55 Gal Drum - Grit Media Level Sensing

[AMS2750]

I have an application where I am trying to detect grit media level (anywhere from "sand" to aluminum powder) in a 55 gallon metal drum with a metal lid. This is a waste drum, so I'm looking for a 90% and 99% full signal to alert the operator that it's time to swap the drum with an empty one.

Right now, I'm using 2 shielded 4mm range capacitive proximity switches. They are hanging from the lid so that 1 trips at about 90% and the other at about 99% full.

This scenario works relatively well, except that:

1. We've had a couple of false positives due to the sensor getting dirty from the grit media swirling around in the drum.
2. This setup requires the proximity switches to hang in the drum, so I have to pull the hanging bracket out of the lid every time I change the drum because the lid can't really be lifted off more than an inch at best to slide the drums in and out.
3. The proximity switches are N.O., so a disconnected cable, etc... will make the system think everything is OK.
4. The "hanging" system is rather cumbersome

I've tried to address those issues by purchasing 2 5-40mm adjustable range N.C./N.O. non-shielded capacitive proximity switches. The goal was to mount both of them in the lid in a way they wouldn't need to be taken out when changing the drum. Then take advantage of the adjustable range to set one to trigger at 90% and the other at 99% and set them to N.C. to make them fail-safe.

I tested this about an hour ago. I purposefully got the proximity switch tip "dirty" with grit media and mounted it on a steel plate to mimick the lid environment. Then I held the assembly over a partially filled drum and by the time I zeroed the proximity switch, I didn't have enough sensitivity range left to make the switch even remotely reliable.

I don't want to keep buying proximity switches just to try something else. I admit, I probably should try to get a shielded proximity switch and that might improve my chances to succeed, but at this point, I'd like a little guidance from the experts here who have much more experience than I using level sensing devices under a variety of applications.

I already know from previous research that sonar is not good in powder. Radar is prohibitively expensive for this application. Optical would probably fare even worse than capacitive proximity switch due to the high "dust" environment. Inductive relies on conductivity - OK for aluminum, but not OK for "sand".

We've ruled out using scales to weight the barrel or anything of that nature.

As you can tell, I need some advice here. Thanks in advance.

Regards - AMS2750

 

[James Mcquade]

Please be aware that a 55 gallon drum cannot be filled to the very top.
You must allow for the drum to expand and material to expand and contract
with temperature changes. How much must be left depends on the product and moisture content.

please investigate the chemical reactions with the drum, contents, moisture, and the possibility of rust from the drum if they are together for time.

james

 

[AMS2750]

Please be aware that a 55 gallon drum cannot be filled to the very top.
You must allow for the drum to expand and material to expand and contract
with temperature changes. How much must be left depends on the product and moisture content.

please investigate the chemical reactions with the drum, contents, moisture, and the possibility of rust from the drum if they are together for time.

james

I should have specified that 99% full is really just an approximation of our cut-off point at which the barrel still has room, but where the blasting process should stop to prevent any more media from entering the barrel.

As to the chemical compatibility, etc..., we've been using this setup for quite some time. The only thing new, is trying to incorporate automated level detection instead of requiring the operator to periodically lift the lid to determine what the level is.

Regards - AMS2750

 

[Andy6]

Dust like material is tough for most sensors. One thing that would work reliably would be a tuning fork style sensor. Problem is, I could see that getting broken when tossed around during a drum change.
Since you're using unshielded sensors. The only thing I can think of would be to try mounting the sensor in some kind of plastic pipe, mounted a bit lower into the drum. It will really come down to some trial and error to get it working.

 

[Doug-P]

Maybe these folks can help. http://www.binmaster.com/products/info/40-3d/4891-acoustic-level-sensor

 

[Tom Jenkins]

If you have shop air available you could use a hose pointed across the face of the switches. Use a solenoid to periodically turn on the air and blast across the switch face to clean it.

 

[L D[AR2P#0.0]]

Wacky idea - use an adapted ball valve float to actuate your sensors.

 

[DeKemp]

What about a rotary paddle switch? http://www.uk.endress.com/en/Tailor...surement/Paddle-level-switch-Soliswitch-FTE20

 

[shooter]

Yep kemp is best solution, however expensive, you can make it simpler with a coil pulling a float up, when the coil does not fall back the bin is full,
Or more info by a ropefloat with a encoder on top. River levels are measured the same way.

Simplest is to use a full bin on a balans and a switch under the empty bin that will stop filling or vacuum.

 

[AMS2750]

Thank you everyone for the feedback. I've learned a few things from the suggestions. In particular, I had never heard of an Acoustic Level Sensor. However, financial reasons are forcing me to take another look at capacitive sensors as it is the least expensive and simplest setup for the application, so I've decided to pursue one more test using capacitive sensors.

Based on my research and current setup, they should be adequately suited for the application once I reduce the sensitivity to eliminate (reduce?) false positives due to dust caking on the sensor. My biggest problem with that solution has to do with the way I have to mount the sensors.

To make a long story short, I've finally made peace with how to mount them. I have a bit of an OCD streak, and it took me all day to figure out a mounting design that should get the job done (there are some environmental and mechanical challenges to the setup) and appease my OCD.

By the way, the 5-40mm range sensors I couldn't get working right yesterday... well, would you know that 40mm != 40cm. Talk about feeling stupid. No wonder I couldn't get anywhere close to the "expected" range out of those.

I'll have to buy some new sensors, so when they come in, I'll update this post with the results of my tests.