AWWA WSC61390

This slide presentation outlines information on using basic water quality sensors to spot deviations from normal in a distribution system, and then mathematically analyze the deviation signals to: trigger when something harmful is present; detect what agent is present; and, quantify the amount in the water. Topics covered include: sensor set selected; "radar" plot of agent deviations; detection algorithm added to trigger; beaker testing; single pass flow testing; trigger signals from nicotine and cyanide; large volume single pass flow testing; problems with initial results; ETV study; loop testing at ECBC; real world deployment; east coast location near water storage tank alarm triggered; plant event defined; learning method for plant events; reduction of unknown alarms via Hach Learning Algorithm; results of learning at a very noisy location; enhanced water distribution system; PipeSonde in-pipe probe; and, ongoing work.