Author Topic: Draught testing challenge  (Read 3292 times)

Offline Pitlamp

  • forum hero
  • *****
  • Posts: 6211
Re: Draught testing challenge
« Reply #50 on: November 16, 2021, 09:01:43 pm »
Some interesting info there. I wonder if anyone has a Hirst spore trap doing nothing for a while. It all looks rather bulky though.

Offline Bob Mehew

  • forum hero
  • *****
  • Posts: 1422
  • breaking knots was fun
Re: Draught testing challenge
« Reply #51 on: November 16, 2021, 09:21:12 pm »
The challenge with lycopodium spore is that each spore is heavier than air and can, when making contact with a surface (especially one which is damp or wet) stick to that surface.  So the spore will be 'filtered' out as the current of air wafts it way through the passage.  Maybe OK for a few tens of metres but hundreds of metres doubtful.  If one has access to a mass spectrometer, then one could use an inert gas.  Argon won't work since it naturally present at just under 1% in air. 

Mercapten is reported to have an odour at a concentration of down to 1 ppb (part per billon or 1 in 1,000,000,000) see https://en.wikipedia.org/wiki/Methanethiol. If I have it correct that is equivalent to 1 gram in 1,000,000 metre cubed or roughly the volume of 23 Gapping Gill Main Chambers (Descent 229 p26).  So it has potential.  But given it is also toxic above 10 ppm (parts per million), safely releasing it presents a challenge.  (Oh and it is also explosive up in the % range.)  One potential problem is that it is soluble in water, so could be 'washed' out.

Offline shotlighter

  • junky
  • ****
  • Posts: 820
Re: Draught testing challenge
« Reply #52 on: November 16, 2021, 10:58:24 pm »
Mercapten is reported to have an odour at a concentration of down to 1 ppb (part per billon or 1 in 1,000,000,000) see https://en.wikipedia.org/wiki/Methanethiol. If I have it correct that is equivalent to 1 gram in 1,000,000 metre cubed or roughly the volume of 23 Gapping Gill Main Chambers (Descent 229 p26).  So it has potential.  But given it is also toxic above 10 ppm (parts per million), safely releasing it presents a challenge.  (Oh and it is also explosive up in the % range.)  One potential problem is that it is soluble in water, so could be 'washed' out.

You really don’t want mercaptan hanging round. Apart from being pretty foul stuff, it’s also a bloody nuisance.
True story, in the late 70’s the analytical instrument company that I worked for had in office on a small town High St. One day we noticed a hole outside with “West Midlands Gas Board” barriers round it. Apparently they were looking for a reported gas leak & over the next few weeks, the hole & workmen came and went. The hole was in a slightly different place each time, as they couldn’t locate the leak. Then one of my colleagues had a moment of enlightenment. The hole digging had started about the same time that we received a customers sample. He'd sent it so that we could do a HPLC quantitive analysis method for him. The compound of interest in the sample was mercaptan, a tiny amount of which is added to natural gas to make it smell.....
We quietly returned the sample to the customer and shortly after, the holes & workmen went away.

Offline Pitlamp

  • forum hero
  • *****
  • Posts: 6211
Re: Draught testing challenge
« Reply #53 on: November 17, 2021, 08:10:36 am »
The challenge with lycopodium spore is that each spore is heavier than air and can, when making contact with a surface (especially one which is damp or wet) stick to that surface.  So the spore will be 'filtered' out as the current of air wafts it way through the passage.  Maybe OK for a few tens of metres but hundreds of metres doubtful.  If one has access to a mass spectrometer, then one could use an inert gas.  Argon won't work since it naturally present at just under 1% in air. 

Hm; thanks Bob. The distance in question is a couple of kilometres in a straight line, so not sure the spore idea is practicable.

Do we have an aerobiologist in the caving community who might be able to advise?

