Howdie,

would be great if I could get some tips from the cloud, ahem crowd intelligence.

The problem: I need to measure the total volume and oil fraction of a changing oil/water mixture.

So far, I plan to have an e.g. 1 m3 tank installed with the mixture entering and after e.g. 0.7 m3 of the tank are filled, the total weight is measured (load cells) and based on average densities the oil fraction calculated. The whole design should be ex-proof and the readings transferred to a nearby station.
What I am missing though is the time for filling these 0.7 m3 to extrapolate 24 hrs volumes.

The difficulties:
1. how do I ensure that after 0.7 m3 the inflow is shut?
2. and also that the time is taken so I can extrapolate for 24 hrs?

Additional 1: I could add an overflow at 0.7 m3 to ensure constant volume, but where do I get the time from?
Additional 2: expected mixture volumes are ca. 0.5-2 m3 per hour with 10-20% oil fraction

For all ideas greatly thankful,
Tom

 

Let me get this right, you are batching into and out of the 1m3 tank
Fill to 0.7, close inlet valve, weigh, open discharge valve.

How about using an accurate level transmitter, radar or magneto restrictive to control the feed in. When it reaches 0.7 close the inlet valve. It wouldn't matter if you overshot a bit because it would be taken care of in the Mass/Volume calc.

Instead of actually weighing the tank you could also use a DP cell then you would have Head/Volume. The flow rate would be the number of fills in 24 hours.

If you want to discuss PM me: roy_matson [at] yahoo [dot] ca
Hope this helps,
Roy

 

Sorry, if I wasn't precise enough:

I also need to know the oil fraction, and here I believe it becomes difficult: the densities are fairly close (oil: 0.95, water 1.03 kg/l) and the gravity separation of the fluid within the measurement interval most likely too small to get a clear interface (lots of foam?). Thus, I believe radar or magneto measurements will not work. Also, e.g. resistivity electrodes are gunking up within a short time - hence I'm thinking about the weight/volume method (using average densities) to get the fractions.

In terms of automatic inlet shut off, I guess it would be the simplest to use the load cells to automatically close.

The DP cell, I believe will not work, because I need both total volume and total mass to calculate the fractions (assuming avg. densities). Is there any other simple method to measure the total volume

Any other (radical different) design proposal highly appreciated.

Thanks, Tom

 

Hi Tom,

Foam is a big problem for any level transmitter. As you know radar relies on a change in the dielectric constant to get a good echo. I think it will disregard foam. BTW, how will you get rid of the foam between batches?

DP should give you mass Head x Area

The problem with load cells is isolating your pipework. I have an idea on how to measure fraction which I will share with you on 1:1 basis, please send me a PM at the address in my previous post so we can discuss in detail at length.

Regards,
Roy

 

Don't you need both mass and volume (level) measurements in order to calculate respective proportions of different density liquids? (thinking aloud here . . . )

DP could substitute for a load cell weight/mass measurement, with the head pressure reflecting the densities of the liquids. (hear the grinding? . . . )

If foam is significant enough to create an error big enough that you can't ignore it, then you'll have to come up with something to take the foam into account. Maybe a diverter valve connected to two tanks so that one tank can fill and then idle so that the foam decays into liquid while the 2nd tank picks up the flow in the meantime? That could be messy, filling to some level less than 0.7 and then trickle flow to make-up to exactly 0.7m^3 level?

For the volume-by-level measurement I'd tend toward a simple level float switch installed at the exact level to trip at 0.70m^3 volume. When the level increases and float switch changes state, shut off the inlet valve, make the mass measurement, then empty the tank and resume filling.

1. How do I ensure that after 0.7 m3 the inflow is shut?

Drive a shut-off valve according to the level measurement. There is an inlet valve ? ?

2. and also that the time is taken so I can extrapolate for 24 hrs?

Drive a counter off the float switch (or level measurement trip at 0.7m^3). Count the number of tank fills over a period of time.

 

Tom

we normally install "COALESCER" , between 2 tanks which gives demarked separation & eliminates foam than any of the measurement concept you would like to implement would work perfectly.

Please note that you can install coalescer in the second tank itself. You can use Radar type xtrs co-axial type & should serve your purpose.

Jari
iconcnl [at] vsnl.net