Scaling question

I work with quite a bit of dairy. whether it be sour cream, milk (skim, whole, 2%, choc. etc.), ice cream you name it. Currently all of the silos and tanks are using a 4-20mA pressure sensor to detect amount of product in tank. They're all currently scaled the same regardless of product, however specific gravity for each product differs. I've been tasked with scaling the value by each product's specific gravity.

My thoughts on doing this:

1. grab product code of product in tank
2. use correct specific gravity depending on product and multiply the scaled output (gallons) by specific gravity.

Or

Should I be multiplying the counts directly from the pressure sensor by the specific gravity and then spit it into the scaling blocks? I feel the difference won't be be that great if at all. Proper way seems like multiplying the counts but I wanted to run it by here first.

Regards,
Phil

How about just using a level sensor? Then you dont have to worry about the specific gravity.

I'm using an Anderson SL5 Level sensor.

Excuse my ignorance, and maybe I'm misunderstanding the application,but are they asking you to calculate the weight using the level and the specific gravity?

MartB said:

Excuse my ignorance, and maybe I'm misunderstanding the application,but are they asking you to calculate the weight using the level and the specific gravity?

Click to expand...

No, I'll try and shed more light on it. Using a level sensor to give amount of gallons in given tank. I realize this sounds straightforward, from what I gather the Anderson SL line of level sensors use a 4-20mA signal as output depending on height in inches of the water column above the sensor. This means they should be factoring in the specific gravity of the water in their choice for materials etc when designing the sensor.

Since I am not using water or a product of identical specific gravity will this not change what the sensor "thinks" the amount of inches above the sensor? Resulting in a different output. That's at least how I interpret the Sensors input to output. If there is somebody who can confirm/explain correctly please feel free to. I would greatly appreciate it.

Regards,
Phil

Your understanding of how the pressure sensors work is correct. You can, indeed, use a different scaling factor for each product. That's how we do it.

HTST operators assign a product (Skim, 1%, 2%, etc.) to a particular tank, which then sets the appropriate scaling factor.

For instance, 12mA may mean there's 10,000gal of water in a tank, but only 19,850gal of Skim, while 20mA = 20,000gal of water and 39,700gal of Skim.

Using SCP function (Assuming you're using AB),

Input = The pnysical input your level sender is connected to
Input min = Analog equivalent of 12mA
Input max = Analog equivalent of 20mA
Scaled minimum = 10,000
Scaled maximum = 20,000
Output = Some Integer (let's say N7:0)

If you know the product has a specific gravity of 0.9925, multiply the value in the integer by that number and store it in another memory location.

You can take this a step further and re-calibrate the sensor every time the tank is used, too. Here's what I do:

• Monitor the input from the pressure sensor. When it rises above a minimum value, use a one-shot to store the metered volume from the HTST to a memory location assigned to "Scaled Minimum"
• Monitor the outlet valve of the tank. As soon as it opens, MOV the analog value of the sensor to "Input Max" and the metered volume from the HTST to "Scaled Maximum"

Bear in mind that I'm assuming you are storing your milk in vertical silos. Also, be aware that the pressure sensors are usually some distance from the bottom of the silo, which means the first analog value above minimum the sensor reports isn't zero.

If all of the operations in your scaling are multiplications the order of operation doesn't matter.

See this for more info:
http://www.engineeringtoolbox.com/static-pressure-head-d_610.html

In you case p2 is always 0 unless your vessels are pressurized.

How about just using a level sensor? Then you dont have to worry about the specific gravity.

Click to expand...

I think that is what he is using. but it happens to be a level gage that uses pressure to measure the level, a very commom method where the liquid does not need to directly contact the measuring device.

Since I am not using water or a product of identical specific gravity will this not change what the sensor "thinks" the amount of inches above the sensor?

Click to expand...

Yes, it will. I understand that your level sensor is a pressure device calibrated to read level. This is a common method and I have used many Rosemont pressure-transmitter-level-gages for that same purpose.

I think all you need to do is multiply your pressure reading by the Specific Gravity of Water / Specific Gravity of Liquid. In other words, if your liquid is heavier than water (S.G. = 1.1), it will produce a higher pressure at the same height, so you need to reduce the level reading by 1/S.G.Liquid = 1/1.1 = 0.909. If it is lighter than water (say S.G. = 0.8), your level reading would be increased by 1/0.8 or 1.25.

Thank you all for the replies,

I just wanted to make sure my findings I did prior to posting could be solidified.

You can take this a step further and re-calibrate the sensor every time the tank is used, too. Here's what I do:
Monitor the input from the pressure sensor. When it rises above a minimum value, use a one-shot to store the metered volume from the HTST to a memory location assigned to "Scaled Minimum"
Monitor the outlet valve of the tank. As soon as it opens, MOV the analog value of the sensor to "Input Max" and the metered volume from the HTST to "Scaled Maximum"

Click to expand...

This is something I hadn't thought of I may have to tinker with this, thanks for posting it!

Phil

Important: divide, not multiply

One divides the hydrostatic level reading by the fluid's specific gravity.

A lighter-than-water fluid with an SG of 0.98 at a level of 100" will give a reading of 98.0"w.c.

To get the corrected indicated level of 100", divide the reading 98.0" by the SG 0.98.

If the level is already converted to gallons, likewise, divide by the specific gravity.

In the end it's just another arithmetic operation, so I don't see whether it makes a difference when it done.

One divides the hydrostatic level reading by the fluid's specific gravity.

Click to expand...

Or multiply the level by 1/S.G. In the end it's just another arithmetic operation!

If you have many products (skim, whole, 2%, choc. etc.) and sometimes the same product might not always have the same density, you probably should think about replacing the pressure transmitters with Radar Transmitters.

danw said:

Important: divide, not multiply

One divides the hydrostatic level reading by the fluid's specific gravity.

A lighter-than-water fluid with an SG of 0.98 at a level of 100" will give a reading of 98.0"w.c.

To get the corrected indicated level of 100", divide the reading 98.0" by the SG 0.98.

If the level is already converted to gallons, likewise, divide by the specific gravity.

In the end it's just another arithmetic operation, so I don't see whether it makes a difference when it done.

Click to expand...

That's right - DIVIDE, don't MULTIPLY (as I stated earlier). Thanks for pointing that out.

Strokur said:

Thank you all for the replies,

I just wanted to make sure my findings I did prior to posting could be solidified.



This is something I hadn't thought of I may have to tinker with this, thanks for posting it!

Phil

Click to expand...

This certainly solves the problems that could be encountered if you have products of many (or variable) densities that could be in a particular tank.

I have seen this problem solved with a "gauging" level-probe in the vessel. When filling, the level reading is stored when the low-level probe gets covered, and stored again when the "gauging" probe gets covered.

Since the exact volume between the probes can be found (i.e. fill it with water), a bit of simple math can automatically determine the density of the product, and use that factor until the tank is emptied again. That deals with homogenous products of varying density.

The higher the gauging probe is, the better accuracy, but the longer you will need to wait until the level reading is adjusted.

One downfall of this method is that the product needs to be of fairly consistent density throughout the filling. If there is any likelihood of a non-homogenous mix, then it doesn't work too well.