Water level detection using non contact sonic sensors

Question

I have been working on a project, where I need to find the water level of underground borewells with a diameter of around 6-10 inches and a depth of up to 100 meters. I have been researching a lot about the methods that I can use for this and after ruling out the contact method of a using a probe I settled for a non contact sonic method where a sound wave would be sent and the time of flight of echo would be used to calculate the distance. Now I was thinking about making a circuit out for it and had a couple of questions in mind:-

1. Which sonic transmitter can I use i.e. a piezo or a simple loudspeaker version?
2. What circuitry would be exactly required to work this out? As far as I know I would produce a PWM of varying frequency until I get an optimum value, fed from a controller and amplify it to feed it to the transmitter.
3. For the receiving sensor I think the condenser mic would serve the purpose, correct me if I am wrong and any suggestions would be appreciated.
4. Where can I find these sensors in order to use them for my project? My local vendors seem to be unaware about any of these.

Here is the link to the discussion a started for knowing about the frequency and spreding angle concepts related to sonic beams.

https://physics.stackexchange.com/questions/408894/sound-ultrasonic-waves-distance-travelled-and-spreading-angle

Answer 1

The easiest way to get started is to cobble a proof of concept (PoC) system out of consumer audio speakers, amplifiers, microphones.

To rephrase what you want to do: send a sound pulse from the top of long and narrow well, the pulse would travel down the well, reflect from the water level, calculate the distance to water from the round trip time of flight.

It's challenging - if at all possible - to make this work in a thin and narrow wells. The challenge is that the sought reflection from the surface of the water will be neither first nor the last reflection you receive. The sound will not travel in a narrow beam like a laser. It will propagate as a cone. At some point, this cone will become wider than the diameter of the well. Some sound is going to reflect back from the walls, and that's going to be the first reflection you receive. You will continue to receive reflections. At some point, you will receive the reflection from the water at the bottom of the well. The reflections will not stop there. You will continue to receive multipath reflections.

Record echoes from well with known diameter and water level. Maybe there will be a consistent feature in the echos which you can clearly identify as the water surface. Maybe you'll see more if get several echoes at several frequencies from the same well.

p.s. If the wells are always straight, a laser rangefinder might work.

Answer 2

I see a lot of problems with this concept:

• first how smooth are walls? If they aren't you can get many unwanted echos
• attenuation will be main problem, especially with reflection as main signal
• speed of sound will change significantly with temperature and humidity
• it will be loud

Now to solutions/your questions:

• I would use loudspeaker (easier to get it working)
• for signal use low frequency sine sweep, for example 1-5 kHz (and use cross correlation for detection)
• I would use pc with sound card and amplifier
• if you use piezo you can also use him as microphone, otherwise condenser mic is good
• for loudspeaker, microphone and amplifier I would look in music shop or similar

P.S. I don't know your exact situation, but I would try to solve this problem by dropping pressure sensor to bottom of bore and read absolute pressure there.

Answer 3

A few months late, but if you are using sound waves in a narrow well you could treat it either as a waveguide and use either time-domain reflectometry or as a resonant pipe and use the resonant frequencies to estimate the distance.

Given your pipe diameter sounds of ~1.5kHz and higher will propagate down the bore, much higher than this and you will start dealing with multiple modes of propagation, much lower than this and you have a resonant cavity instead of a waveguide.

I would put a water-resistant speaker sealed against the pipe at the top and a water-resistant microphone further down the bore at a known distance, so that you can better estimate and correct for sound propagation speed. Given changing temperatures and humidity levels these will not be constants. Remember that a speaker emits sound from both sides, so it either has to be in an enclosed rigid box or its top surface has to be in open air while the driving surface seals the well.

• time-domain reflectometry.

Emit a single pulse (e.g., one or a few cycles of a 2kHz sine wave) and measure the received signal. This would consist of several attenuated repetitions of the transmitted signal. An auto-correlation would give you the time delay between the multiple echoes, the delay between the transmitted signal and the first signal will provide you with the needed information to calculate sound speed.

• resonant frequencies

Emit a continuous sinusoidal frequency that is slowly swept through a few octaves (e.g, 100Hz-1.6kHz). Note that in this case the intent is to produce a standing pressure wave; so, just like in an organ, the wavelength is given by the length of the pipe.

The pipe will act as a resonator that will produce a comb of response peaks at frequencies directly proportional to the wavelength of the signal supported by the pipe. Although some resonant frequencies will be the same, this comb of frequencies will be uniquely determined for any length of pipe, and its maximum common denominator will give the actual pipe length.

As before, the delay between the excitation and the signal with respect to the distance between the microphone and speaker would allow you to calibrate the measurement.