Smart water network management with in-pipe leak detection robots

Abstract

In this thesis, I created methods and designs to implement smarter, more autonomous water distribution networks (WDNs) and also improved the robots which will travel within the WDN's pipes to better differentiate pipe leaks from bumps in the pipes. Starting from the unit of the in-pipe leak detection robot, I investigated ways to make its soft leak sensors able to differentiate between pulling (due to leaks) and bending (due to bumps), and showed how a new design of adding fabric to the soft sensor allows the sensors to differentiate bending from pulling. Zooming out to the larger picture I looked at feasible ways these robots could be used throughout a cities' WDN, and created cost analyzes to compare futuristic methods of WDN management with current methods of district metered areas (DMAs). However, going from our current state of minimally instrumented pipes, to pipes with many valves to direct in-pipe inspection robots is a big leap, and thus I also created a method to help evaluate the cost trade-off of valve placement and the optimal spots for adding valves in the case where it was ideal to place valves on only some of the intersections of the WDN..