PSV Sizing: Double Jeopardy?? 1

[Canoman]

When sizing a PSV, there are a number of cases that one must consider. Blocked exit is just one of them. I am looking at the blocked exit case for sizing a relief valve on a trayed amine absorber. The feed is a header system in which a few streams combine before flowing into the absorber. Of the 5 or 6 feed streams, only two of these are at a higher pressure than the vessel is designed for (100psig). We're taking into account a 10% overpressure, and these two streams are at a higher pressure than the overpressure also.

The lean amine pump has a deadhead pressure of around 235 psig. It's normal discharge pressure is pretty high because the lean amine travels around 3000ft of straight pipe distance, so when you add an estimate for frictional losses through fittings, it's a pretty long equivalent length. Add 3 filters and an air cooler into the line, and there's considerable pressure drop to overcome.

So, we have two sources of possible overpressure due to a blocked exit. When designing for a blocked exit case, would it be considered double jeopardy to consider both the vapor outlet and the liquid outlet to be blocked? Do you normally consider all exits being blocked for this relieving case? If I assume that all exits are blocked, then eventually, the column will fill to the top with liquid, and there will be both liquid and vapor relieving. This will, however, take a while. Due to the increased destination pressure, the flow of lean amine will decrease as you change positions on the pump curve. The column is around 60'-0" tall and 5'-0" in diameter, so it would take a while of relieving vapors before liquid got to the top of the column. I'm not even sure that I have to consider both happening at the same time.

Any thoughts would be greatly appreciated.

Thanks.

Canoman

 

[SooCS]

Before you can blocked vapor and liquid outlets is double jeopardy you should check to make sure there is no posibility of both the outlets can be closed simultaneously. Things that you should check;

1. lost of instrument air to the system, does both outlet lines share the common instrument air supply?
2. computer control system, what if it lost computer signal from the computer control or shutdown system, will both the outlet line be closed.
3. what if the vapor outlet is off spec or vapor downstream equipment fails, does it shutdown both the outlets.
4. what if the amine regen system fails, does both the outlets required to be shutin.
5. what if gas/fire is detected in the area, does it require to shutin both the outlets.

I hope you get the idea. But I do recommend you is to do an HAZOP with your design engineers and operators to answer your concern.

 

[CHD01]

Double Jeopardy could apply, check your system as suggested.

Columns such as this could have two relief valves; one for liquid case installed at the bottom of the column; and the other for vapor at the top of the column. If so, remember that the liquid head will ADD to the set point of the valve.

Two valve are used if there are two relief scenarios differing greatly in flow that would cause a problem for just one valve; and also if the column is susceptible to pluggage in trays, etc.


The more you learn, the less you are certain of.

 

[ChemE001]

You may be looking at a multiple failure if your scenario does not cause both outlets to be blocked. Do check your system as suggested above. You only have to consider the worst case scenario if only one outlet is blocked per scenario. Find which one gives the largest required orifice size.

However, you may have to look at two-phase flow in this case. If the column can become liquid full and you have a vapor release at the top of the column then two-phase is likely. You can perform a two-phae flow test using the vapor volume fraction,"a", the vapor and liquid density and surface tension, the vessel volume and cross-section to find the bubble rise and superficial velocity. Then based on the flow regime you suspect either bubbly or churn-turbulent you must check the flow model vapor-liquid disengagement,Yr, with the actual vapor-liquid disengagment,Y = .

Get the Free Mach II Reader and Regression Tool and look at the Two-phase flow tab. Choose two-phase flow test and run the two-phase flow test. You will also find in the DIERS folder a two-phase flow example taken from the literature by Harlod Fisher the chairman of the DIERS User Group. To get the Free Mach II Reader and Regression Tool to to

http://wwww.digitalsolutions.org

and choose downloads, programs and Mach II with installer. You will download the Mach II Reader and Regression Tool along with a folder of over 200 DIERS benchmarks examples for single and two-phase flow design.

Happy hunting.

 

[Montemayor]

Canoman (Chemical):

I highly recommend you follow SooCS advise and hold a "mini" HazOp with your Operators and at least one control systems engineer. I get the strong impression you haven't gotten the whole thrust of the Acid Gas Removal Unit's operation under your belt yet. For example, without any further information about your process or design, I can foresee that you would probably not have a valid case in worrying about the Lean Amine being blocked at the absorber outlet. I also can see that you are being confused by your Lean Amine Pump service. Let me explain why:

1) As in any Acid Gas Removal unit, this one probably has a closed Amine solution loop. In other words, the Rich Amine out of the Absorber bottom goes to an Amine exchanger, on to the Amine Stripper/Reboiler, on to the Amine exchanger (as Lean Amine), and on to the Lean Pump and the Absorber. There are level detectors and controls on the absorber and the Reboiler. Therefore, if the Rich Amine is blocked at the Absorber outlet, the Amine loop will be broken and the reboiler will be pumped dry (by the Lean Pump) or the level detector will shut the unit down due to Amine flow failure. The over pressure at the absorber due to the Lean Pump's discharge will be MITIGATED by existing instrumentation. Even if you don't have the above standard instrumentation I've described, you would have to have sufficient Amine solution in the reboiler to fill your Absorber - which I strongly doubt because I've never seen an Amine system that allowed this. But, I may be wrong.... please tell me.

2. The resistances in the Lean Pump discharge do not play a role in the ability to pressurize the absorber to the dead head pressure level. What I mean by this is that if your pump were able to receive sufficient Amine solution, it would continue to pump until it reached the dead head level -235 psig - regardless of the resistance. The pressure drop resistances only slow down the delivery of the ultimate pressure - but the pump would continue until it realized 235 psig at its discharge flange. The pressure in the absorber would then be the 234 psig minus the hydrostatic head to get to the top of the Absorber (regardless of the straight pipe distance).

You state that "The feed is a header system in which a few streams combine before flowing into the absorber". However, this doesn't tell us the difference between the gas inlet feed and Amine Solution inlet feed. You should always state the two different feeds as well as the two different outlets. There is a big difference because one is a compressible fluid while the other isn't.

Double jeopardy is defined as occurances that are caused by UNRELATED (or INDEPENDENT) causes. However, as SooCS and CHD01 have noted, if your instrumentation controlling the blocked outlets of both gas overhead and Lean solution out are both controlled by the same pneumatic supply (and the valves are FC -Fail Closed), then you must consider that this is possible and not Double Jeopardy. An instrument air failure is the relationship between both outlets and it would cause both to fail closed at the same time. Of course, you may have electrical instruments that alarm and shut the unit down should the Absorber level increase to a dangerous level - like above the gas inlet. This would mitigate the scenario.

We could go on discussing your scenario through the night, but I think you've got the idea. A mini HazOp will be more effective than our further details because it involves the real, interested parties in the safety of your unit - the Stake Holders. Best of luck in your effort to improve your operation.


Art Montemayor
Spring, TX