This system detects leaks by listening to the water flowing through the pipes

This system detects leaks by listening to the water flowing through the pipes

Arduino Team — July 27, 2022

Damaged and leaky pipes are not only a nuisance to clean up, but they can also create major inefficiencies in water distribution systems, resulting in loss of both water itself- even and electricity needed to sanitize and pump this. Over the past decade, water pipe detection systems have been upgraded to include state-of-the-art sensors capable of pinpointing the exact location of a leak. Due to their high price, Manivannan Sivan have designed their own leak detection system which can be produced at a much lower cost.

Sivan's project involves placing two microphones next to a pipe and reading the acoustic signatures they pick up. For this task, he chose a single Arduino Portenta H7 and an accompanying Vision Shield because of its pair of built-in mics and fast processor. He then took samples for no water flow, water flow without leakage and water flow with leakage. The resulting machine learning model achieved an accuracy of 99.1% and a loss of only 0.02.

After deploying the model on its board and placing it near a pipe, the Portenta now had the ability to identify when the pipe began to leak - and potentially warn someone one thanks to its wireless connectivity, should Sivan decide to have it add that feature.

For more details on this project, read its description here on the Edge Impulse blog.

This system detects leaks by listening to the water flowing through the pipes
This system detects leaks by listening to the water flowing through the pipes

Arduino Team — July 27, 2022

Damaged and leaky pipes are not only a nuisance to clean up, but they can also create major inefficiencies in water distribution systems, resulting in loss of both water itself- even and electricity needed to sanitize and pump this. Over the past decade, water pipe detection systems have been upgraded to include state-of-the-art sensors capable of pinpointing the exact location of a leak. Due to their high price, Manivannan Sivan have designed their own leak detection system which can be produced at a much lower cost.

Sivan's project involves placing two microphones next to a pipe and reading the acoustic signatures they pick up. For this task, he chose a single Arduino Portenta H7 and an accompanying Vision Shield because of its pair of built-in mics and fast processor. He then took samples for no water flow, water flow without leakage and water flow with leakage. The resulting machine learning model achieved an accuracy of 99.1% and a loss of only 0.02.

After deploying the model on its board and placing it near a pipe, the Portenta now had the ability to identify when the pipe began to leak - and potentially warn someone one thanks to its wireless connectivity, should Sivan decide to have it add that feature.

For more details on this project, read its description here on the Edge Impulse blog.

What's Your Reaction?

like

dislike

love

funny

angry

sad

wow