Making Informed Decisions about Customer Complaint Surveillance Alarm Rates Using the Alarm Estimation Tool

摘要

After September 11, 2001, the U.S. Environmental Protection Agency (EPA) was charged with improvingrnthe security of the Nation’s water and wastewater infrastructure. Legislation, Presidential Directives andrnfunding have spurred utility vulnerability assessments, research and exercises that have furthered thisrncountry’s readiness to prevent and respond to water security concerns. This paper will focus on CustomerrnComplaint Surveillance (CCS) as part of a comprehensive contamination warning system developed byrnthe EPA’s Water Security initiative (WSi). The paper will highlight an Alarm Estimation Tool (AET)rndeveloped for a CCS component at any utility, based on the initial CCS system at the Greater CincinnatirnWater Works (GCWW), which was the initial pilot utility under a Cooperative Research andrnDevelopment Agreement between EPA and the City of Cincinnati.rnAdvanced statistical analysis of historical data, to determine appropriate thresholds, at which to notifyrnutility staff of unusual customer call types or volumes, may seem intimidating for utility personnel. EPArnhas developed a simple tool, which allows a non-statistician to experiment with thresholds to experience arnrange of alarm rates. An alarm threshold having an accompanying frequency or rate for alarms to berninvestigated over time, requires a utility to balance the level of effort (LOE) in investigating false alarmsrnwith the probability of not investigating a rare, occurring true alarm.rnUsing scan statistics, macros are used in a spreadsheet allowing a user to input a column of date-timerndata, and then set thresholds for various time analyses. The result is a graph showing the number of alarmrnoccurrences in a data set, based on the user-inputted thresholds. The thresholds can be changed to viewrnother alarm occurrences in the data set. This paper explains and demonstrates the AET.