Level Measurement problem

Hi. I'm curious as to how to measure, reasonably accurately,
the level in a silo for a range of liquids of different densities. The instrument being used is a pressure sensor at the
base of the silo. (10 meters of head = 1 bar {Water})
I take the 4-20mA signal and scale it (0 - 45000 Ltrs). The problem is that inaccuracy is occurring as the range of liquids the silo holds vary in density (Specific gravity from 0.9 to 1.2 compared to water which is 1.0).
Basically all i need to know is how to go about calculatiing this. I know there are instruments out there that
handle this kind of scenario but I have to stick with what I got.Recession and all that! The silo is cooled so temperature does not vary much. Any help would be appreciated.

Thanks in advance

You have to know the liquid's specific gravity (SG) and provide the SG value to the PLC for its correction calculation (which, of course, changes for each material)

1 bar (water) = 10.00 m (water) level

formula: measured head pressure value of water (bar) / SG = physical level (meters)

for SG = 1.1: measured 1.000 bar / 1.1 = 0.9090 m physical level
for SG = 0.90: measured 1.000 bar / 0.90 = 1.111 m physical level


Dan

danw said:

You have to know the liquid's specific gravity (SG) and provide the SG value to the PLC for its correction calculation (which, of course, changes for each material)

1 bar (water) = 10.00 m (water) level

formula: measured head pressure value of water (bar) / SG = physical level (meters)

for SG = 1.1: measured 1.000 bar / 1.1 = 0.9090 m physical level
for SG = 0.90: measured 1.000 bar / 0.90 = 1.111 m physical level


Dan

Click to expand...

I agree with Dan.

Endress + Hauser have just introduced an on-line SG measurement instrument based up their Liquiphant vibrating fork level switches supplemented by temperature measurement and a fancy controller. You could take the output from this instrument into you existing control system and continuously correct for changes in SG using Dan's formula. However this system isn't cheap (lots of 1,000s of Euros)and you would end up with two instruments to support. And they are a real pick to setup. Therefore I would suggest you would be better swapping to an alternative level transmitter that uses Time of Flight as measureing principle (e.g. ultrasonic or radar) as these will be independant of changes in SG. Depending on what outputs and display you need you should be able to source something for less than a 1,000 Euro.

Best regards,
Corby

Thanks Guys. I have got my hands on the SG values for the different products I am using and should be able to use the calculation Dan suggested. I think the Radar solution is probably
the way to go but I have 10 silos so cost becomes an issue.

Thanks again guys.

We use guided microwave sensors onsite, by Vega (similar to the radar ones, work on time of relection). Work well, and get you round all the problems of SG. Like a lot of these things initial cost will be higher, but life time cost will probably be lower as there will less issues with them.

Is the liquid to each silo always the same SG, or does the liquid vary in each silo? If it is the same, only one coversion for each silo is needed.

However, if they vary, and you also have control of the distribution of liquids to the silos, you could then use the distribuiton setup to substitute the correct correction when the silo is filled.

I've had very good success using ultrasonic level transmitters in 6 meter tanks. They may require stilling wells to be accurate though. If you decide to go that route, a little research can prevent some problems with implementation.

Whoops, my stated physical levels are incorrect, off by a factor of 10 (an order of magnitude).
1 bar = 10m water, not 1m water

should be:

for SG = 1.1: measured 1.000 bar / 1.1 = 9.090 m physical level
for SG = 0.90: measured 1.000 bar / 0.90 = 11.11 m physical level

Dan

There are around 30 different SG liquids which can occupy any of the silos at any one time. Jim, by "distribution to the silos" you mean maybe applying the calculation to the flowmeter on the fill line i assume.
While this does make sense I need to sort the level in the silos first as it gets complicated with the array of flow meters filling the silos.
I am doing some homework on radar and other non-SG specific methods of level measurement. Can you explain what you mean by stilling wells Jim?

Thanks for the update Dan also

If you want to find the SG of the liquid in the silos, all you have to do is to install another "cheap" pressure tansducer at a known distance from the transmitter (0.5mts)already installed at the bottom of the silos. After both transmitters are under the liquid, you can read the difference between the two transmitters, and the value you read is the SG. You can use this value to scale the level from the bottom transmitter.

ThePortLion said:

There are around 30 different SG liquids which can occupy any of the silos at any one time. Jim, by "distribution to the silos" you mean maybe applying the calculation to the flowmeter on the fill line i assume.
While this does make sense I need to sort the level in the silos first as it gets complicated with the array of flow meters filling the silos.
I am doing some homework on radar and other non-SG specific methods of level measurement. Can you explain what you mean by stilling wells Jim?

