5 replies to this topic

[riven]

All; attached is a PnID of the affected area that I wish to discuss.

It contains
1. A pressure vessel D-01; max temperature 200C, max operating pressure 25 barg, volume 1.3L, contains: 95% ethanol 5% water (for the purpose of this discussion)
2. A heating/stirring plate manually controlled via TT-01 and TC-01
3. A safety relief valve SV-02 set at 25 barg
4. A level control system containing LT-01, LC-01 and P-01.
5. A feed sampling system at V-10 whereby a dead volume can be removed.
Inside the pressure vessel is a membrane which can selectivity remove water. On the permeate side of the membrane, a vacuum is held (approx 10 mbara).
For the purposes of the discussion we ignore the other items.

The process is to fill the vessel to a safe level and heat it up using X-01. Upon heating the liquid will expand and the pressure will rise. This process will keep the liquid in the same phase as the temperature increases beyond the boiling point of the liquid.

What confuses me is that in the event of a large level drop via most likely operator error at V-10, failure at V-10, membrane breakage, seal failure on top flange, level control system failure, or other (ideas appreciated), this could result in a large amount of liquid flashing off to produce an unsafe pressure rise inside the vessel (This has not been considered in the Hazop so assume SV-02 would not relieve the pressure quick enough; I can calculate if this is the case). Am I correct?

I stress that I only want a qualitative discussion as if I enter calculations I must produce some. Any and all comments appreciated.

Attached Files

•  Doc5.doc   87KB   44 downloads

[shan]

It is true that liquid level drop will result more vapor flashed out. However, vapor flashing is because of lower pressure not higher. Therefore, there will be no such thing “unsafe pressure rise” at all.

[riven]

It is true that liquid level drop will result more vapor flashed out. However, vapor flashing is because of lower pressure not higher. Therefore, there will be no such thing “unsafe pressure rise” at all.

The associated level drop will cause the volume of the liquid and hence the pressure in the container to decrease. The liquid will still be above its boiling point and thus will dropping pressure and constant temperatue, it can flash. Is this realistic?

Edited by riven, 16 September 2009 - 07:52 AM.

[joerd]

The associated level drop will cause the volume of the liquid and hence the pressure in the container to decrease. The liquid will still be above its boiling point and thus will dropping pressure and constant temperatue, it can flash. Is this realistic?

That's right, the liquid will flash. The amount of flashing will be such that the pressure becomes equal to the equilibrium vapor pressure again, which depends on temperature - and since the temperature didn't change, the "final" pressure doesn't change, and is the same as before the event.

[shan]

It is true that liquid level drop will result more vapor flashed out. However, vapor flashing is because of lower pressure not higher. Therefore, there will be no such thing “unsafe pressure rise” at all.

The associated level drop will cause the volume of the liquid and hence the pressure in the container to decrease. The liquid will still be above its boiling point and thus will dropping pressure and constant temperatue, it can flash. Is this realistic?

My comment is based on the scenario that you presented. When the gas volume increases due to liquid volume drop, the pressure has to decrease according to the gas law P1V1=P2V2. The lower pressure will result more liquid flashing to reach the equilibrium at the constant temperature. However, it is impossible to have “unsafe pressure rise”. The higher pressure will result more vapor condensing not flashing.

[riven]

Thanks for the input. It has cleared the issue in my head at least.

My path from here will be to turne the level control to ensure there is always liquid in the vessel (vital for operation) and in the case if leak to external atmosphere, a BLEVE does not occur.