Measuring water flow rates

[darklord]

Maybe not very scientific, but what is the current best method for measuring flow rates of streams? I want to measure amount of water going in sinks and coming out of risings...see if they all add up!

Stream beds are a series of variable width/depth pools and falls, and I can't see any accurate way of floating a bit of paper down for a cu/sec measurement!  I'm sure there must be a more modern method...

DL

 

[langcliffe]

Salt dilution techniques?

 

[graham]

v- notch weirs.

 

[TheBitterEnd]

In industrial applications height over weirs is used (e.g. http://www.lmnoeng.com/Weirs/vweir.htm) but for decent levels of accuracy straight sided channels are required upstream and downstream. Where weirs aren't practical one method is to measure the cross section and the velocity at various points throughout the cross section is measured and work out flow from that.

That said I remember doing and experiment at college using conductivity to measure flow in a stream, find a reasonably straight stretch with a reasonably uniform cross section, put some salt in upstream and measure conductivity downstream plot conductivity against time.

(Just noticed that Graham has also posted re. wires)

 

[langcliffe]

That said I remember doing and experiment at college using conductivity to measure flow in a stream, find a reasonably straight stretch with a reasonably uniform cross section, put some salt in upstream and measure conductivity downstream plot conductivity against time.

= salt dilution techniques also as above (the alternative to measuring the conductivity is titrametric analysis).

 

[TheBitterEnd]

Not wanting to get into some hair-splitting debate but the way we did it was != salt dilution.

The salt was just acting in a way like bits of paper on the surface, except it was in the water body and therefore a better indicator of velocity. A non-turbulent stretch of stream of reasonably uniform cross-section was chosen (ideally with plug-flow). Salt was tipped in upstream and downstream a stop watch was started and conductivity measurements taken every few seconds. As the salt passed through, the conductivity peaked and the time at which peak conductivity occurred was taken as the mean time of travel for the salt. The distance between where the salt entered and where the conductivity was read was measured, as was the cross-sectional area and hence flow rate could be calculated. The actual dilution of the salt was not critical.

 

[langcliffe]

Not wanting to get into some hair-splitting debate but the way we did it was != salt dilution.

OK - fair enough. I can see that working with "a non-turbulent stream of stream of reasonable cross-section", but I can't really see it as being particularly effective for stream beds which "are a series of variable width/depth pools and falls" as the OP required which is the question I thought you were answering.

 

[SamT]

Flow measurements are incredibly difficult to do accurately without high quality good V notch weirs, smooth walls etc. data loggers etc. - even then, they are a little bit of a rough science.

Do you proposed data loggers etc, so you can compare sinks and resurgences within the same time frame. (no use measuring the inputs one day, and the outputs another, if the water levels are rising/falling).  Also - there will be the time lag between sink and resurgence to consider.

So, you have to ask yourself, with the degrees of uncertainty involved, will your results prove anything with regards to adding up input and outputs.



(not suggesting its not worth trying though  )

 

[bograt]

To get acceptable vol/time values, an accurate measurement of volume is required, this will entail catching all the water in the stream and channeling it into an easily calculable cross section (square, triangle, etc.), there are then many methods of measuring the speed of flow through this channel -  flow meter, drum and stopwatch etc.

The capture of the water is the biggest problem, an alternative to concrete engineering might be inflatable barages? If a wooden, or plastic, or metal rectangular leet is let into it with a fixed depth guage, the volume of flow will be easily calculable. Alternatively set a v-notch into it, this will require a more involved calculation.

 

[langcliffe]

Salt / Gulp dilution techniques were designed to get an estimate of stream flow without requiring knowledge of the cross-sectional area of the channel and the speed of flow. Here's is an example of an undergraduate field study exercise undertaken in White Scar Cave: http://www.es.lancs.ac.uk/people/nickc/104/case16a.htm

This paper by the same author reviews stream discharge measurement techniques: http://www.es.lancs.ac.uk/people/nickc/104/case16.htm

 

[SamT]

Salt / Gulp dilution techniques were designed to get an estimate of stream flow without requiring knowledge of the cross-sectional area of the channel and the speed of flow.

And, from everything I've ever heard/read - fairly in-accurate.

 

[langcliffe]

And, from everything I've ever heard/read - fairly in-accurate.

Yes - it is inaccurate compared to structural engineering techniques, which is not often practical for the amateur hydrologist. When we did it under controlled conditions, the accuracy was +-10% in a reasonable sized stream. Can you think of a practical more accurate technique?

I think that your well-made point about correlating the stage of all inputs and the output is probably of more significance as far as the accuracy is concerned.

 

[bograt]

Hmmm ..... This technique could answer the OP's requirements though, the conductivity readings in and out would indicate the ratio of in/out flows, comparing these ratios for various ins could add up to the total out, and also possible indicate any unknown ins, without needing to know the actual volumes, the measurements would indicate the dilution ratios.

 

[SamT]

Can you think of a practical more accurate technique?

No - its a very difficult measurement to undertake.

and...

I think that your well-made point about correlating the stage of all inputs and the output is probably of more significance as far as the accuracy is concerned.

Ergo - if your going to do it at all, it should be done as accurately as possible otherwise the exercise is a little pointless.


8)

 

[darklord]

Hmmm....!! Thanks for all the input and observations.  Looks like it would have to be a fairly determined 'study' with some instrumentation needed (for salt analysis) and/or engineering for weirs. Obviously I was looking for a nice, simple technique... and it seems there isn't one!

It's just that you quite often hear cumec and cusec figures bandied about for various risings (and sometimes sinks), but without any background on the measurement techniques.

I'm thinking pretty small Dales type streams here, not massive French vauclusian fountains or anything!  Rudimentary though it might seem, I did wonder about damming streams (even though there would be a bit of leakage) and using a bit of pipe to channel water into a fixed size container. Presumably the time taken to fill the known size container would give you at least a relative flow rate?

 

[langcliffe]

Ergo - if your going to do it at all, it should be done as accurately as possible otherwise the exercise is a little pointless.

I really don't understand this, but I expect that you're right.

 

[bograt]

Rudimentary though it might seem, I did wonder about damming streams (even though there would be a bit of leakage) and using a bit of pipe to channel water into a fixed size container. Presumably the time taken to fill the known size container would give you at least a relative flow rate?

No, this would only give you the flow rate of the pipe used to fill the container and this would depend upon the head of water above the pipe inlet and the drop of the pipe, you would need to feed the whole stream unencumbered into the container.

 

[darklord]

you would need to feed the whole stream unencumbered into the container.

That would be a bit tricky......!

 

[TheBitterEnd]

There are things like this

http://www.bamo.eu/international/s4_page575.htm]See [url]http://www.bamo.eu/international/s4_page575.htm[/url]

For a small-ish dales stream it would not be beyond reason to cobble something like that up from scaffold boards, especially if you did it at low flows (if we ever get any dry weather). Obviously getting the shape of the venturi correct is an issue but something similar could be done with a V-notch. This could then be sand-bagged into the stream bed.

The point about simultaneous readings could be handled by having several of these and a person at each to read them at fixed times. There again, given a dry spell, I wonder how much variation you would get over a day if you read one at a time?

 

[langcliffe]

The point about simultaneous readings could be handled by having several of these and a person at each to read them at fixed times.

That would work fine except for the resurgence, where there will be a distinct time lag in the majority of cases.

There again, given a dry spell, I wonder how much variation you would get over a day if you read one at a time?

Very little - unfortunately there is very little water to measure in a dry spell, and a relatively high proportion of it tends to hide away. I suspect that the uncertainties will increase under such conditions.