large L/D ratio HX sizing

[vesselguy]

Hi all,
Are there any negative issues that I should be aware of in regards to slender HX, ie., shell length is much greater than diameter, lets say L/D > 5.

I have no concern with thermal design for long and slender hx shape but from a mechanical perspective, too long and slender a vessel the more challenge it is to support it on two saddles without getting into large repads around the shell. Are there other concerns?

Would it be correct to justify use of a hair-pin with shorter length in such instance?

thanks.

 

[gr2vessels]

There are many other concerns, like fabrication tolerances (might end up with a deformed shell), maintenance requirements (space for bundle removal), long flimsy bundle, special lifting arrangement for bundle handling, additional costs (more strakes, more weld). The structural weakness might offset the benefit of thinner materials. Overall, bad design, avoid if possible.
Cheers,
gr2vessels

 

[moltenmetal]

I'd say that depends greatly on whether or not your L is beyond 20 feet in length.

Long, slender bundles are common in small exchangers. I'd rather deal with the issues raised by gr2vessels than deal with multiple pass shells or low tubeside velocities.

 

[vesselguy]

Thanks to both for your input. I should've qualify that my question is pertaining to a fix tubesheet design, hence no pulling of bundle. I agree with moltenmental about limiting removable bundle to 20'.

gr2vessels brought up a good point which I did not thought of. large L/D shell can have significant deflection to mess up fabrication tolerance and would be difficult to extract a bundle on a removable bundle due to scraping of baffles caused by shell deflection.

I don't understand moltenmetal last sentence. It is because one is trying to develop the bulk stream velocity and that's why they end up with long bundle or multiple passes. I'm just trying to understand why you mention low tubeside velocity.

 

[moltenmetal]

Sorry, guess I've got tubeside on the brain. Yes, with small exchangers the focus is usually on the tubeside velocity, but it works the same on the shellside. As the note implies, if the option is between a long, thin bundle or multiple shell passes on a shorter bundle, I'll take the long, thin bundle every time.

 

[electricpete]

I am nowhere close to being familiar with hx design approaches....Just throwing in my 2 uninformed cents fwiw:

It seems that long tubes if not properly supported would be more susceptible to vibration which can cause fatigue.

Design against vibration is a tricky business and a big part of it is avoiding resonances. When flow is present, there can be broadband excitation and a variety of discrete-frequency excitations. I would think the logical approach is to attempt to place the lowest resonant frequency above most of the excitations which requires stiff design (relatively short distance between supports).


=====================================
(2B)+(2B)' ?

 

[tickle]

Cost is the main driver for keeping the diameter down (smaller tube plate and fewer tube holes). If multiple passes are required to achieve the required length, then you do not get true counter current flow and you loose heat transfer efficiency. But the length is usually constrained to standard tube lenghts (20ft/6m).

Of course there are standard designs for multi-tube hairpins (Brown Fintube) that offer cost advantages and do work.

Vibration is dependent on un-supported tube length, but also baffle spacing, baffle cut and tube/tube sheet attachment. When designing HEs I found that high cuts and spacings gave better thermal design and lower overall cost than short spacings and free area. TEMA has requirements for maximum baffle spacings and most STHE sizing programs predict vibration.