Offline T pot 2

  • obsessive maniac
  • ***
  • Posts: 496
Re: Draught testing challenge
« Reply #54 on: November 17, 2021, 04:01:23 pm »
During copper and lead mining when compressed air was used to operate the drills in deep mines the alarm used to vacate mines was the smell of eucalyptus, which was drawn into the compressed airlines on the surface and delivered deep underground


Offline Chocolate fireguard

  • obsessive maniac
  • ***
  • Posts: 454
Re: Draught testing challenge
« Reply #55 on: November 17, 2021, 06:20:55 pm »
How about using some poly sheet and a large fan at one end or the other to try to engineer a draught reversal? Only a small change in pressure may be needed and no fumes to worry about other than from your generator for the fan.
No one has commented on this, but if it was my project and I really wanted to test a link between caves that far apart I would give it serious thought.
It would cost a bob or two and need some organising, but could work.
The main advantage is that the stuff you are putting in at the top end (air) doesn't need to travel all the way to the other end, it just has to pressurise the system so the flow rate out of the bottom end increases measurably.

Offline Pitlamp

  • forum hero
  • *****
  • Posts: 6211
Re: Draught testing challenge
« Reply #56 on: November 17, 2021, 07:41:57 pm »
That's a really interesting suggestion Chocolate Fireguard - but we suspect the system in question has a significant number of entrances, so it may be too leaky for that to work effectively enough. If it was only a single conduit that'd be a cracking idea.

Offline Bob Mehew

  • forum hero
  • *****
  • Posts: 1422
  • breaking knots was fun
Re: Draught testing challenge
« Reply #57 on: November 17, 2021, 07:49:56 pm »
So if we are thinking of a 1,000,000 metre cubed volume of passage way, one would only need to pump in an extra 1,000 m3 of air to get a 1 milli bar pressure rise.  Air fans moving 4000 m3 per hour are available.  The BMP 085 pressure sensor has a noise level of down to 0.03 mBar so 1 mBar should be easily discernable.  The major assumption would be there are no leaks between the two ends which are sealed.  But I would expect the pressure change with respect to time at the other end of the system away from the fan would be spread out compared to that by the fan.  What I don't have data on is the rate of natural pressure change.  Is all that correct?

Offline Pitlamp

  • forum hero
  • *****
  • Posts: 6211
Re: Draught testing challenge
« Reply #58 on: November 17, 2021, 07:51:32 pm »
That all sounds sensible to me Bob.

Offline Steve Clark

  • regular
  • *
  • Posts: 59
Re: Draught testing challenge
« Reply #59 on: November 17, 2021, 08:57:52 pm »
From my limited experience of pressure testing buildings, the external factors can be quite large. Wind being the worst. If the two entrances (and unknowns in between) are in different locations, the wind gusting over the fell will affect each at different times.

Temperature is also a factor. Obviously limited solar gain, but there may well be convection dependant on the average cave temp vs. external temp. This varies with time and location.

I think probably the best way would be to run the fan with a pre-determined sequence (eg 1min on, 1min off, repeat x times, then reverse the direction and do it again). Then you could data log the digital manometer on the other end and see if you can pick out the signal by Fourier or similar.

Offline Steve Clark

  • regular
  • *
  • Posts: 59
Re: Draught testing challenge
« Reply #60 on: November 17, 2021, 09:29:57 pm »
Also, when doing the rough calcs, remember that it’s not just adding 1000m3 of air to a sealed container of Xmilliom m3 causing a simple pressure rise. The other end (and possibly other  routes) are open. The pressure change will just cause flow due to massive leakage. Detecting the change of pressure due to flow change needs to be much more accurate. (1mbar is the stagnation pressure for about 30mph air).


Offline Chocolate fireguard

  • obsessive maniac
  • ***
  • Posts: 454
Re: Draught testing challenge
« Reply #61 on: November 17, 2021, 10:12:34 pm »
I think the way to go would be to measure air speed at the lower end - I wouldn't expect any real rise in pressure as it would remain essentially the same as that just outside the entrance.
Unless the result was dramatic I expect some sort of data logger would be needed.
But lots of other entrances and the large distance probably mean it's a non-starter.

