Public water supply and sanitation systems are fundamental to normal functioning of a society. Uninterrupted access to water is an integral part of our lives, and willingly pay for it. In the days of growing interest in AI for predictive maintenance, we have a chance to significantly reduce repair costs by preventing breakdowns in the first place.
Our Client, a public sector operator with many years of experience in the area, came up with a brilliant idea to assess condition of pipes by analyzing sound using IoT transmitters. In a nutshell there are certain sounds that allow you to tell a ‘healthy’ pipes from those that may leak very soon. Long hours of different sounds were collected to feed the ML sound recognition algorithm.
When the solution is well polished, breakdowns will no longer be accidental. But even now a PoC provides quite accurate predictions which mean that we are moving into the right direction.
April 8, 2020