how to detect ice to water formation

[zambre]

Ice to water sensor

(how to detect if water is in liquid form or solid ice form)

I am working on a project where I need to detect if the water in a tank has frozen and it is ice. Many times the ice forms on the surface but there is water under the ice slab.

initially, i thought if using multiple temperature sensors at various depth. but i am not sure if any off the shelf temp sensor can be useful. secondly zero degree temp does not confirm if the water is in liquid form or solid ice form.

I am not sure using a temperature sensor will give me accurate data whether the water is in liquid form or in ice form. I will appreciate if you can give me suggestions to detect the state of water, whether it is ice or liquid water..

Thanks

[adafruit_support_bill]

A temperature sensor would not be conclusive. Water often exists in a 'supercooled' state before it crystalizes.

You could take advantage of the fact that sound travels faster through ice than through water and use ultrasonic transducers to measure it.
The refractive index changes also as water freezes, but I can't think of a practical way to measure that in-place.

[zener]

As you stated it is not uniform, so the question is: Ice where? If you just want to know if your sensor is in ice or water then I would think of something that tries to make a motion of some kind and detects if it can't. And it can't make too much heat or that could throw off the test. This approach might go along pretty well with Bill's ultrasonic idea. But maybe look at the amplitude received vs the speed. You could try both.

[adafruit_support_mike]

You could probably do something with a series of capacitive sensors.

The dielectric constant of liquid water is about 80 times higher than air, mostly because water molecules are electrical dipoles.. one side has more positive charge and the other side has more negative charge. When you apply an external electric field, all the dipoles swing around to line up with the field, and the energy necessary to make that happen (for a given amount of water) is the dielectric constant.

In ice, the water molecules are locked in a crystal pattern that makes it harder for them to swing in line with an external electric field. That reduces the amount of energy the ice can store, so the dielectric constant drops from 80 to about 20 or 30.

If you put a series of capacitive sensors at different levels in the tank, you should be able to get a self-calibrating set of level-and-material-phase readings. Sensors with air between the plates will see a capacitor value of C, sensors with ice between the plates will see a capacitor of strength 20C to 30C, and sensors with liquid water between the plates will see a capacitor value of 80C. The exact value of C doesn't matter, only the ratios between the air value and the liquid/ice values.