380 VAC Line-to-Line, but the potential to ground might have an RMS of 600 VAC - the voltage on an oscilloscope would look like a 380-volt ripple on 500VDC.

I'm surprised that you would ever see a DC voltage riding on the AC in a full wave rectifier, whether performed by free wheeling diodes or a valve (IGBT) arrangement. I have not actually see this in any of the equipment I work with.

However, if that is the case, from a shock hazard potential one of the big differences is that AC is almost always expressed in RMS terms whereas DC is absolute potential. I would make the attempt to read the shock hazard table using the worst case of both because there's not a great deal of difference at that voltage.

For arc flash, it's a much more difficult proposition. The major difference between DC and AC is that with AC, the actual arcing waveform is more or less a truncated, ringing square wave because the arc extinguishes at each zero crossing of the incoming power and then does not reignite until the voltage is high enough for this to occur. In the case you describe, we have no zero crossings so it would be inappropriate to add the DC and AC contributions separately. The AC component confounds the situation because the assumption that DC is...constant, does not apply. That being the case right now the DC equations are effectively just power x time x the energy flux through a sphere at a given working distance assuming 50% of the energy is released as radiant energy. I would model it using the DC equation from Annex D.8 using the RMS voltage which would be the DC equivalent with the AC component.

It is actually AC riding on DC (like a huge 480-volt, 60 Hz ripple), and it is really quite common for inverters depending on how the DC Bus is set up, I have seen inverters use a positive - negative bus arrangement (where the common connection between the two busses may actually grounded) and also use just a single DC bus. Inverters have come a long way, and we tend to think of them generating an AC voltage, when they are actually chopping up a DC voltage. I am not sure if that is significant in regards to arc flash or not.

You would see the AC riding on DC voltage on the inverter output to the motor. Inverters derive the output to the motor from a DC bus. The motor sees an alternating current. With a six-step inverter the voltage looks somewhat sinusoidal, but the current does not. With a PWM inverter, it is hard to see anything sinusoidal about the voltage, but the current that the motor sees is sinusoidal.

Inverter-duty motors require insulation rated for the bus voltage, which may be 650 - 720 volts. The effective voltage may be only 480 VAC, but the PWM waveform is peaking at bus voltage. If you try and use a standard motor on an inverter, it will only be a matter of time before the motor fails because the insulation is not rated for high enough voltage.

While the motor current crosses zero 60 times a second at base speed, and phase-to-phase on a scope, you will see zero volts, the actual voltage at the point where the motor crosses zero current might be 300 VDC phase to ground (The actual voltage to ground would vary depending on several factors).

Inverters have circuitry and logic to detect ground fault and phase-to-phase shorts as well. They often annunciate a fault without the accompanying loud noises and they generally will not produce an output when they detect a fault.

Attached is a simplified diagram of an inverter. I’ve only shown one phase of the output.
There are essentially three voltages that come into play:
The voltage feeding the DC BUS, 1000VDC voltage from a DC source,
The AC output voltage, 315 VAC RMS voltage you measure phase-to-phase,
And the 660 VAC RMS voltage you would measure phase-to-ground.
The output is actually AC riding on DC (like a huge 480-volt, 60 Hz ripple, actually a chopped-up DC voltage.
Now, as far as arc flash goes, would you do the calculations for each voltage and rate the enclosure based on the worst-case?

In the case I am working on, it is going to get a bit complicated as there are panels that access just the DC – those are pretty easy to calculate. There are panels that just access the AC – again, those are pretty easy to calculate. But then there are the majority of panels that are an intermingling of voltages. Some are partially shielded by interior panels, and backplates