mflam…
Assuming a strongly looped and well sized distribution system with relatively low head losses, it is very possible that the flow in the pipes you are concerned about already reverses during emergency conditions (i.e. fire flows) and maybe even during normal operation.
Water systems are typically modeled for extreme conditions to establish an envelope of operation. But, what happens in the middle of the envelope? It is likely that flows sometimes reverse in the loops as demands continuously vary.
For modeling purposes, demands are usually assumed to be consistent by type, location, and time, but in reality demands continuously vary. While assumed consistency suffices for global modeling, it is not actually the case on a connection by connection basis or a neighborhood by neighborhood basis. For example, I can guarantee that your water use pattern and rate is as unique as your next-door neighbor's, and this applies system-wide. A neighborhood with lots of school-age children and working adults will probably show a much stronger morning peak than a neighborhood full of retired people. These little nuances can add up to temporary reversal of flow in a looped pipe from the normal direction. Only the dead ends are guaranteed to have one direction of flow.
Fred
==========
"Is it the only lesson of history that mankind is unteachable?"
--Winston S. Churchill