Fracta
Simplifying Pipe Maintenance and Repair with Machine Learning

Today, the United States is home to several hundred miles of drinking water pipeline systems. The chances for breaks and leakages of those pipelines are high, primarily due to aging infrastructure and wear and tear, among other reasons. These are conventionally dealt with by excavating the soil and physically evaluating the condition of water pipe systems. However, the time, cost, and labor in digging up and replacing them demand herculean efforts from utility companies.
Can this challenge be addressed more proactively?
Hideki Okada, CEO of Fracta, says, “Yes!” “We are here to solve the hardest riddle of water utilities with our affordable digital solution that accurately predicts infrastructure failures, thereby preventing unnecessary, expensive and long-haul pipeline inspections.”
Fracta is a California-based leading provider of AI-based software platforms that monitors and analyzes the break patterns, target leakages, and failure scenarios of pipe networks. Owned by Kurita Water Industries of Japan, the company helps utilities prioritize efficient water pipe replacements. Its cutting-edge technology helps them better understand the underground conditions and improve the overall reliability of water mains. “By leveraging our utility network database, we bring to the table more of a data-driven pipe condition assessment that goes beyond traditional pipe age and break history models, to identify failure risk of pipe segments” says Ted Sindabizera Ntwari, Product Manager of Fracta.
Fracta’s platform comprises condition assessment solutions that leverage machine learning(ML) and advanced analytics technologies: the likelihood of failure (LOF) and the consequence of failure (COF). These advanced tools calculate the correlation between historical breaks and local conditions of pipes. Fracta’s LOF prediction model lists out the individual risk factors of pipe segments and the probability of their failures in scores. Utilities can therefore verify planning and identify the pipes that need to be replaced, ensuring good pipes are left in the ground. Moreover, the COF data-driven model analyzes the financial implications of pipe breaks by creating their failure scenarios. It makes utilities manage direct and indirect costs to improve recovery planning.
Fracta’s BRE tool visualizes the overall business risk exposure using dynamic graphs and charts to connect the dots between the probability of break events and their financial effects. Further, the company’s Job Planner tool combines the LOF and COF results and creates optimized repair and replacement plans to restore the weakest parts of water networks. This assists utilities to be more proactive rather than reactive in optimizing the capital improvement project (CIP) planning.

We are here to solve the hardest riddle of water utilities with our affordable digital solution that accurately predicts infrastructure failures, thereby preventing unnecessary, expensive and long-haul pipeline inspections

On top of that, the company’s onboarding process takes only about four weeks which begins by collecting the basic pipe data and the break history of client utilities. Its environmental database then derives over one hundred and fifty parameters related to pipe failure. Going beyond just delivering solutions, the company also provides consulting services with their in-house GIS experts and data engineers. They use reconstruction algorithms to automate the restoration of high-quality pipe data in a couple of weeks. “Our core service is opted for the triple bottom line of People, Profit, and Planet on focusing beyond the revenue loss,” prides Ted.
Steering ahead with a focus on reducing the nonrevenue water (NRW) loss, Fracta’s success stories are worthy of note. Hideki recollects an instance where they helped the Greater Johnstown Water Authority reduce 20% of their NRW within a year. It was challenging for the client to identify the precise location of water loss which had been a greater risk for community safety and business operations. To this end, Fracta’s data engineers stepped in with their LOF maps in providing Johnstown a more granular and nuanced view of where water loss was likely coming from. They also leveraged ML algorithms and evaluated historic break events across various combinations of environmental variables at each pipe segment. Ultimately, Fracta had helped the utility manage the pipeline integrity and brought about a keen saving against NRW loss. “Our service came out as a positive uptake for Johnstown to better understand its network of assets and management in moving forward,” adds Hideki.
Similarly, the public-owned utilities such as San Francisco Public Utilities Commission (SFPUC) and East Bay Municipal Utility District (EBMUD) have supplemented their current pipe replacement methodology with Fracta’s LOF software. The clients benefited from determining the right pipes to replace and then protecting natural resources through reduced water loss. In line with these, Topeka of Kansas has also used Fracta’s BRE tool for CIP planning that helped the utility save nearly $80-100k per year with a strong ROI.
Scripting many such success stories to its name, Fracta has a well-planned blueprint for expanding to Europe as well as promoting its platform among Spain, Italy, Germany, and France. Across Japan, the company envisions coming up with a product that better visualizes their pipe network conditions. Further, it plans to add new features in terms of leak detection and infrastructure bills that will provide funding for utilities to replace lead service pipes. “What’s more crucial is, we play an active role in improving our product more in terms of AI, ML, and Data Science to solve a global issue,” concludes Ted.