THE FIVE STEPS TO HELP MINIMIZE YOUR WATER MAIN PROBLEMS.

Fracta follows a five-step process to assess the condition and risk of water distribution mains beginning with data review for quality, completeness and readiness for machine learning.

Fracta uses a patent-pending data processing tool to assess, clean and normalize data. The tool can identify missing water main parameters such as material and install year, rationalize disparate data sets coming from multiple sources and recognize incorrect data values. The condition and risk assessment process begins after the data is cleaned.

FIVE-STEP PROCESS:
DETERMINING WHICH PIPES TO REPLACE AND WHICH TO SAVE
1. WRANGLE, import and geolayer water main, geographic and environmental data
Fracta wrangles the utility data, which is a data science term meaning the data is transformed into the data type and form that the machine-learning model understands, and then imports it into the Fracta system. Approximately 1,000 geographical and environmental variables, potentially impacting the condition of a water main, are layered over the utility service area.
2. predict likelihood of failure using machine learning
Fracta trains and validates the Machine Learning algorithm that then calculates the correlation between the parameters and historical failures and builds a model of the system. The output of the Machine Learning analysis is an accurate prediction for Likelihood of Failure (LOF) for all water main segments in the distribution system.
3. CALCULATE the financial consequence of failure
Once LOF is determined, Fracta calculates the Consequence of Failure (COF) value using parameters and values derived from basic water main segment data, as well as environmental variables. The values calculated for each water main segment are then categorized and transformed into dollar amounts based on individual utility cost structures. A translation table is needed to transform the raw analysis figures into dollars. Fracta’s research will guide some default values for all conversion, but translation values can be edited based on individual utility needs.
4. Determine the business risk exposure
With the predicted LOF and calculated COF, Business Risk Exposure (BRE) is determined by multiplying the LOF percentage by the COF dollar amount. BRE is calculated for each water main segment and aggregated for a total amount for the entire distribution system.
5. VISUALIZE and apply results to support asset management decision making
Armed with the LOF, COF and BRE, utilities can now focus on asset management and risk mitigation strategies. Engineers, planners and management can now make quick, accurate and affordable decisions regarding:

– Rehabilitation and replacement
– Changes to operating conditions
– Further condition assessment
– Rate increases, cases and studies
– Leak detection targeting
– Valve maintenance and exercising
– Educating boards and commissions

"We gather all the environmental data, mix it up with the utility data, run our algorithms to provide the likelihood of failure for each section of the pipe and map it out using the web.”

– Takashi Kato, Fracta CEO and Cofounder