Here are a few thoughts:
A dead-end pipe with zero demand at the downstream end will have zero flow.
A looped pipe with the same (or nearly the same) HGL at each end will have zero (or nearly zero) flow. This case is not common, but I've seen it often enough. In a looped system, this would be a transient condition in real life because demands are constantly fluctuating in ways we can't model.
A pipe with a check valve that has a higher HGL on the downstream side of the check valve will have zero flow because this condition will close the check valve. This can also happen with certain types of control valves.
A pipe that has been set in the model to be "closed" will have zero flow.
The pipes on the suction and discharge sides of a pump that is set to be "off" in the model will have zero flow.
A pipe with a typo for the diameter (e.g. 1" instead of 10") will have only a tiny fraction of the flow you expected. So, not zero, but very close to zero.
I may be missing some other situations, but this should get you started.
==========
"Is it the only lesson of history that mankind is unteachable?"
--Winston S. Churchill