Online Mr Dinwiddy

  • regular
  • *
  • Posts: 60
Re: Draught testing challenge
« Reply #62 on: November 17, 2021, 10:33:12 pm »
I think Lycopodium power still merits further consideration. I agree with Bob Mehew's assessment that spores will get stuck to surfaces but (a simple google search shows that) one gram of Lycopodium powder contains 4.89*107 spores. That gives considerable scope for some to get stuck on cave surfaces and others to make it through to traps.
This paper https://royalsocietypublishing.org/doi/10.1098/rspa.1967.0005 suggests that under normal meteorological conditions spores dropped from 50cm height can be transported 1km (median range) and dropping them from 10m gave a range of 10km so they are highly mobile in air. A cave with a draft, even a small one could potentially transport spores some distance. I think it is possible to get hold of dyed spores so you can distinguish between your test and naturally occurring spores from club mosses growing in the Dales. That paper also talks about radioactive tracers but that gets over complicated and I like the simplicity of lycopodium spore testing.

Offline Pitlamp

  • forum hero
  • *****
  • Posts: 6211
Re: Draught testing challenge
« Reply #63 on: November 18, 2021, 05:22:29 am »
Thanks folks, some really useful information above.

I feel like I'm slightly getting out of my depth here; just wondering if anyone who knows what they're doing might be interested in taking this little project on, next spring / summer? Apart from the interesting riddle we're trying to solve, if a method could be devised which is successful, this would then be available for others to use in similar circumstances in the future.

If anyone's interested, maybe PM me? (It's only a few miles from Ingleton, by the way.)

Offline Bob Mehew

  • forum hero
  • *****
  • Posts: 1422
  • breaking knots was fun
Re: Draught testing challenge
« Reply #64 on: November 18, 2021, 09:03:54 am »
I started with the premise that the two ends were sealed as per digscaves statement.  I agree that running the fan in some sequence would be the way to do this and accept that external factors could play a significant role.  It is a trivial matter to link a BMP085 to an Arduino mini computer to data log pressure over some time period.  Though the power source might be a challenge for long term deployment as well as protection from humidity.  But it leads me to wonder if it might be worth trying this out on two known linked entrances without a fan to see if there is some relationship in pressure variation when deploying a sensor inside each entrance and another sensor away from each entrance to capture natural changes.  What are the flaws?  And any pointers to fourier signal processing?

Another concern about the lycopodium spore approach is how do you capture the remaining spore at the other end?  Using dyed spore might also foul up the system from being able to use dyed spore for water tracing.

Offline Steve Clark

  • regular
  • *
  • Posts: 59
Re: Draught testing challenge
« Reply #65 on: November 18, 2021, 11:22:20 am »
I'm no expert in Fourier transforms and haven't used one in anger for 20years.

The principle is to convert a signal (eg. pressure vs. time) into a frequency plot. Allows you to see what frequencies are most common in the signal. Most common example is a graphic-equaliser on a 1980's hifi.

With a known input signal frequency (eg. 1min on / 1min off). Cycle time of 2min = 1/120Hz = 0.008Hz. You can run a fourier transform on the recorded output signal and see if you can see this frequency in the plot. And also not see it at other times.

With large and long datasets, it's possible to see interesting frequencies in the data. Day/night frequency vs. Tide/moon frequency may appear for example. Also very easy for other noise to appear eg. weather.

You need to pick a sufficiently fast sampling frequency, such that it includes the input frequency and ideally much higher. Eg. if you're only sampling every 10mins, you can't see a 2min long waveform in the data. Sampling once per sec would be more appropriate.

