Effect of DP on PVs 1

[NovaStark]

Most heat exchangers or pressure vessels that I have come into contact with are design for either full shell and/or full tube side pressures.

In the case of heat exchangers when fouling is concerned, can a high differential pressure damage tube-to-tubesheet joints or any other location?

(By differential pressure I mean with respect to one side only, so on the tubeside, the inlet is say 100 psig and the outlet is 80 psig.)

 

[jte]

Good observation! You msy be interested in a paper which addresses this issue in a bit of detail. In a nutshell, ignoring the change in pressure from one side of the channel partition plate to the other is not in all cases inconsequential. The state of stress at the (low pressure channel half) to (partition plate) to (high pressure channel half) junction is not simple nor always insignificant. Having said that, I'll say that I'm not aware of loss of containment issues attributed to this; it is more of a code compliance and complete design margin issue.

But how many times has a partition plate been torn off with nobody ever thinking "I wonder what the stress in the channel was when this tore off?" How many thermal efficiency losses are completely attributed to "fouling" when in fact some of the efficiency loss is simply tubeside fluid bypassing the tubes due to a partition plate failure?

 

[SnTMan]

Not to disregard jte's concern about the pass plate, which is real, as each side of the exchanger is usually designed for the full design pressure, I'd find it hard to believe any amount of differential pressure due to fouling would cause damage to the exchanger, except for fixed tubesheet exchangers and then the mechanism would likely be differential thermal expansion in excess of what the exchanger is designed for.

Regards,

Mike

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

 

[NovaStark]

@SnTMan

Do they normally set the inlet and outlet temperatures to be different with respect to mechanical design to determine thermal stress ? Most I've seen just set one temperature and pressure for thickness calculation.

Now the actual designer may actually do all that and analyze with FEA in their internal documents but that's probably propriety information that they don't give with their units.

 

[SnTMan]

NovaStark, fixed tubesheet exchanger design must account for differential thermal expansion between shell & tube sides. Much more to it than just thickness calculations. Refer to Part UHX-13.

Regards,

Mike

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

 

[NovaStark]

Thanks for the reference.

But like I said, most I see don't show the thermal stresses part as the OEM doesn't normally send that part.

 

[SnTMan]

NovaStark, these are not really thermal stresses, stresses due to temperature gradient. They are mechanical stresses due to difference in expansion between the shell and the tubes caused by the difference in temperature between the two sides. Only applicable to fixed tubesheet exchangers.

What kind of exchangers are you talking about? Floating head, U-tube, something else? What is your Code of construction? Do they fall under TEMA?

Regards,

Mike





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

 

[jte]

SnTMan- If I'm understanding NovaStark correctly, he's asking about the difference in temperature within the channel. Say the inlet is 300 deg and the outlet is at 200 deg. Your channel design temperature will be set at maybe 325 deg. Fine. But you have 180 deg of channel at a different temperature than the other 180 deg. So the concern is with regards to the gradient within the channel "shell course". This would arguably be a thermal stress due to a temperature gradient within the single pressure retaining component. I've not though much about this so besides intuitively thinking that there could possibly be a configuration where this might be a concern (maybe a thicker than average channel), I'd be curious as to the combined effect of a temperature gradient coupled with the moments induced by the partition plate.

 

[SnTMan]

jte, once again, I don't really know WHAT the OP is talking about. Started with differential pressure due to fouling...

Well, I just re-read his last sentence: "(By differential pressure I mean with respect to one side only, so on the tubeside, the inlet is say 100 psig and the outlet is 80 psig.)"

I guess I took it to mean, well, I'm now not sure what.

Could be about pressure drop across one side only, in which case the pass plate is a concern on the tubeside, a long baffle would also be a concern on the shell side. Can't think of anything else offhand that I'd be too worried about.

Really, I don't often see a specified allowable pressure drop over about 10 psig. Operate over that allowable and the unit performance is affected due to (I suppose) pump limitations, flow rates being off and so forth. This is the process engineers' side of things, not really mine

And yeah, the temperature gradient from one side of the channel to the other is not routinely considered during design of the run-of-the-mill shell & tube exchanger.

Well, perhaps we will learn more if we stay tuned

Regards,





Mike

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

 

[NovaStark]

Hey all,

Let me try to explain better.

Say we have a brand new heat exchanger (fixed tubesheets) and I'm only considering the tubeside for now.

