Poly beads level detection

[gophert]

If you can lift the silo, Load cells are a great option.

https://www.futek.com/application/load-cell/Silo-Measurement

 

[BobW]

Point level in liquids or granules?

Their inability to completely address static discharge is concerning (specifically for plastic pellets)...
https://www.drexelbrook.com/-/media...te-electrostatic-hazards-in-granular-bins.pdf

We generally refer those concerns to their engineering support services. They've been very responsive to these kinds of application related questions (as most instrumentation companies are).

We have also used the vibrating probe type sensor, that was mentioned above, with good results, but found them to be not as reliable as the RF admittance or capacitance type probes. The one thing that I learned the hard way when I started in this field, is to avoid any instrument that involves any kind of moving parts. So, while there have been several clever suggestions offered involving moving parts, they are doomed to have a very short life.

 

[Les Jones]

Here is a suggestion that has a slight possibility of working. It is based on the resonant frequency of the air in the silo between the top of the beads and the (I assume.) open top. The velocity of sound in air is about 340 metres per second. As the first resonance occurs with the air column equal to a quarter of a wavelength the empty silo would resonate at 340 meteres per second/( 4 x 10 metres) = 340/40 = 8.5 hz. The idea is to drive a loudspeaker at the top of the silo with a swept frequency staring at about 8 hz. There would also be a microphone monitoring the sound amplitude. A microcontroller would record a large number of samples of the amplitude and frequency. At the end of the sweep it would take the sample with the highest amplitude and from the frequency of this sample would calculate the length of the column of air.

Les.

 

[dr pepper]

The silo is outside and has a sealed top.
That said Les wouldnt your principle still work?, certainly in a sealed type loudspeaker the volume of air affects the resonance.
The poly is taken from the tank with a vacuum conveyor system, however there is a delivery hopper at the base of the silo which isolates the vacuum from the tank, as obviously we dont want the thing to bu sucked inside out.
I might need to compensate for air temp, as on a cold day it might be below zero in there, and on a hot day the sun gets it over 40, certainly gets to hot to sit on top!
The only thing is a mic that will go down to those frequencies, the speaker itself might have to be used as a mic.

 

[Les Jones]

The biggest potential problem I see is if the beads absorb sound. If the beads are polystyrene foam then they would probably absorb sound so it would probably not work. If they were hard solid polystyrene then there is at least a chance that it would work. I agree on probably having to use a loudspeaker as a microphone. when I posted this idea I thought it may prompt a response from members with a better knowledge of acoustics than I have. I think a closed vessel would have a half wavelength between the ends as opposed to a quarter wavelength for an open vessel so the lowest frequency would then be 17 hz. I was pleased to here that my first suggestion was not as daft as I thought it might be.

Les

 

[Pommie]

I know it's polystyrene beads but they still weigh over 50 tonnes. The pressure at the bottom must be immense. A rubber bladder with a pressure sensor might be worth exploring.

Mike.

 

[dr pepper]

Yes rough calcs its about 40 psi, that has 2 problems to solve, one is the bladder would have to be thick enough to be able to compress the air within without the sides touching, and I dont know where I'd be able to have such a thing made, assuming I was to have something the entire surface area that is, maybe a smaller one would work.
Sounds like a similar idea to a load cell and a plate to measure a portion of the area in the bottom of the tank, I've never seen anything like that done.

 

[Pommie]

Do footballs still have bladders? Could a hot water bottle suffice? Maybe a football itself could be a bladder. The available solutions are immense.

Mike.

 

[dr pepper]

I think I mis-explained, originally I thought you meant a bladder the full area of the base.
Various smallish industrial bladder type assy's are available from air bellows to hydraulic accumulator ones, conveniently with a fitting on the end.
I like the idea of say a 12" square fastened to the bottom with a load cell attached, the hole in the bottom of the silo is 12", any larger than that and it would be a big job, as the whole thing is welded up apart from a vent at the top, the hole for the original guage and the hole just mentioned at the bottom.

 

[Cicero]

Just out of interest, what kind of resolution are you wanting?

 

[alec_t]

I like the idea of say a 12" square fastened to the bottom with a load cell attached

That should work. It's a sufficiently large area that there is little likelihood of the beads forming an air-trapping 'tent' over the plate (which would reduce the pressure on the plate).

 

[BobW]

The problem with load cells/strain gauges is that they drift. They're fine for measuring tank levels as part of a short term batching operation where the tank fills and then completely empties frequently. When it reaches the empty part of the cycle, then the load cell is re-zeroed, so that the drift errors are regularly corrected. When used for long term level measurement they are unreliable due to the drift.

 

[dr pepper]

Well that kinda poo poo's that then, still good to know however.
The tank remains part full all the time, or the plant would shut.
So the way to do that the would be to use the bladder idea.
Or something I came up with is a 'rope' of 20 capacitive proxy's dangled into the silo.

 

[Cicero]

This is a bit out there, but what about a 10m strip with 20 or so LED's evenly spaced, gives 50cm resolution.

Flash the led's individually in a Time Division Multiplex kind of fashion. Use a single phototransistor to pick up the flashes, in sequence. Point the LED's towards your phototransistor.

Whichever LED's signal you dont receive, you can assume are covered. Flash for as long as required to filter out the 'angel-hair' optical interruptions.

 

[gophert]

Well that kinda poo poo's that then, still good to know however.
The tank remains part full all the time, or the plant would shut.
So the way to do that the would be to use the bladder idea.
Or something I came up with is a 'rope' of 20 capacitive proxy's dangled into the silo.

Drift is barely an issue on a 50 ton silo and 20 ton refills.
You will be refilling at least every two weeks, I assume.

I don't think you need even kilogram resolution. You just need an alarm/notification to let you know the silo needs to be refilled. You are not using it to dispense, right?

 

[dr pepper]

5% tolerance is good enough, your right this is just a 'fuel guage', refills can be 2 times a week to 2 weeks.
However the silo is hardly ever empty, wont drifit accumulate, or will it just hover around a few %.
The proxy's seem like a good idea, ready made and can be just dangled in from the top.

 

[gophert]

5% tolerance is good enough, your right this is just a 'fuel guage', refills can be 2 times a week to 2 weeks.
However the silo is hardly ever empty, wont drifit accumulate, or will it just hover around a few %.
The proxy's seem like a good idea, ready made and can be just dangled in from the top.

Load cells were used all over our plant. I've never heard of so much concern about drift. You can deal with it. Get the right meter or build your own. https://www.ti.com/lit/ds/symlink/lmp8358.pdf

 

[dr pepper]

Done that, used a amplifier (Hx711) with the arduino, the output string is compatible with usual software, the operator zeros it all the time.
Been on hols for a bit, will return with some ideas next week.

 

[dr pepper]

Latest development.
Rigging up a test rig I found a capacitance probe seems to work, sticking 2 wires into a bin of poly beads varies the capacitance fairly linearly with the amount of probe submerged.
So this may work for measuring the level in a silo, I just need to see if different poly's have wildly different dielectrics.
Not sure whether to run 2 probes in the tank, or to use just one probe and the aluminium tank as the second probe.
The tank itself is connected to mains ground (and a lightening conductor) so could be guage system 0v.

 

[alec_t]

Sounds promising. One conductive rod/strip plus the silo wall should be more robust than two wires. What was the wire spacing?