Hi

We have many DPTs installed as flow transmitters along with orifices on pump discharge lines @ liquid ammonia service. The impulse lines are filled with glycol or some other material to create a wet leg and not to allow liquid ammonia to enter the impulse lines. But what we have observed, that with the passage of time, this wetleg liquid gets carried over and disappears. I have a feeling that since these DPTs are installed on pump discharge lines, therefore the venture effect is created whenever the pump is pumping the liquid ammonia. This is causing the wetleg liquid to get carried over with the process fluid.

So if this is the case, what engineering modification/practice you can suggest to solve this problem?

Thanks in advance!

 

Routine refilling of all such wet legs is commonly practiced in such cases.

 

Why did you use the buffer solutions? The SS impulse lines are very resistive to ammonia corrosion so I would get rid of all the buffer solutions and let the ammonia fill the impulse lines.

Also your theory of the venturi effect does not hold any water. Try and suck the liquid out of straw when the bottom end is blocked by your finger. On an impulse line you get the same effect in that if you have suction against a dead end in a tube, you will not get the liquid out of the tube. This can also be seen in vacuum level application when we do wet legs on them.

What I think is happening is similar to what we have found in the past when we have tried to use glycol as a buffer in DP level applications. The process liquid and buffer start to mix and after a while there is no more glycol in the impulse line, only water. Since ammonia is soluble with a lot of liquids, especially water and possibly glycol as well this could be what is causing this. Anyway like I said SS is great as a metal to use if you have concerns w.r.t corrosion.

If you still want to use these buffers your buffer should be a liquid that is first of all heavier than the 6,3 sg of the liquid ammonia and the two liquids must never be able to mix under any circumstances. Then the buffer liquid will work.

 

Thanks for the reply. The reason we are using buffer solution is because we are measuring flow of liquid ammonia. Therefore the impulse lines are downward (as liquid ammonia is there). So do you think that the vaporizing effect of ammonia liquid will cause the problem?

Regards