Insights

How machine learning changed the game for pipe condition assessments

Every day, water pipe failures lead to damage, disruption, and water loss. It’s a problem that, unfortunately for the water utility industry, shows no signs of stopping. Even for organizations with extensive pipe networks that are usually quick to react to failures, the effects of unexpected pipe breaks – water loss, damage, service interruption, repair costs , and a water utility’s reputation – can be significant and longer-lasting.

Up until recent years, responses to pipe failures have been reactive, often involving repair crews working late in the night and struggling with flooding and inclement weather, not to mention utilities paying 10X in unexpected repair costs vs. regularly scheduled replacement. For the water industry to truly evolve, these reactive solutions are no longer good enough.

No more guesswork

It was clear that a more nuanced approach to identifying the likelihood of failure (LoF) was needed. And just like that, the asset management world was brought into the age of machine learning (ML).

ML incorporates data-driven models based on key inputs, including past failure data, to help organizations make decisions about at-risk and healthy pipes. It’s exactly what Arcadis Gen’s Water (AI) Pipe Predictor (WAIPP) app is based on to help predict the LoF for existing pipe networks so water utilities can avoid many, if not most, catastrophic breaks and premature pipe replacements. WAIPP harnesses the power of data analytics to help prioritize pipe replacement by evaluating prior breaks, material, age and other key factors.

ML and AI have proved to be more accurate than traditional human analysis in pinpointing which pipes to leave in place and which pipes demand attention. Not only does this make AppliedInsight’s WAIPP a highly effective alternative to a data scientist, it’s a more cost-effective solution too. Our app is perfect for organizations who find themselves with limited capital to spend on ageing water infrastructure, as they get all of the expertise of a professional data scientist, without the costly consultancy bill.

What’s more, ML software can actually “learn” from current and past data in order to predict future events. Our app finds patterns in prior pipe failures to accurately predict the LoF for all pipes in the near future. This means the more data you input, the “smarter” your insights will become over time.

Conclusion

There’s never been a better time to consider an ML and AI technology such as Arcadis Gen’s the Water (AI) Pipe Predictor for your water organization. WAIPP’s algorithms will not only help you to prioritize decisions about water pipes but also help you to stay within your infrastructure budget too.

Now you and your organization can gather insights needed for robust, data-driven investment decisions. You can also take advantage of more than 20 years of artificial intelligence and machine learning expertise, rolled into one single app that has been specifically designed for the water sector.

To learn more you can request a demo of WAIPP from AppliedInsight today.