Using Predictive Modeling To Identify Lead Service Lines

Published on by in Case Studies

Using Predictive Modeling To Identify Lead Service Lines

When it comes to the Lead and Copper Rule Revisions (LCRR), the number one thing causing uncertainty for most utilities is creating a service line inventory. Water systems are searching for creative solutions and efficiencies when it comes to their inventory in order to meet compliance requirements.

LCRR’s inventory requirement
LCRR requires all water systems to develop an initial service line inventory by October 2024 or demonstrate the absence of Lead Service Lines. Lead Service Line Replacement Plans must then be created as a result of your inventory.

Inventories must also be made publicly available and must identify service line material for both private-owned and utility-owned portions as:

Historically, most utilities have only been responsible for the public or utility-owned portion of the service line, but LCRR is adding private site verification to utilities’ plates as well.

Under LCRR, utilities may use predictive models to identify service line materials more systematically: “Such predictive models could also inform water systems in how they can approach LSLR in a more efficient manner. The EPA encourages but does not require this practice as it allows consumers with lead status unknown service lines to be informed sooner about their service line material.“

What is Predictive Modeling
Predictive Modeling technology works by analyzing historical and current data to generate a model that helps predict certain information, like the material type of unknown service lines.

With Predictive Modeling from 120Water, agencies and water systems employ existing records to inform our advanced data science. We use machine learning to identify the most probable locations of lead service lines, allowing you to generate your inventory more cost-effectively and plan for LSL replacements without arbitrarily potholing and excavating.

How 120Water Applies Predictive Modeling with the Lead Service Line Probability Finder
The 120Water Predictive Modeling process helps customers develop service line inventories through an integrated approach of full-service management, easy-to-use software, and advanced machine learning. The entire process is handled by the 120Water team, and all the results and data are available in accessible dashboards throughout the process. 

1. Develop Preliminary Inventory

The 120Water team will identify all sources of potential records including:

All the data will be assessed through a QA/QC process and all relevant data is loaded into the 120Water software.

2. Run the Model

If there is no known lead (but the customer knows they have lead somewhere in their system), the Model provides an initial inspection list of most likely spots to have lead.

If there is confirmed lead, the Model is run to provide an initial predictive set.

3. Deliver Results of Initial Model Run

Results are broken into neighborhood clusters with different degrees of confidence by neighborhood inclusive of overall hit rate expectation as well as overall coverage.

120Water delivers a verification list for neighborhoods with a higher hit rate as well as a list of sites for inspection that will be used to further enhance the model’s accuracy.

Attached link

https://120water.com/predictive-modeling-to-identify-service-lines

Taxonomy