i’ve following the edge impulse for over 2 years now, and finally decided to take the plunge as i think i have a case that might benefit from tinyML, so pelase allow me to explain:
ESP32 arduino framework
Hall effect (Water meter Pulse counter)
so basically I’m hooked up to my water meter through it’s pulse counter using interrupts which reports the pulses as water flows through, where each pulse for this meter is 1pulse/m3. so far so good…i timestamp the pulse counter and report back to my dashboard i can see how many m3 of water have flown (or consumed) based on time…
now the idea is i would like to know how do i go about training a model that can learn to detect water pipe leaks meaning when there is a leak i would assume the model would noticed that a constant rate is being flown that doesn’t change over a period of time (defined or learned by the model). which would mean a leak… simila to NILM algorithm that is used for Power.
can someone please help me understand how to plot this or maybe point me to an example i can learn from or any readings online.