liquid seal overpressure protection 2

[mm88]

Hi,

At our plant we have several tanks protected against overpressure by a liquid seal.
These liquid seals consist of a small pipe in another pipe with a larger diameter and a closed bottom end. (I've added a basic sketch for clarity)
In case of overpressure the liquid will move upwards and at a certain overpressure the gas will start flowing through the liquid .

Calculating the pressure at which the liquid seal opens and allows gas to flow through is rather straightforward but there have been some discussions about the scenario where a large amount of gas has to go through the liquid seal.
A larger flow means higher gas velocity through the liquid seal and possible fluid entrainment. In other words: two phase flow.
Does anyone have any suggestions in regards to calculating the increase in pressure drop over the liquid seal cause by this two phase flow?
(allowable gas velocity, liquid seal diameter,...)

Thanks in advance!

 

[LittleInch]

The key is the word seal, not valve or relief. These are not designed to allow any significant flow and if they do you will fairly rapidly not have very much liquid left... You own't really have two phase flow - you'll have gas with some liquid drops in it

I suggest you find another way to flow gas in the event of a relief scenario - PV valve, bursting disc, bent pin etc. All depends on what your unknown system is.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

Liquid seals are not reliable for the reasons you've mentioned: liquid entrainment into exit gas stream , rapid loss of level, corrosion in the well / dip leg etc.

As suggested, try a low dp backpressure control valve with low dp check valve; and
add an RD or buckling pin operated RV on a parallel bypass line across the backpressure PCV in case there are some operating scenarios that may be beyond the response capability of the backpressure PCV.

In some critical cases, you could add, on another parallel bypass line , an on-off quick acting gap acting fail open SDV that operates on PT feedback from upstream.

 

[mm88]

Thank you both for the replies!

I agree that some of the liquid seals will not be adequate as a measure against overpressure, depending on the scenario's, and will have to be replaced.

I was wondering though, up to which gas velocities can liquid seals be considered as a reliable overpressure protection? I've been looking for literature on the design of liquid seals but haven't found any good references so far. Do you know of any?

 

[LittleInch]

Zero m/sec. The lack of data should tell you something. Velocity is irrelevant, its flow that's the issue. More than the odd bubble every minute is too much.

None of the liquid seals are good for overpressure, not just "some"....

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

And also add a low pressure trip for the system (upstream of these devices) in case the RD has blown and there is potential for flashback. Some of these pressure transmitters may need to be made 1oo2 voting and trip loops configured in ESD-PLC in case these are in critical process safety service. The low dp check valve should be a tilting disk type.

 

[moltenmetal]

A liquid seal can be as reliable as any other type of overpressure device if it is maintained properly. If selected or designed wrong, it will take a LOT of maintenance. Selected and designed correctly, you can even successfully use them as a vent path for inerting gas or the like.

They have the advantage, or disadvantage, of also allowing vacuum relief. Of course in that case, you'll draw a slug of the liquid in the liquid seal back into the tank, so it must be selected to make that safe.

In the case that the flow through the seal is too high, it will simply blow the liquid out and become an open connection to atmosphere. If back-flow of atmospheric gas after the relief event is unsafe, then indeed these units are unsafe for that application- but then, a rupture disc would also be unsafe.

If you select the wrong MOC, the wrong liquid for use in the seal etc., of course they won't work. But as a means to protect a small atmospheric tank which needs a small pad pressure for inerting etc., they're better than a lot of other options you might select.

 

[LittleInch]

The biggest issue I see is that if these seals "blow", and remove part of your liquid seal, then the relief pressure essentially falls even more unless you know it has happened and religiously fill up the liquid level back to the same point. If you had a significant sized reservoir attached to the seal and the overpressure was short term then it might work Ok, but too many unknowns here to say much more.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[mm88]

Moltenmetal, do you have any suggestions as to how I can check whether our liquid seals were designed correctly? (I know this is not an easy, if not impossible, question without seeing the flowsheet or installation drawings, but any suggestions as to how we can proceed in making our plant inherently safe would be greatly appreciated)

LittleInch, most of these liquid seals are equipped with a level detection which detects when the liquid seal is blown out and generates an alarm(sometimes before the seal is blown).
Measures have also been taken to ensure that the liquid seal is full.
So during normal production the liquid seal will be full, and when the seal is blown because of high pressure this will be detected. In some cases this automatically shuts down production.

