Add a temperature sensor (DS18B20) to the pipe and measure constantly at 1 sample per second and keep an eye on the delta T.
If water starts to flow you will possibly see a different temperature gradient than when there is no flow. Probably temperature drops more suddenly but it really depends on the local situation.
You can prototype this for not too much money, only investment is a sensor.
First step is to tell us what is in the water line when there is no water? If the water starts flowing, what is being displaced? Is the answer air? If so, measure the change in air pressure in the line. Where does the air go when water flows? There must be a vent that let air out, but keeps water in. How does that work? Does air come back in when water flow stops?
@gilshultz pressure range is not known, uniform, or steady. I remember in my younger days water wouldn't climb to the roof without a pump. Now Pump is usually only needed in rare days, supply pressure is such that the water climbs to the roof itself. But of course if we decide to build another storey and place the tank higher, pump might be needed regularly again.
Maybe it is displacing air near the tank (because when the flow stops, the water drains out of the horizontal pipe into the tank), but is it displacing air anywhere else?
Regardless, you could put a microphone at the top of the tank and have an Arduino listen for the noise of water splashing into the tank.
Or just monitor the level of water in the tank (e.g., using a waterproof ultrasonic sensor at the top of the tank) and note when the level increases.
@Paul_KD7HB I am concerned with the overhead tank supply. When the supply comes in municipal line, it flows into our private line where it goes through a water meter and through a pump before rising to the overhead tank.
@paulpaulson that is again a contact solution, which would require plumbing work. essentially that would involve installing a tap output at ground level and in place of a tap attaching a pressure sensor. I would like to avoid involving a plumber if I have to...and if I have to, then perhaps a turbine sensor would be more useful. But again, trying to do it in a no contact way.
@DaveEvans the open end inside the tank, when air is displaced, would this open-end have any sound when water is there on the ground level but not in enough pressure to rise above? Edit : And if a microphone is put at its mouth, I guess it would obstruct the flow of water when eventually flow happens, plus would get wet and go bad.
I guess there is no definitive way except a contact sensor? I thought vibration sensor but a lot of things can cause vibrations except water. Temperature sensor also would fail with PVC, or if the pipe is just wet with water.
But what makes me even more curious, sometimes, in the height of summers when the demand of water is so high that water does not reach even our domestic line, we have to do hit and try to see if the pump can pull anything from the mains line. Is it too ambitious to hope for a way to detect water in the mains line from something like 20 meters away? Any possibility?