We used it once to see if we could find the source of some interference on the comms cable to a set of traffic lights in Preston. Whenever we turned up to site it didn't happen. Turned out to be a motorcycle policeman who liked leaning on the control cabinet with his radio. He didn't lean on the cabinet when the engineers came to site. When they hid around the corner, it became obvious.
« Last Edit: November 18, 2021, 11:38:15 am by Steve Clark »

Offline Graigwen

  • forum star
  • ****
  • Posts: 619
Re: Draught testing challenge
« Reply #66 on: November 18, 2021, 11:45:21 am »


Another concern about the lycopodium spore approach is how do you capture the remaining spore at the other end?  Using dyed spore might also foul up the system from being able to use dyed spore for water tracing.

Lycopodium spores evolved to get everywhere. they will get everywhere!

Offline Bob Mehew

  • forum hero
  • *****
  • Posts: 1422
  • breaking knots was fun
Re: Draught testing challenge
« Reply #67 on: November 18, 2021, 12:59:50 pm »
Right.  I think https://pythontic.com/visualization/signals/fouriertransform_fft shows the basic process (ignore the coding).  Let's assume we have 4 sensors, A outside entrance 1, B inside entrance 1, C outside entrance 2 and D inside entrance 2.  So the challenge is taking data from the 4 sensors, applying the Fourier transform to each data set (and presumably sub sets over specific time periods) and see if one gets similar frequency components.  Hopefully one might see the same frequency components from sensor B & D but not from A & C.  There is a possibility that the sub sets might show up a time slip between sensors B & D and also A & C.  I guess the first thing is to try it out on a simple top & bottom entrance and hope wind conditions will provide the variable signal.  Any suggestions for a simple but long system which one could place detectors in without the need for gear?  Imposing a signal needs some thought.

Offline Bob Mehew

  • forum hero
  • *****
  • Posts: 1422
  • breaking knots was fun
Re: Draught testing challenge
« Reply #68 on: November 18, 2021, 01:00:17 pm »
Lycopodium spores evolved to get everywhere. they will get everywhere!
Except assuming Sod's Law, not the detector.

Offline Chocolate fireguard

  • obsessive maniac
  • ***
  • Posts: 454
Re: Draught testing challenge
« Reply #69 on: November 18, 2021, 01:17:20 pm »
Thanks folks, some really useful information above.

I feel like I'm slightly getting out of my depth here; just wondering if anyone who knows what they're doing might be interested in taking this little project on, next spring / summer? Apart from the interesting riddle we're trying to solve, if a method could be devised which is successful, this would then be available for others to use in similar circumstances in the future.

If anyone's interested, maybe PM me? (It's only a few miles from Ingleton, by the way.)
What is the vertical separation of the 2 sites?

Offline Pitlamp

  • forum hero
  • *****
  • Posts: 6211
Re: Draught testing challenge
« Reply #70 on: November 18, 2021, 01:23:47 pm »
Any suggestions for a simple but long system which one could place detectors in without the need for gear?  Imposing a signal needs some thought.

Is Rowten Cave long enough for an initial test? Or would Great Douk be a good site? (This has the advantage of an open pot dropping in - so maybe tests could be done from both upstream & downstream of this point, to see if the presence of a large entrance fouls up the method(s) of choice.
« Last Edit: November 18, 2021, 01:43:45 pm by Pitlamp »

Offline Pitlamp

  • forum hero
  • *****
  • Posts: 6211
Re: Draught testing challenge
« Reply #71 on: November 18, 2021, 01:42:35 pm »
I don't think I ever gave details in this topic of the reason I originally asked the question. Austwick Beck Head has been extended this year, mainly by diving. Sump 1 is short and in dry weather there is a thin airspace. There has been an inward draught here right through the warmer months of 2021. But there should have been an outward draught in hot weather, if ABH is the lowest entrance of the system it drains. (Sump 2 beyond is never unroofed and the draught goes into a choked bedding above water between the sumps, which is yet to be investigated properly. Sump 1 is not safely passable without diving gear by the way, as the route through involves complete submersion, whether or not airspace exists, via a couple of awkward restrictions.)

There must be another entrance which is low enough to generate a convectional draught. Contenders include Blind Beck Cave (22 m lower) and Beck Head Stream Cave (32 m lower). BBC has a feint outward draught in warm weather; BHSC has a strong outward draught. Could either of these two air currents originate from ABH?

