5 replies to this topic

[AndyK]

Typically when sizing a liquid outlet nozzle from a separator, I use a maximum velocity of 1 m/s as the basis. This seems to be the industry norm and referenced in many company standards including Shell DEPs. I’m looking to increase the capacity of a separator and the only bottleneck is the nozzle velocity. The nozzle is on the bottom of a three-phase separator and fitted with a vortex breaker. I presume the maximum velocity is to prevent coning and potential gas entrainment in the liquid. I would have thought that 1 m/s cannot be an absolute limit, particularly where a vortex breaker is installed, however I can’t find any reference material that supports a higher acceptable velocity than 1 m/s. Does anyone have any reference source for higher acceptable velocities? I would rather not have to go to the detail of CFD modelling to prove the higher velocity is acceptable.

[Art Montemayor]

Andy:

I’ve never concerned myself with sizing the nozzles on a vessel independently – i.e., without taking into consideration the piping connected to those very nozzles. And I am 100% certain that no other process engineer would do it differently. The reason for this is that the pressure drop related to the attached piping is of concern and is usually the parameter one uses to size the line and, consequently, the nozzle itself. In other words, it is done the other way around.

I have never heard of an industry norm that sets nozzle velocities. I am certain that the Shell “DEP” (Design Engineering Practice?) is citing a generality and not a “standard”. In other words, when one has no other guideline by which to set the nozzle sizes, one would normally pick a conservative velocity – in this case 3.3 ft/sec as opposed to the average design result which is usually in the range of 5 to 10 ft/sec for liquid velocities. But when you are designing a process vessel, you are subject to the allowable pressure drops within the process you are working in.

Think about your application from a common sense point of view: why would you select a lower velocity through your nozzle and then be forced to employ a concentric reducer to be able to weld or flange to a smaller diameter pipe? You would be forced to do this because I can assure you that a liquid velocity of 3.3 ft/sec is not permitted in some cases due to excessive laminar flow that has a propensity to deposit solids within the line – especially when dealing with slurries. As I’ve stated, you will confirm out in the field that 5-10 ft/sec is the normal range of liquid velocities resulting when one designs the associated piping based on an allowable or optimum pressure drop. As you get further and further into fluid flow applications you will find that it is the pressure drop that normally sets the criteria for designing the pipe diameter (and the resulting velocity). At the acceptable pressure drop, the velocity is checked to see if it falls within the acceptable range for the type of fluid and you go with that. Pressure drop is the primary criteria used because it is the one parameter directly related to the power ($$$$) required to transport the fluid. And you will find that money ($$$$) is usually the principle motive for optimization and “good” design. I certainly do not see any logical reason for contemplating CFD modeling for this application.

[gvdlans]

The Shell DEP (Design and Engineering Practice) on Gas/liquid Separators indeed states that the diameter of the liquid outlet nozzle shall be chosen such that the liquid velocity does not exceed 1 m/s. The minimum diameter is 0.05 m (2 in). The nozzle shall be equipped with a vortex breaker. They do not give an explanation for this requirement (in Shell terms the wording "shall" means a requirement, where "should" means a recommendation).

[joerd]

My guess is that the 1 m/s is related to the recommended max. velocity in pump suction lines, which is also 1 m/s according to Shell, if I'm not mistaken.
However, if you're not on a Shell project, you would be free to use your own judgment, and pressure drop calculations will get you a long way towards that. I would say between 1 and 2 m/s is a good rule of thumb.

[AndyK]

In my experience it is the norm to size the liquid outlet nozzle independently of the attached pipe using different criteria. I would also say that is normal that the liquid outlet nozzle is larger that the attached pipework rather than the same size, due to having this low velocity criteria for the nozzle. The 1 m/s is a Shell (and other operators’ design requirement, not recommendation), regardless of whether the vessel is feeding a pump or not. Even Campbell's Gas Processing hanbook quotes 1 m/s as the sizing basis.

[JoeWong]

In my experience it is the norm to size the liquid outlet nozzle independently of the attached pipe using different criteria. I would also say that is normal that the liquid outlet nozzle is larger that the attached pipework rather than the same size, due to having this low velocity criteria for the nozzle. The 1 m/s is a Shell (and other operators’ design requirement, not recommendation), regardless of whether the vessel is feeding a pump or not. Even Campbell's Gas Processing hanbook quotes 1 m/s as the sizing basis.

AndyK,

Have you ever ask yourself or other expert why design for 1m/s ? Why not 0.5 m/s or 1.5 m/s ?

If the liquid outlet is feeding a control valve, velocity > 1 m/s may only induce high pressure drop and partial flashing...control valve still can be handle 2 phase flow...what's wrong with vel. > 1 m/s ?

If your liquid feeding a pump, velocity > 1 m/s cause high pressure drop. But if you have sufficient static head and NPSHa still higher than NPSHr of the pump...what's wrong with vel. > 1m/s ?

SHELL DEP also limit the gas outlet with momentum criteria. I have experienced once where the gas momentum criteria result a very large outlet nozzle (it exceed the allowable ratio between nozzle and drum diameter) and it can not be constructed...what shall we do ?

Yes.....DEP mentioned that the liquid outlet nozzle shall be designed for 1 m/s. It stated "shall" means mandatory to comply per DEP...but you also have to understand that you still can raise derogation if you prove that technically it is sound...and you may get approval from PRINSIPAL for this kind of specific case...

I do agree with Mr. Montemayor. The criteria is rule to be comply for normal circumstances but may be deviate if reasonable. RULE IS SET BY HUMAN !

If you really have no choice but the velocity is not too far from 1 m/s and your analysis shows that you are technically sound...you may raise derogation to the PRINSIPAL...

Good luck.

JoeWong