Thanks for the update Dan also

Click to expand...

What I meant by distribution was that if you are also controlling the valving and pumping to/from the silos and also know the liquid being moved by that particular valving setup, you would also know the SG of the liquid being pumped to that silo. Using a lookup chart, you could then apply the correct SG, or actually the correct correction to the raw pressure reading for use in the scaling function for that silo.

Stilling well is usually a pipe, sometimes with perforations, inserted from the top of the tank/silo with a mounting flange for the level transducer. It's purpose is to provide am udisturbed (or "still") surface target for the ultrasonic waves to reflect from. It will prevent false readings from agitation, foam, etc. It also must be large enough in diameter to provide proper operation of the transmitter - the manufacturer will usually be able to provide that info. I mentioned it because it must be figured into the cost of fitting some transmitters, and can be a factor.

1)

nonuke said:

If you want to find the SG of the liquid in the silos, all you have to do is to install another "cheap" pressure tansducer at a known distance from the transmitter (0.5mts)already installed at the bottom of the silos. After both transmitters are under the liquid, you can read the difference between the two transmitters, and the value you read is the SG. You can use this value to scale the level from the bottom transmitter.

Click to expand...

This is known as 'hydrostatic tank gauging'.

Given that it involves installation of a 2nd transmitter at a precise elevation above the lower transmitter, most existing facilities do not see the advantage over changing to a top mounted technology, like radar, ultraonic, capacitance, whatever.

2) Stilling wells
Stilling wells are needed in some very turbulent mixing tanks, but have the problem that the non-contact radar and ultrasonic technologies 'beams' will bounce or echo off residual material on the inside of the stilling well.

If there's a reason to use a stilling well on a bulk storage tank I can't think of it at the moment.

3) radar & ultraonsic
both technologies are now 4th to 6th generation and can work very well when chosen properly and installed properly. Where and how they're mounted makes a big difference in performance. If you're serious, you really need to work with a vendor to pick the right technology for the materials and tanks you've got.

Dan

Thanks Guys for all the info. I will, for now, just go with applying the SG for a particular product to the raw pressure reading. I forgot to mention that the operator enters a product
tag to the silo on the Scada when it has filled, so I can then modify the raw pressure reading to suit. The solution with the second pressure gauge is an interesting one especially now when
cost is more important than ever.
I have lots of questions about the radar/ultrasonic methods. e.g. the silos are conical at the base, so how effective
would these methods be at low levels... I'm not starting a new thread on this, it's just one of the many questions i'll have to get answered.
Anyways Thanks again for all your help.

ThePortLion, your concerns using radar or ultrasonic measurement with conical tank bottoms equally apply to using a pressure transmitter in the bottom of the silo.

This measures "head" or height of the liquid, just as a radar/ultrasonic ranger does - so you would need to linearise the reading at the bottom end of the silo in either case.

Also you have not stated if your vessels will have any top pressure, if so, you will need a second pressure transmitter at the top to offset the additional pressure on the bottom sensor.

Remember that the volume in the conical changes exponentially (cubic), because the volume of a cone is given by the formula ((4/3) * pi * r^3).

At a cone angle of 90°, for example, the volume goes up by the following factors

at 1 unit depth, r=1 : 1^3 = 1
at 1 unit depth, r=2 : 2^3 = 8
at 1 unit depth, r=3 : 3^3 = 27

etc.

Accurate silo dimensional data would be required, or for best accuracy, fill each silo through a flowmeter and read off the head pressure at specific volumes entered. This is a time consuming, and therefore costly, commissioning task, factor this into your costings.

Very often, this sort of scenario is avoided by using flowmeters in and out of the silos, and determining volume in each silo by accumulating + volume in, and - volume out. You already have "Input" flowmeters, you said.

An enhancement to that technique often uses a "guaging" level probe installed in each silo, and correcting the empirical volume data as the probe gets covered/uncovered. The guauging probe level would need to be determined once per silo by filling it through one of your flowmeters.

I've also worked on systems using load cells on silos/tanks, measuring the fill volume by weight and adjusting for SG. This can work well with complicated shape storage vessels, or for vessels with "inclusions", which produce many shapes of linearisation curves.

You have many choices to make. Each method has it's merits, and cost overheads, it all depends on your desired accuracy and repeatability.