 

[LittleInch]

good - it would have been nice to know that up front. a question I meant to ask on your original diagram and that was which is connected to the tank? I assume the central tube, but would be nice to know.

what sort of inches water column are you working with?

Are they all separate seal pots or are some roof sleeves?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[mm88]

The central tube is connected to the tank.
In one instance we're working with 100 inches water column (250mbar)
I'm afraid I am not familiar with the term roof sleeve, what do you mean by this?

 

[LittleInch]

Sorry, I think its called a rim seal and was used extensively on telescopic gas holders. See this and scroll down a bit.

gluedideas.com/Encyclopedia-Britannica-Volume-10-Part-1-Game-Gun-Metal/Gasholders.html

A pot is a better idea especially if you have level switches to tell you when the seal is gone.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[mm88]

Thanks for the link, at our plant we only use liquid seal pots.
The problem right now is that HAZOPs usually require proof that these liquid seals prevent the tank from being pressurised above its design pressure. At low flow this is not a problem as the gas will bubble through the liquid seal at its "set pressure".
For some of these liquid seals however, we've found scenario's where the flow through the liquid seal seems high enough to blow out the liquid seal and creating a higher pressure drop during the process.

As I understand it, the next steps would be:

-Prove the pressure drop over the liquid seal is low enough so that the pressure inside the tank does not excel the design pressure when the liquid seal blows​

-If we fail to prove the liquid seal is adequately designed, add a rupture disk/safety valve to the tank sized for the higher flow of gas.​

any advice (preferably on the first step)?

 

[moltenmetal]

mm88- that is the right process.

As to step 1: ensure that the outlet piping and seal chamber can pass the relief/overpressure scenario flow without overpressuring the tank in the absence of the fill liquid, i.e. when the pot is empty. If it passes that test, you're only worried about a brief transient state when there is two phase flow of relieving gas and liquid from the seal pot in the discharge line. If you're close to the limit without the liquid, you'll be over the limit with the liquid and you need to start designing plan B. If you're well below the design pressure without the liquid, a brief spell of entrained liquid likely isn't going to push you over. But if you're close, you'll have to make an assumption about the rate at which the liquid leaves- assume some period of time for the liquid to leave, say 1 minute, and use the DIERS method or a conventional 2 phase flow method from fluid mechanics to calculate that brief, peak overpressure. A properly designed liquid seal will never have its exit tube completely filled with liquid, so the flow of pure liquid isn't a case you need to consider. If that brief peak pressure is above 110% of MAWP but below the hydrotest pressure for the tank, you'll have to decide whether or not a retrofit is necessary, since you should be reasonably confident that exceeding 110% of the MAWP for a few seconds by that small margin is unlikely to cause the tank to rupture. Depending on the contents and size of the tank, you can decide just how conservative you need to be.

 

[mm88]

Ok, thank you! This has been a great help.

 

[georgeverghese]

The other reason for poor function in liquid seals is biofouling / low temperature relief streams that could freeze the seal liquid(this could even be low ambient temp)/ high flow gas relief streams.

For the last of these, there are some recommendations in Perry for the design of sparger pipes in the liquid - gas chapter. Briefly it says that superficial gas velocity through the cross sectional area of the seal drum should not exceed 0.2ft/sec for non turbulent flow. Else another solution may be a larger seal drum?

 

[e43u8]

The following design features from "PRESSURE SAFETY DESIGN PRACTICES FOR REFINERY AND CHEMICAL OPERATIONS" are usually incorporated for liquid seal:

1. Continuous water makeup and overflow on the seal loop, to ensure that
the seal is always made during normal operation, and reestablished after a blow.
2. Adequate winterizing, where necessary, to prevent freezing of the seal.
3. Safe disposal of the effluent seal water, considering possible contamination by process fluids.
4. Specific criteria which govern the acceptability of discharging process fluids to atmosphere.
5. Contingencies by which liquid hydrocarbon could be discharged through the atmospheric vent must be positively eliminated.
6. The vent line must comply with safe practices to prevent flashback and snuffing requirements