The City of Huber Heights, Ohio, operates a water distribution system of 205 miles of predominantly cast iron and ductile iron pipe. While their system was already experiencing a relatively high break rate before 2019, the break rate from 2019 through 2021 doubled the prior 10 years. The City needed to unravel the mystery of what caused the spike in breaks.
Utilizing the City’s GIS pipe attribute data and break database, and leading-edge failure forecasting software that predicts break rates, the City tested their theories about possible causes of the spike in breaksThe City had commissioned three new booster stations, decommissioned one of its two water treatment plants, and commissioned a new softening process at its primary plant, all of which impacted flows, pressures, and water quality over the last five years. They obtained answers using robust analytical tools that allowed for significant interaction.
After completing the analytics, a proactive watermain replacement strategy was recommended to significantly reduce breaks and the associated impacts, providing customers with a more reliable drinking water supply. Through the customized analysis in the break prediction software and GIS, recommendations for operating the distribution system and treatment plant to better control finished water quality and pressures were possible, and the break rate is already declining without any pipe replacement.
To build the replacement plan, the City chose not to rely on industry standard values for predicting pipe life because those values can vary widely and lead to wildly inaccurate predictions of when pipe should be replaced. Instead, the City leveraged its comprehensive pipe attribute data set and empirical break database to identify its riskiest pipes.
The City’s GIS contained over 99 percent population of attribute data for installation date, diameter, and material, and working with the City, estimates of the remaining attributes were conducted. The 12 years of break in the database were all associated with a corresponding pipe. A significant effort was undertaken to ensure the accuracy of break data, with the City performing a review of watermain break work orders for the full decade.
Asset and break data were imported into the online break-prediction platform, specifically designed to make pipe-by-pipe failure predictions leveraging machine learning algorithms. Before performing the analyses, the software guides the user through a series of quality control steps to verify and improve data quality and assure the most accurate predictions. Faulty data is flagged, and the user can either correct the data quality issues or remove them from the analysis.
Predicted break information is coupled with each pipe’s consequence of failure data generated based on proximity to roads, water, structures, and service to critical customers. With a listing of pipes prioritized by risk, the City evaluated the resultant break rate and risk associated with various proactive watermain replacement investment levels and zeroed in on an affordable annual expenditure that will stabilize the break rate.