Being able to trace such draughts, in the manner of tracing water, would be a very useful technique to develop for cavers generally. Maybe a BCRA field day, bringing folk together who are interested, might be of value?

Offline Leclused

  • forum star
  • ****
  • Posts: 594
    • The Avalon Blog
Re: Draught testing challenge
« Reply #72 on: November 18, 2021, 02:02:27 pm »
I don't think I ever gave details in this topic of the reason I originally asked the question. Austwick Beck Head has been extended this year, mainly by diving. Sump 1 is short and in dry weather there is a thin airspace. There has been an inward draught here right through the warmer months of 2021. But there should have been an outward draught in hot weather, if ABH is the lowest entrance of the system it drains. (Sump 2 beyond is never unroofed and the draught goes into a choked bedding above water between the sumps, which is yet to be investigated properly. Sump 1 is not safely passable without diving gear by the way, as the route through involves complete submersion, whether or not airspace exists, via a couple of awkward restrictions.)

There must be another entrance which is low enough to generate a convectional draught. Contenders include Blind Beck Cave (22 m lower) and Beck Head Stream Cave (32 m lower). BBC has a feint outward draught in warm weather; BHSC has a strong outward draught. Could either of these two air currents originate from ABH?

Being able to trace such draughts, in the manner of tracing water, would be a very useful technique to develop for cavers generally. Maybe a BCRA field day, bringing folk together who are interested, might be of value?

In that case I would suggest to locate the position on the surface and see what is there.

We have done this several times already. There are two options

- using avalange beepers (ARVA)
- using an ARCANA system (some info https://www.unilim.fr/fablab19/wp-content/uploads/sites/8/2018/10/notice_arcana_fablab19.pdf )
                                         or http://wittamdoun.free.fr/documents/radiolocation.pdf )


Arcana is very accurate and is the ultimate solution to pin point location and depth of a point in a cave in relations with a point at the surface.

We are currently digging on a spot to make an entrance in one of our projects in France. The arcane gave 5m in depth :-)

Dagobert L'Ecluse (SC Avalon - Belgium)
http://www.scavalon.be
http://scavalon.blogspot.be

Offline Steve Clark

  • regular
  • *
  • Posts: 59
Re: Draught testing challenge
« Reply #73 on: November 18, 2021, 02:15:21 pm »
Right.  I think https://pythontic.com/visualization/signals/fouriertransform_fft shows the basic process (ignore the coding).  Let's assume we have 4 sensors, A outside entrance 1, B inside entrance 1, C outside entrance 2 and D inside entrance 2.  So the challenge is taking data from the 4 sensors, applying the Fourier transform to each data set (and presumably sub sets over specific time periods) and see if one gets similar frequency components.  Hopefully one might see the same frequency components from sensor B & D but not from A & C.  There is a possibility that the sub sets might show up a time slip between sensors B & D and also A & C.  I guess the first thing is to try it out on a simple top & bottom entrance and hope wind conditions will provide the variable signal.  Any suggestions for a simple but long system which one could place detectors in without the need for gear?  Imposing a signal needs some thought.

When we test buildings, they use long thin surgical tubing to sample pressure at particular locations. The tubes come back to a centralised location for the sensors. You can use very accurate differential manometers this way. This also means you could put the sensor pack on the surface where it's easy to change batteries & get a 4G signal.

They use this method in Wakulla (northern Florida) for sampling the flow rate at the resurgence down at -90m. Volunteer divers fit & maintain the flow rate sensor & water sampling tubes to the surface and the State park / USGS folks can do all the technical work on the surface in a cabinet. The state/scientists get to do science that would otherwise cost hundreds of thousands of dollars to do with commercial divers and the dive team get a permit to dive & explore the Wakulla system. There is rumour that some of them may even have enjoyed it.......

 

Main Menu

Forum Home Help Search
SimplePortal 2.3.5 © 2008-2012, SimplePortal