Demister Pad for Shell and Tube Heat Exchanger

[Morts81]

Gday Forum Members

We have a current issue where we are having to replace existing undersized TEMA AES S&T Coolers (air compressor inter-stage coolers). The existing have a demister pad installed in the DN400 outlet nozzle and we intended to keep this configuration, however, during the design phase it was identified that the velocity through this nozzle was significantly above the maximum required for the demister to work.

We have engaged an experienced demister supplier who has confirmed that we need ~1.3m^2 demister area to ensure that the velocity is sufficiently low to knock out the water in the air.

Our intent was to accomodate this demister as part of the bundle, we had envisaged a 1000mm diameter vessel with the demister integrated into the bundle sitting above the top row of tubes. However, the design review showed that we need to have a 300mm gap between the top row of tubes and underside of demister for the velocity between the tubes to be brought down sufficiently before entering demister. To do this, the diameter of the vessel needed to be increased to such a size that HTRI would not accurately model the exchanger because of such a large dome space above the top row of tubes.

Having looked at what is offered by suppliers, demisters are predominantly provided in outlet nozzles with little regard to whether they will be successful in operation or not. My qusetion is, does anybody have experience with integrating demisters into S&T bundles and what did that design look like? Or in peoples experience, is an external knock out drum or similar the way to go?

Appreciate any assistance that can be provided.

 

[SnTMan]

Morts81, I have rarely seen a demister incorporated into a bundle for reasons you mention. Obviously if a certain level of separation is to be met the demister has to be sized and installed properly.

Whether located in the shell or a drum is a matter of "packaging".

Incidentally, I find your statement "HTRI would not accurately model the exchanger because of such a large dome space above the top row of tubes" curious as, while I do not run HTRI myself, I have seen numerous no-tube-in-window designs and kettle designs run in HTRI by very experienced people.

Regards,

Mike


The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[marty007]

If you must have the demister integral to your exchanger, there are a couple of other options rather than increasing the diameter of your whole unit. To keep the shell tight to your bundle all you need to to is fix the outlet nozzle. This could be done is a few ways:
[ul]
[li]Locally increase the diameter using something like a vapour distribution belt.[/li]
[li]Make the outlet nozzle large in diameter and a little longer than normal (whatever is needed from the demister supplier)[/li]
[li]Increase the shell diameter just at the outlet. (This makes it a little difficult to complete the ASME calcs (if that's your code), since ASME assumes the two tubesheets are of the same design).[/li]
[/ul]

Would any of these work?

Cheers,

 

[r6155]

Use hand calculation
Regards
r6155

 

[georgeverghese]

Wouldnt this then be just a matter of extending the shell length so that the vapor exit is beyond and downstream of the tube bundle ?

Trying another approach; Is this is wiremesh demister? If so, a double pocket vane type demsiter pack may be used with higher K factor and hence make this drum smaller in dia ? There may be higher liquid carryover, but centrifugal compressors can manage with upstream scrubbers fitted with vane packs also.

 

[Morts81]

Thankyou for all your responses

SnTMan, I believe the issue with HTRI is that because this is a baffled construction the program doesn't like the large dome space. For kettles and similar this is not an issue.

marty007, we investigated the options you have put forward. Increasing outlet nozzle is unfeasible, we currently have DN400 but we need a surface area of 1.3m^2 to bring velocity down so this would mean a diameter of ~1.3m. We certainly looked at increasing the diameter of the vessel locally, it would have to be 1300mm up from 600mm, our client didn't like this. I will certainly investigate the vapour distribution belt you refer to, thankyou.

r6155, our client would unfortunately not accept this.

georgeeverghese, this is an AES bundle, we did investigate the approach you refer to with a U-Tube bundle and included the demister in a removable end but the client didn't like this approach as it seemed to be a novelty in their words.
We are currently integrating a wire mesh demister, yes. The vane type separator approach was investigated in the line after the cooler but our proces department felt the sizing was extremely rigourous and based on that sizing we would still have a cooler with a similar diameter.

Appreciate all your responses guys.

 

[marty007]

Mortis81,

Just one note when you look at the vapor belt, the ASME code has specific rules. They are not hard to follow, I'm just point out that they exist. You can find the section in paragraph: UHX-4(c).

Cheers,
Marty