Liquid Level Measurement in Agitated Tanks

Your customer wants ±1mm accuracy in an agitated tank? That's completely crazy.

Tell me about it! Well, the radar guys have software that takes into account the readings from the agitated mass and gives a reading within +/- 3 mm. But that's all they are willing to guarantee in agitated as well as in still conditions.

We could give a SOP to the customer that says - stop the agitation and uses this gizmo - you get the level with accuracy of +/- 1 mm. Question - what gizmo?

Alternately I was thinking if we can measure very accurately the flows into the vessel (coriolis flow meter? don't know any thing about it - just looked it up on the net).

Load cells and a density transmitter seems to be the way to go. Not knowing the liquid, the temperature or the pressure prevents further suggestions being made.

Line diameters and flowrates would be valuable too.

Are these vented tanks?

Is there an inert gas blanket?

+/- 1mm? At what temperature swing and what density swing?

Define <...corrossive liquids...>. All liquids are corrosive to some materials; one wouldn't put hot water in a vessel made of sucrose, for example.

Guru,

Right! alwayes hated using hot water in those fancy sucrose vessels!

These are vessels in SS or glasslined

The operating parameters :

Temp - Room temperture ( 30 deg C) ro max 80 degree C
Operating pressure - 3 bar
Process constituents - solvents + WFI + product mass

Ph - mildly Alkaline: in one case acidic where it is a glasslined vessel

Do not have info on density mass but let me know how it could be relevant and I will ask.

There is no nitrogen blanketing, tanks are not vented to atmosphere.

Terrific! Thanks a lot. ...

You have not defined "agitation".

HI PWSlack,

A bit more on your recomendations: "Don't use ultrasonic if there is froth or an agitated surface, or a risk of droplets condensing on the sensor or if the pressure is not ambient."

Question is why? Maybe you have experience in using ultrasonic sensors but I wanted to check with you a device I located on the net - www.solidat.com. One of the products featured, GaugerGSM claims .."Embedded algorithms extract false echoes, compensate for temperature and handle turbulent targets".

Would also appreciate letting me know why pressure different from ambient could give false readings? I can guess droplets interfering with the reading of the reflected waves...

Thanks again

<Resubscribes>

<...recomendations: "Don't use ultrasonic if there is froth or an agitated surface, or a risk of droplets condensing on the sensor or if the pressure is not ambient."...>

'Cos it won't work!

<unsubscribes>

Thanks Iris. This is good. Sartorius is the gold standard for load cells and if we can use a density meter as suggested by PW Slack it may be good enough....

Dear Mr.Ash55,

As I understand from your statement, agitating vessel means LEVEL CANNOT BE MEASURED CORRECTLY.

If so, you can provide a gauge glass, outside the vessel, like as in Boiler, where the liquid will reach its level and stay undisturbed, and level sensor can be clubbed to this. Perhaps some working difficulty may come and to be overcome.

This applies to cylindrical vessel only -vertically mounted. Let us see responses from CR4 Members.

DHAYANANDHAN.S

If you cannot measure it while it is being agitated, measure it when it is not being agitated. Meter the liquid as it put into, and taken out of the tank, and the difference is what is in the tank. You will need positive displacement flowmeters to get that sort of accuracy.

Dear Mr.jhhassociates

Your view may not work for contiuous flow and agitation. It will be OK for batch type filling, drain/discharge after agitation.

Pl. let us know from you on the continuos flow/agitation.

DHAYANANADHAN.S

Hi Ash,

I work in a facility that uses agitated tanks between 250L and 10,000L we use load cells and a calibrated sight tube on the outside of the tank. This works for everything from WFI to Buffers containing ethanol / NaOH etc.

Aslong as the inprocess testing is adequate to prove the contents are the correct concentration / ratio etc then there should be no problems.

Hi only1yeti,

Yes, there are graduated view glasses on the side. Two problems with it:

1. It cant show the level all the way down and up (sight glasses on side of vessel as built now).

2. Even if such extreme levels are not required to be met, the customer needs the data to be electronically be available from two sources and verified electronically (CFR 21). So we will need the level markings to be picked up and converted to a level electronically.....

Thanks, that's a very interesting piece of information that we can use when discussing with the customer... do you know where I could get hold of the relevant extract that allows it?

try googling the Eudralex (european Drug / Medical Device GMP requirements) that is prob a good place to start in a highly regulated industry like this.

for the rules of CFR 21 part 11 go here --> http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?CFRPart=11

for the Eudralex material --> http://ec.europa.eu/health/documents/eudralex/ is your best bet.

You know if there was some way the calibrated sight glass on the side of the vessel could be read and validated with the load cell readings electronically, we may have a solution.

I was actually thinking of having webcams that will take a shot of the level through the calibrated sight glass as a way of validating the load cell reading. So, although the validation is not auto, there is an electronic record of the time when the load cell reading was done and the cam shot.

Got to talk to the customer in a couple of days with some solutions.

What do you think -dumb idea?

Have a glass level gauge beside the tank. Measure level here. The level averages out nicely. You are likely to get much better reading and tolerance.

Here is a link to ER guided wave tank level monitoring that is at .5mm accuracy.

We use the 5300 unit in our agitated tanks and have had excellent results with no failures since installation 5 years ago.

It is critical to secure the immersed probe so that it cannot get tangled up in the agitator.

Is this a batch-fill process? If so, you can get your second measurement from a flowmeter with an appropriate totalizer. This would also exhibit the advantage of representing liquid level in the same units as the load cell system (i.e. true mass).

If you can have a side tube at the bottom, the agitation will be damped and the level will be steady at the average value.. Measure this level using a suitable device.

Perhaps a radiation gauge may work as it can sense total material under the path and for averaging you can have larger oppurture and sensing area.

>I was actually thinking of having webcams that will take a shot of the level through the calibrated sight glass as a way of validating the load cell reading. So, although the validation is not auto, there is an electronic record of the time when the load cell reading was done and the cam shot.

caveat emptor: A US pharmaceutical advised me that they could not validate bitmap screen shot images of recorded trend data on a CIP cleaning process. The reason - Photoshop - it's possible to alter an image for fraudulent purposes.

I suspect that video files are as subject to editing as static images. You might make inquiries as to what it takes to validate a video file if you pursue that route.

I agree Validating unsecure records while trying for CFR 21 compliance can be a minefield.

one thing we do onsite for various pieces of equipment is to print and sign results from balances etc so with load cells an auto output to a printer may work. The other route to go down would possibly look to integrating it into a LIMS (Laboratory Information Management System) if the company already uses one as they tend to already be CFR compliant.