Case Study

Recovering Revenue Using Data-Driven Meter Replacement Program in Indiana

By addressing ongoing apparent water loss and its associated economic, social, and environmental costs, the City of Fort Wayne identifies more than $250,000 in recoverable revenue.

AUTHOR

Related Topics

Water Network Management

Smart Metering

Apparent Water Loss

Non-Revenue Water

In 2019, the City of Fort Wayne’s Water Division decided to perform a check on its meter fleet. Specifically, the utility wanted to evaluate the effectiveness of their existing meter replacement program – which relied on random testing and factor-based criteria such as meter age and throughput to prioritize replacement.

THE CHALLENGE

They suspected that taking an innovative approach involving machine learning and analytics might help them more accurately identify underperforming meters and locate water lost as a result. This could be a big benefit to the utility since apparent water loss can account for about two percent of a utility’s top-line revenue.

THE SOLUTION

Xylem deployed their Revenue Locator solution, a cloud-based, SaaS subscription platform. The solution leverages machine learning and analytics to provide utilities with data-driven operational guidance to identify metering inaccuracies, prioritize field maintenance activities, implement efficient meter testing and replacement programs and maximize revenue recovery.

Over the course of the program, the Revenue Locator solution analyzed all of Fort Wayne’s 2-inch meters. The data that was gathered successfully demonstrated which meters had inaccuracies and provided the quantifiable impact of how much volume or revenue would be lost if the meter were not repaired or replaced.

Before working with Xylem, we were managing our meter fleet and making inspection and replacement decisions without true visibility

Ben Groeneweg, Utility Asset Management and Sustainability Manager

RESULTS

  • Identified recoverable revenue of $264,871 over two years
  • Provided an 8x improvement in meter inaccuracy identification over traditional methods
  • Pay-back period of just over one year
Featured SWAN Members: