Golden State Water Company (Golden State Water) is an investor-owned regulated utility operating 41 water systems, serving over 1 million customers across 81 communities in California, and managing approximately 2,900 miles of water main. The utility has relied for many years on their in-house engineering planning staff to develop pipe replacement projects using components such as GIS, break and leak history, hydraulic and statistical models, and risk analysis.
THE CHALLENGE
Golden State Water was challenged with a time-consuming process for creating pipe replacement construction packages. The output from the “traditional” risk model required a complex manual review process for grouping and prioritisation. Their continuous pursuit of best practices led to a desire for an even more streamlined, data-driven methodology for identifying high-risk pipe segments for replacement – one that strikes the best balance among multiple competing objectives. The new solution needed to ensure that the most efficient replacement projects are selected for each community to ensure premium water quality and reliability.
THE SOLUTION
Golden State Water partnered with Trinnex to deliver Data Science-as-a-Service (DSaaS) for this project, including developing a risk prioritisation model and an optimised project packaging tool. The tool uses advanced analytics to help water utility managers convert multiple existing risk assessments to streamlined, compact, construction-ready projects.
- GIS-based Risk Modeling: Trinnex updated the risk approach to a more robust model that analyses the Likelihood of Failure (LoF) using factors like age, material, and break history, and the Consequence of Failure (CoF) based on parameters such as pipe diameter and flow rates.
- Spatial Grouping/Clustering: A spatial clustering algorithm was applied to high-risk pipes to group them into clusters, considering criteria like size differential, connectivity, compactness, and total length.
- Multi-Objective Optimisation (MOO): This automated, algorithm-based search combines clusters to find the optimal solution to multiple project objectives, allowing Golden State Water to customise the solution by establishing benefits (e.g., replace pipes that have high LoF/CoF scores to reduce main breaks) and penalties (e.g., construction impacts such as traffic and community distruption).
RESULTS
Trinnex delivered a repeatable and defensible process for selecting replacement packages to ensure that utility goals, including budget and customer community needs, are met.
- The key deliverables included the comprehensive risk model and GIS layers, as well as ranked pipe replacement clusters.
- The solution revealed new ways of considering replacements and increased construction efficiency (buildability).
– Mark Insco, Engineering Planning Manager at GWSCThe [DSaaS solution] optimisation often uncovers strong candidate pipes – or combinations – we had not previously prioritised, adding real value to our pipe replacement planning process.
Future Steps
- Provide comprehensive training and a final report, equipping staff to independently manage and adapt the prioritisation framework moving forward.
- Enhance data with additional inputs for future analyses to help reveal patterns of pipe failures.
- Explore potential technology integration with work management and metering systems.
