Estimating flow rate from noise

jarvist

New member
We're dusting off our plans to build some dataloggers for installation over a year in an alpine cave. It's easy enough to buy sensors to do temperature, pressure and humidity (though it might saturate), but one thing we've considered trying to do is estimate the flow rate of the cave by recording the level of noise due to water falling down a pitch.

Does anyone know of any work done to relate the dB of the white (or pink?) noise of a streamway to flow rates?
 

Muddy Funkster

New member
I'm no scientist, but there's lots of reasons that noise would increase on one pitch when compared to another that are not exclusively to do with flow rate / volume of water. Passage morphology in the pitch. Accoustics created by the space. Roughness of rock surface in the pitch. Depth of water that the falling water lands in. etc...
It may be you can get some roughly approximate readings by doing something like this but the sheer number of variables in a natural environment like a cave pitch make it hard to fully trust in the results produced.
That being said I often do use ambient sound as a gauge of how wet an approaching crawl is going to be. On the last trip down Aggy's main stream, we sat by the second choke sign above the wet crawl and could not hear the water 10 feet below us. Normally it's very noticeable, and sure enough we didn't get all that wet going through. I would have thought that measuring flow rate based on sound level in a phreatic passage would yield more accurate results. As I said I'm not from a scientific background but it does seem that there are not as much entropy stuff going on when compared to falling water? What do others who know more about it think?
Matt
 

Bob Smith

Member
I suppose you might be able to guess that for a particular pitch flow on day X was greater/lesser than day Y, but comparing different locations would be tricky.

Interesting thought though.

 

jarvist

New member
Bob Smith said:
I suppose you might be able to guess that for a particular pitch flow on day X was greater/lesser than day Y, but comparing different locations would be tricky.

Absolutely - this is what we're thinking, so we can understand the temporal arrival of flood pulses with surface rain data etc. If we had multiple ones logging regularly in the same cave, you should even be able to see the wave-front of water passing through the cave. During the winter and thaw it would be interesting to see how the temperature changes with passage of snow melt / warmer spring rain etc.

I would suspect that as long as nothing v. strange is occurring on a given pitch (i.e. when flow is high enough it starts hitting the wall askew rather than tumbling down the pitch), higher flow will result in higher noise levels.

I also suspect that as flow is directly proportional to the amount of power being converted at the pitch from potential and kinetic energy of the water into thermal energy of the water and noise, that within certain limits the flow will be proportional to noise.

But I'm sure someone must have looked at this properly? I've had a bit of search and found that acoustic rain gauges are extensively used at sea, where you can do all kinds of clever things to infer the droplet size from the frequency of the noises produced!
 

ian.p

Active member
If you can calibrate the noise levels with observed flow then that might give you a way of comparing flow at multiple sites.

there are a couple of potential problems in the case of the sea the rain droplets are hitting a relativly flat open surface the acoustics of a cave environment are going to be much more complex and i doubt you will get a linear trend between noise and flow rate so extapolating the data beyond the range which you can calibrate the instruments to is unlikley to be reliable...
 

graham

New member
ian.p said:
If you can calibrate the noise levels with observed flow then that might give you a way of comparing flow at multiple sites.

there are a couple of potential problems in the case of the sea the rain droplets are hitting a relativly flat open surface the acoustics of a cave environment are going to be much more complex and i doubt you will get a linear trend between noise and flow rate so extapolating the data beyond the range which you can calibrate the instruments to is unlikley to be reliable...

I wrote a reply earlier which promptly vanished, but made essentially the same points:

Firstly it is unlikely that the relationship between noise and flow volume will be linear.

Secondly, the complexities of actually calibrating this will be prohibitively difficult.
 

TheBitterEnd

Well-known member
Would it be better to measure the water depth at the pitch head (e.g. from an ultrasonic on the roof/wall) and use weir flow calcs to estimate flow? If circumstances allow it may be possible to install a V-notch or similar
 
re measuring water depth. There is a report of work done at Dan-yr-Ogof using a datalogger by Stuart France in Chelsea SS Newsletter Volume 53 No 5/6 May/June 2011.
 

jarvist

New member
Thankyou for all of your comments.
Using noise as an alternative to ultrasonics is what I'm thinking of as I suspect passive noise measurement will be a lot lower power to operate (we'd want this to log over a full year with no intervention), and it means we can put a sealed data logger in a safe phreatic side passage, rather than risk it getting swept away at a constrained pitch head (the only time we ever seem to see streams really). Installation would also be a lot easier, as the logger(s) would just need stashing somewhere safe rather than belay points + attachments.

The pitch I'm mainly thinking about (and alpine pitches in general) are so tall relative to flow rates that water falls more or less like rain, and lands on broken boulders, so if anything the sound might be more easy to analyse than droplets hitting the ocean surface.

