Tank and pressure measurement

I have a tank where they (other company) placed a 0-6 Bar Pressure transmitter. The Tank can have 20000 liters of butter max. There is a tracing around the tank for warming up the tank so that the butter stays a little liquid. The density of butter is 0,86 kg/dm3.

The Pressure transmitter gives a signal from 4 to 20mA to an S7 PLC.

So 4mA = 0Bar
And 20mA = 6 Bar

When they delivered 5000 liters butter, I had a signal of 4.741mA
So this gives us:
Range = 20mA - 4mA offset = 16mA
Signal = 4,741mA - offset = 0,741mA
(0,741mA * 16mA) / 6 bar = 0,278Bar, or 278mBar

Because of the range 0-6 Bar we have problems for showing the right volume in the Tank.

When they have 1000 liters remaining in the tank for example, then 0,741mA / 5 = 0,1482 mA + offset = 4,1482mA.

So I think they should place a pressure transmitter with a lower rangen, like 0-1 Bar for example. Or measuring ultrasonic from upstairs. De density of butter in combination with the transmitter 0-6 Bars is giving probs right now. When they should see 3000 liters, this can also be 1000 liters or 4500 liters...

Any suggestions on what I can do ?
Is the Pressure transmitter usable ?
Or is the fault more a problem because of the tank not being lineair (filled and allmost empty on a certain curve) ?

Please help,

Thanks in forward,

Regards,
G

Or is the fault more a problem because of the tank not being lineair (filled and allmost empty on a certain curve) ?

Click to expand...

Yes this will give you incorrect readings.

Dont mess around with the PT in this application. The Ultrasonic is the way to go.

ok

Okay, but I'm not the guy who placed a pressuere transmitter.
Is pressure more for liquid products like paint and so... ?

RheinhardtP said:

Yes this will give you incorrect readings.

Dont mess around with the PT in this application. The Ultrasonic is the way to go.

Click to expand...

What is the shape of the tank?

Obviously the Span of the PT is too big. Most PT's can be easily re-spanned in the field. Consult the user manual and contact the mfg. to check that the sensor part of the PT will be accurate of the range you want to use.

Is the PT tapped into the bottom end of the tank ? If so then you may need to calculate the "bottom head" volume to give an accurate reading. I just use the triangle shape equivalent and that's close enough.

Another thing is the density vs temperature. If this changes much, you may need a TT (temperature transmitter) and calculations.

All in all, a ultrasonic sensor is better, but I understand that your hands may be tied.

hello

hello,

little drawing

Bob O said:

What is the shape of the tank?

Click to expand...

Would seem to me a presure transducter would be a nightmare on this aplication. Just the material density changing which I think with butter would be fairly common would cause a bad reading. Also the fact that butter can solidfy so easily would be an issue as well.



I would push the ultra sonic as hard as possible or a weighing system. The ultra sonic would be the cheaper of the 2.

Pressure Measured by Transmitter = 9.81 x density x Height (h) of Liquid in Tank.

For a measurement of 6 bar this would imply a tank level of:
h = 6 x 100 000 / 9.81 x 860 = 71.1 metres.

If your tank maximum level is less than this then you need to re-range the pressure transmitter accordingly.

I would have expected the quality of your butter to be such that the density remains fairly constant so that a pressure transmitter would be OK for the application. As you can see from the formula above your transmitter output will vary whenever the density changes. If variations in density are small then small variations in measurement may be acceptable.

The transmitter body may need to be heat traced so as not to get clogged up. However not all transmitters can handle heat tracing. A transmitter with the diaphragm flanged to the tank directly rather than connected via small bore tubing would be the better type to use here.

A better option may be to use some form of non-contact measurement as already mentioned by others if your finances permit. This could be ultrasonic, radar etc.

The inverted cone at the bottom of your tank should be out of your range of measurement (i.e. minimum level should be at / above the top of the cone) to avoid problems with linearity of measurement at low tank levels.

I'm not sure ultrasonic etc. are the answer, as condensation of fats oils and greases offgassed by the butter could coat the face and screw up readings.

What is your objective? If it is custody transfer ($ involved) your system isn't accurate enough. If you want to know when to order more butter you're probably just fine. If you are measuring output for a recipe, maybe OK, maybe not.

I agree that a diaphragm seal is the way to go, and you should ideally have a transmiter that is at about 75% full scale when your tank is full.

This is a simple geometry problem. Calistodwt gave you the formula for height, so you need to go back to fundamentals for the volume. You need a two part formula, one for the conical bottom, and when the level gets above that then one for a cylinder.

To start with you have the wrong sensor, you should be using a level sensor such as an Anderson SL http://www.andinst.com/dairy-level-products.php These level sensors have temperature compensation. Because of the non linear section of the tank, you will need some sort of look up table in the PLC code. What I have done in the past is meter in water into the tank and record how many litres accumulated for every " of water column the level sensor measured until the level reached the linear section of the tank. For each increment in " I have a 'K' factor . Then what I do is simple lookup table in PLC so that between 0 - 1" use a certain K factor , K[1], 1- 2" K[2], that is applied against the reading from the sensor until the linear section of the tank is reached.

Websites such as this http://www.wemactanks.com/tankcalc.htm help with conversion of " of water column to USGallons (1 USG = 3.8ltr)

What is volume of conical section of tank vs the "right cylinder" ?? section of tank. If a 100:1 ratio I would move the pressure sensor and worry about the right cylinder section only where I have linear relationship. To protect the pump from running dry if needed I would put float switch at bottom of tank.

The only remaining issue is to correct density for temperature. I think I would use minimum operating temperature as a constant and control at above this temperature. I think this should be most conservative (unless I have my brain in reverse).

Maybe if you could find one a detector more closely matched to your tank would be good idea.

For sure I would not rely on this as an accurate method for making batches and would have separate weighing / measurement method.

Dan Bentler

I am also in the Dairy industry but only handle Milk & Cream, we use Endress & Hauser instrumentation which with the correct software (Endress & Hauser Fieldcare - i think??) you can place the transmitter into a teach mode where you can fill the tank with fluid (or butter in your case) and even if the tank is non-linear then the device compensates for this.
The pressure/ level transmitters we use are Deltapilot FMB70.

The problem with the base of the tank being conical (for hygiene puposes) we get around by having a low level sensor which is a Liquiphant M ("tuning fork sensor"), as fluid makes this sensor a fixed value e.g. 600Litres is added and whatever the level transmitter picks up is added to this.

Hope this helps,

LH

Load cells plus 2 capacitive probes for high and high high levels.

put the tank on loadcells always give accurate readings.
you could use the pressure sensor with a few cylinders as feet for the tank.

muusic_man said:

Load cells plus 2 capacitive probes for high and high high levels.

Click to expand...

I concur. Load cells would give you the quantity in a more reliable form. We have had great success with Emery Winslow load cells in our very demanding environment.