Seems like instead of laying siege to one of their big cities, the best way to quickly bring them to their knees would be to cut off, or significantly reduce their fresh water.
And if the Romans DID defend these points, how many soldiers would be there? I assume not very many... So why weren't there more stories about enemies of Rome destroying their water wheels/ aqueducts?
Actually Roman Cologne had a qater supply over a distance of roughly 100 km. Mostly the track would be underground, only when valleys needed to be crossed there were aquaeduct bridges as they are known from other points. But, even the underground parts had acsessing points in regular intervalls, so that basic cleaning could be done easily. A few of those have been excavated, one of them is accessible today. The archaeological research for the water supply line is relatively good, and an modern engineer predicted the course of it after very careful measurements of the landscape.
As a Roman freshwater supply system has to maintain a gradient line, he predicted the course very precisely and when new bulding activities or roadworks take place, the Roman water line is ususally very near where he guessed it to be^(1).
That said, we believe that on the western side of Roman Cologne near a gate there has been a water tower, and from there on, the water could flow down into the city to fill the Thermae, and all the wells in the city. The water in those was constantly running, flowing over the rim, taking some of the debris from the streets and into the sewers.
This is relevant, because there has also been archaeological research in the sewer systems. The layers of debris can be dated. But, it should be noted, that during an active water supply, there wouldn't accumulate much debris, because of the constant flow of fresh water through he wells, and of course the rain that was also flowing into the sewers. However, the layers in the sewer system suggest a remarkable drop in the water flow in the beginning/mid fourth century. (Excavation picture can be seen here: http://www.museenkoeln.de/archaeologische-zone/default.asp?s=2980&tid=424. Remarkable are the thin layers at the bottom, with a huge block of dark grey layer above, and then a bigger, inhomogenous filling topmost.)
355 the city of Cologne was conquered by Frankish warriors, though from 274 the province was considerably less safe than before, and a lot of raiding vici (small settlements) and villae rusticae (farmsteads) took place. So, thinking is, that indeed the main freshwater system to Cologne was destroyed in that time, but as far as I am aware, a direct destruction layer on the water line that can be dated to that time is still missing.
1: Klaus Grewe, Atlas der römischen Wasserleitungen nach Köln, Cologne 1986.