Machine Learning - the future of data-driven decision validation for Infrastructure

100,000 miles of pipe + 250,000 breaks = 1 World Changing Model

machine learning

Using Technology To Solve A Global Issue

Fracta is a cutting edge condition assessment solution that uses Machine Learning to assess the condition and risk of drinking water distribution mains.

The Fracta solution shifts asset operation and management from reaction to prevention. It helps avoid disruptive water main breaks, lower non-revenue water (NRW), better target leak detection, valve maintenance efforts, and educate key stakeholders on the true cost and risk of their aging water main infrastructure.

Our Process

Comparing Traditional Methods To Machine Learning

Existing methods of forecasting pipe prioritization are effective, but fail to take into account many recent advancements in artificial intelligence.

title

Data Quality

Environmental Considerations

Model Intelligence

Accuracy

Current Practices

Individual Variables

Limited

Non-Learning Model

Local Utility Network

Fracta’s Machine Learning

Multi-channel Variables

Extensive

Machine Learning

Global Utility Network

Technology

Our Approach To Training Data

A model is only as good as the data used to train it.

Model predictions get stronger as more data is processed. Fracta uses nation-wide geospatial information as well as tens of thousands of miles of water mains and historical break events to calculate LOF and COF to determine Total Risk. As time goes by, more data is collected and the accuracy of LOF and COF improves. By rehabilitating or replacing the worst pipe, utilities should see their break rates decrease over time.

Pipe and Break Data

Environmental data

Population data

Soil condition

Weather data

100+ more

Cloud-based Machine Learning

The Fracta solution is a cloud-based software application that can be connected to important software applications used by water utilities, including Enterprise Asset Management (EAM), Computerized Maintenance Management Systems (CMMS) and Hydraulic Modeling.

Fracta can complete Likelihood of Failure (LOF), Consequence of Failure (COF) and Business Risk Exposure (BRE) assessments for an entire water main distribution system in 1-2 weeks.

The results are delivered in a Software as a Solution (SaaS) system and visualized using dynamic graphs and charts. New data can be uploaded and modeled several times per year, enabling a dynamic, near real-time assessment of the system.

Enabling The Shift From Reactive To Proactive

Utilities and their consulting engineers perform desktop and physical condition assessments of buried water mains as part of their infrastructure asset management. Physical condition assessments tend to be slow, expensive and labor intensive. Multiple physical measurements are required for correlation and confirmation. Beyond the tested pipe, it’s difficult to extrapolate this data for system-wide recommendations.

Desktop condition assessments primarily rely on age, break history and the experience of the engineer evaluating data. The results are often based on best professional judgment, which is a generally accepted engineering practice, but it can be subjective. It’s also been shown that age-based analyses are significantly less accurate, more time intensive and more expensive.

press & media