tbtkdz, so you want to pose questions for the sake of arguing with the answers? Okay, I'm cool with that. Remember, the title of this forum is "Electrical Theory and Applications."
Of course, when the phases are balanced, there is no neutral current (ignoring harmonics). Is not the difference between two or three of the same number zero? That the difference is zero does not negate the validity of the math.
And also of course, at the moment of zero-crossing of the sine wave, no current (ignoring reactances) is flowing in either direction, nor into, or out of, the earth. We're talking two steady-state moments here; near the positive and negative peaks.
When one models a pair of current loops that share a conductor, the currents are diagramed and studied individually, even if that means showing two opposing currents in the shared conductor, and the resulting current is the mathematical difference.
However, your posts suggest that you believe that the instantaneous polarity of the alternating current affects whether any current is diverted to earth instead of directly back to the supplying transformer's terminals.
That one conductor is grounded in no way alters the current flow from its intended path. (Remember, we're talking theory here, and ignoring impedances and leakage currents.) In practice, we are never to intentionally use the earth as a conductor.
The reason a solid neutral is required from the source to the main disconnect, even if there are no line-to-loads, is to assure a low-impedance path to the source for the proper operation of over-current/ground-fault protective devices.
Neither half of a sine wave is "higher" or "lower" than the earth as far as current flow is concerned. A negative earth (relative to the hot wire) absorbs no more current than a positive earth does. Only a rectifier can cause the halves of an AC wave to flow differently.
In theory, no circuit current should ever flow through the earth, like an EGC, but in reality, the whole planet is like a giant equipotential grid, and some current is always in the earth, but that's not what we're discussing here.
The main difference between a system EGC and the earth is that an EGC is intended to carry fault current during a ground fault, but the earth shouldn't even then, because it is shunted by a low-impedance EGC and grounding electrode system.
As you obviously know, voltage is the difference between two points, and a circuit between those points is required for current to flow. No current flows into the earth because there is no potential difference between the neutral and the earth; the neutral to the source sees to that.
So yes, alternating current flows back and forth, like a see-saw, but the two halves of the waveform should be the same, and take the same conductive pathway. Again, only a rectifier can cause a difference between the positive and negative wave halves.
Therefore, the polarity at any given moment has no affect on the current flow's path. Whether the electrons are moving from earth to hot, or hot to earth, they take the route through the transformer's secondary, through the load, and back to the secondary.
Whether that circuit involves a grounded conductor or not is of no relevance. Only a ground fault, whether bolted or high-impedance, involves an outside path. Hopefully, that path does not involve a person. That's why we have GFCI's.
Whew! My fingers are tired!
