AI Pressure-Transient Detection for Water Networks

The problem

Pressure transients, better known as water hammer, are short, sharp pressure spikes that travel through water distribution networks. They are easy to miss with conventional monitoring, yet repeated spikes fatigue pipes and trigger failures. Utilities could not see them, so they could not manage them.

What we built

• An AWS-native data platform that ingests high-frequency (>100 Hz) pressure feeds and serves transient-risk views within about 15 minutes, giving water utilities visibility they previously did not have.
• Detection and clustering algorithms with >97% reliability that detect transient events and cluster them, so engineers can tie invisible spikes to asset failure and trace them to their sources.

Impact

• Worked with Greater Western Water to identify and mitigate transient sources, unlocking roughly $3 million in avoided reactive work.
• Extends asset life by an estimated 5–10 years at transient hot spots.
• The results underpinned SpiralData’s $1.5 million scale-up capital raise.

Learn more at transientdetect.io.