Flow to parallel separators - flow regime

[Bala0610]

We have 2 separators operating in parallel - with symmetrical piping etc. However, we experience a 1 : 2 liquid flow distribution at approximately same gas rates, despite the u/s symmetrical piping arrangement.

The control system aims to control the vapour outlet flow from the vessel seeing the higher liquid rate, so that the gas can be made to preferentially flow into the other separator. Though this helps to improve (not correct !) the situation with respect to liquid flow, it causes much different gas rates.

I think the low operating flow rates could be an issue with the vertical inlet leg experiencing inadequate velocity and hence, not being able to "wash" the liquid into the separator efficiently. This can be seen when the horizontal pipe shows flow regime change from annular to stratified at about 75% of design flow rate...so if the line goes on to become vertical, would it not result in slugging ?

I have seen designs where the separator inlet is split into two vertical legs of half cross sectional areas, with isolation valve on one, to be shut down in times of low flow. Basically the intention is to keep the inlet velocity high enough at low flow conditions (start up, high turn down). Is there a technical basis in published literature or common engineering practise that I can use to implement such an arrangement ?

I'd appreciate if anyone whos experienced this problem of flow disturbance or engineered these systems successfully could share your experience.

Thanks

Bala

 

[zdas04]

Fluids will go where they will. I've never seen parallel separators that had similar gas and liquid flows. The difference can be significant (I've seen cases where one separator took 2/3 of the gas and 1/4 of the liqud). A small thing like one vessel being fed from the branch of a tee and the other coming from the outlet can make a big difference.

The real question is "what is it hurting?" If the gas and liquid flows are both within the design range why do you care if one gets more liquid than the other? I never spent a nickel to try to balance untis that both were within the measurement range. It looks funny on the measurement reports, but I never found that to be a bad thing.

David

 

[Bala0610]

Thanks David.

It gives me comfort to hear that this phenomenon is not uncommon and do agree that unless it hurts, no action need be taken. Whats the worst that could happen (beyond this imbalance - slugging, internals damage ?)

While running at about 70% of design capacity, the level is pretty close to HH and hence, we are apprehensive of increasing the load.

On the other hand, if the low 70% load is the reason for mal distribution (flow regime related), then probably the problem will fade away as we increase the plant throughput. We have (for a variety of reasons) been unable to check it at higher rates - so dont have actual operating experience.

The issue is of concern to me because we are building another system, with bigger vessels and inlet lines but similar to current flow rates (to begin with), and may possibly see this repeated (or worse?)

Thanks & Regards

Bala

 

[SNORGY]

Bala0610:

Are these two-phase or three-phase separators?

It will be even more difficult to achieve a symmetrical liquid flow split if there are two liquid phases present coincident with a slow feed velocity.

To me, your third paragraph has probably hit on something. People often downsize lines to move up into better flow regimes; amine regenerator feed comes to mind.

Would it work if you tied the liquid level control logic into the vapour outlet control logic and thereby tried to control individual separator throughput on level instead of on gas flow?

Regards,

SNORGY.

 

[Bala0610]

Thanks Snorgy

Separators are 2 phase.

That indeed is the scheme. The liquid flows are measured and compared and then the control logic attempts to reduce gas flow by throttling the vapour outlet of the vessel where the liquid flow is higher (while keeping the other vapour outlet fully open).

I was looking for engineering guidelines or published literature where the twin half-sized-inlets is discussed or recommended for better flowability under turndown conditions.

Best regards

Bala
Balaji

 

[davefitz]

In the boiler industry, 2-phase flow splitters were developed in the 1950's- 1960's. IN the case of branching a single stream into 2 streams, the single inlet pipe would be verticle, and either teminate in a splitter tee( 2 horizontal outlet branches), or use a spltter wye. In the case of a wye, there would be added an upstream splitter plate for 3 diameters upstream of the wye. The velocity in the single inlet pipe must be high enough to ensure dispersed bubble flow to minimize the unabalance in phases entering the 2 outlet branches. Use a baker plot or equivalent to determine the 2 phase flow regin .

 

[dcasto]

In the late 70's we had a class on two phase flow. We watched an 16mm film (yep that old) where they had a lab with glass piping (plastic?) They would mix different ratios of water and air and watch the flow. They were honing in on split flows. Generally what the film showed was that the liquid perfered to go out of the side of a tee. It was something like 90% of the liquid took the turn when the split was varied anywhere from 20% to 80% of the total flow was routed out the side connection.

It went on to demonstrate how you could take advantage of this and build a system where most of the liquid went with a 20% side draw.