I think a key thing to remember with science, is that if you come back with the only data, then you have the best data in the world  :)

Does anyone have any good links / recommendations for DIY anemometers / low wind speed / cave draft measurements? There's some technique with a heated wire in a cylinder isn't there? Might be beyond our power budget though.
 

Chocolate fireguard

Active member
This all sounds really interesting. Dynamic microphones are cheap enough on eBay. Why not get one and try it in Britain? They are the best in damp conditions (though wrapping in a sandwich bag may be a good idea!) and they don`t need any power supply. The linearity thing is perhaps something you have to live with - there`s rarely anything linear about noise and its generation, including the scale(s) we use to measure it and our response to it. But as you say it is sensible to assume that when more water starts going over a pitch it is going to get louder.
In a year the hot wire anemometer you refer to will chomp through any power supply you can sensibly provide. The one I used to use (some years ago now I must admit) needed to have its 2 "U2" type batteries replaced every couple of hours, even though we switched it off as much as we could.
If I were wanting to make something to do the job I would start with a small dc motor, (eBay again?) put some vanes on it and see what I got across the windings when I blew on the vanes. Same principle as the dynamic mic and so no power supply needed. But you might need some big vanes if there`s not much of an airflow!
 

ian.p

Active member
I think a key thing to remember with science, is that if you come back with the only data, then you have the best data in the world 
still dunt necessarily make it usefull though!
 

Bob Mehew

Well-known member
graham said:
ian.p said:
I think a key thing to remember with science, is that if you come back with the only data, then you have the best data in the world 
still dunt necessarily make it usefull though!

(y)

But surely Jarvist's objective is not to measure flow which he perhaps erroneously mentions in the title of the thread but to correlate "...the temporal arrival of flood pulses with surface rain data etc...".  Even so one would have to calibrate for no / soft / loud noise verses flow conditions.  But how about something like http://www.ebay.co.uk/itm/Flow-Scout-Meter-Thermometer-Water-Liquid-Cooler-System-5V-/190652524833?pt=UK_BOI_Electrical_Test_Measurement_Equipment_ET&hash=item2c63c6a921 modified with magnets at the 3 tips and a small coil sensor so one is sensing an emf pulse generated by the system?  (I give credit to Chocolate Fireguard's ensemble of techniques for using magnets.)
 

dmcfarlane

New member
Why not use a pressure logger?  I have used the Solinist versions (""Levellogger") in 'white water" streams in Costa Rica- they are sensitive to < 0.5 cms water level, have a ten year battery, store tens of thousands of data points.  The device is totally submersible, so you can put it inside a piece of vented steel pipe epoxied to the rock surface, or straight into a 1" hole drilled in the rock (they are about 6 inches long and 7/8ths diameter).  You do also need a "barologger" to correct for barometric pressure, so its not a cheap option, but they work extremely well.

DM
 

jarvist

New member
Bob Mehew said:
he perhaps erroneously mentions in the title of the thread but to correlate

Ah yes, my apologies. I certainly wasn't trying to suggest you could go directly from dB of noise --> flow rate with no other information. But I do still suspect that noise power level will be proportional to flow rate, though quite possibly with nasty non linearities and transitions. Which is a complicated way of saying that 'if the pitch sounds more noisy, more water is falling down it'.

Price is a certainly a strong factor, as this is entirely for my own entertainment!
It certainly seems possible to put together a data logger with accurate barometer, temperature and microphone for a parts cost of around ?30-40, based around the Arduino / Atmel AVR platform.
A friend also suggested the really cute idea of getting it to record a segment of ambient cave noise at Midnight on the new year.

I think an easy way of building a 'propeller' anemometer would be to start with a large (new!) brushless computer fan, then directly tap the coils / use the hall effect switches built into it. The problem is that cave wind speed is so low that you have very little force at all to deal with, you'd need to find something with very good bearings to have it spin at low flow rates.
The good thing with a fan is that it's easy to see which way the flow is! With a hot wire anemometer, you would probably need a second hot wire purposefully shadowed from the wind in one direction.

Still, I honestly think that with a bit of design and engineering you could build a little data logging box that'd be able to extract a lot of interesting information including audio remote sensing, and do some serious science. I'd certainly be interested to know how the cave draught varies with weather systems + seasons, and understand a little more of the hydrology + nature of rain driven flood pulses vs. yearly snow melt.
 

TheBitterEnd

Well-known member
Sorry but not quite given up on the water depth idea yet (and I'm sure you have been round this a few times yourself) but what about something like a bike innertube in the stream bed with a tube going back to the high'n'dry data logger with a pressure sensor on the end.

The innertube could be in a box (e.g. ammobox) that could be filled with rocks in the cave - it would still need attaching to something stop it going over the pitch but at least your data logger would be well out of the way.

Whatever you end up with, good luck with it, I think it's great that you are even attempting it.
 
Top