Let's say the tubes and tubesheet is operating to 80 psig and 100F. So my documentation shows me mechanical calculations for the tubes and tubesheet based on these to both ASME 8 Div.1 and TEMA.

Now my exchanger has been in service for a year and heavy fouling is occurring so now my inlet is getting clogged such that the inlet pressre is now 120 psig and my outlet is now 75 psig. So across the tubeside there is a DP = 120-75 = 45 psi.

My question is, while 45 psi may not cause a tube rupture or the tubesheet to crack based on pressure alone but can it have any other detrimental effects such as buckling of tubes due to impingement on baffles, damage to the tube-to-tubesheet joints, etc.

My second question, is that if my HX is designed to TEMA, I see it calculates thermal differential expansion for the shell and tubes and tells me which is higher. Can I use this as basis to determine that if I calculate a tube expansion greater than this that I will damage my tubes ?

And well I guess now based on my second question, if I have one side as a floating tubesheet if I calculate a thermal expansion to be greater than the stop on the floating head side, can I use this as a basis to say I need to clean my exchanger ?


(Let's pretend that I have some way to measure all these temperatures accurately)

 

[SnTMan]

NovaStark, as I said before, other than perhaps a pass partition or long baffle, increased pressure drop alone across the tubeside or shellside is unlikely to cause any damage to the exchanger. Assuming of course the design pressures are not exceeded.

Of more (much more) interest, especially in the case of a fixed tubesheet exchanger, is what is happening with the fluid temperatures. Fouling is a thermal resistance. It prevents heat transfer from the hot side to the cold side. Hot side exit temperature will approach the inlet temperature. The same is true for the cold side. In the extreme, the inlet and outlet temperatures for each side are equal, no heat is transferred.

For floating head or U-tube exchangers, no big deal, since differential thermal expansion between the tubeside and shell side is not a design consideration, assuming clearances are such that no "collisions" occur within the exchanger. By the way, I've never seen "stops" on a floating head exchanger, got an example to share?

For fixed tubesheet exchangers, differential thermal expansion is a big deal, due to the stresses imposed on the shell, tubesheet, tubes and tubesheet joints, both tensile and compressive. The differential expansion is almost entirely due to the difference in the so called "mean metal" temperatures of each side. These are a representation of an "average" of the inlet and outlet temperatures. The differential is also a pretty strong function of the metallurgy of the exchanger. Pressures alone have a negligible effect.

So, for a given fouling, and supposing you know the associated flowrates and so forth, a new operating point can be calculated and the mean metal temperatures obtained. The fixed tubesheet design can be evaluated to see if the imposed stresses are within allowables. If so, operate away.

As you said, TEMA gives a calculation for differential thermal expansion as does Part UHX. These are based on metallurgy and the mean metal temperature for each side of the exchanger. As the operating point changes, so do the MMT's and the differential expansion. So in general, no, you cannot use the differential expansion as a basis for when to clean.

The key measure is the performance of the exchanger. When it no longer transfers the required amount of heat whatever is the cause, fouling or not, must be addressed. And by the way, fluid inlet and outlet temperatures are routinely monitored with acceptable accuracy on the vast majority of operating heat exchangers. That together with known flowrates is the ONLY basis for calculating performance.

Enough for tonite

Regards,

Mike

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

 

[NovaStark]

Thanks for your insight SnTMan.

I may have mixed a common term used here with the actual term with when I referred to "floating". I was mainly referring to one end being allowed to slide a set distance (so one end is fixed onto a concrete base and the other end is allowed to expand a fixed distance calculated by the designer). Sorry about that mix up.

Which then like you said, it shouldn't be a problem as it is still allowed to move.

But yes we do use heat exchanger performance but due to fouling, some people tend to get panicked with overheating a bundle because of the fouling. (which again, as you said, once the performance is unaffected it should be fine).

 

[georgeverghese]

@NovaStark: You've left a sting to the tail on your last sentence. Are you saying that

(a) Due to excessive pressure drop / fouling on the tubeside, to keep up HX performance, you have to increase shellside flow to the extent that you have a concern on what mean metal temps may be on the tubes and the tubesheet ?
AND
(b)That you have a HX with different design temp for shellside and tubesides?

 

[NovaStark]

@georgeverghese

Due to pressure drop as a result of fouling, there would be a DP of around 20-30 psi while the design data might say the allowable is 5 psi. So due to higher DP, less flow is allowed through the unit so what tends to happen is that to keep the unit online they would decrease the flow (this HX is in a loop so the flow through the tubeside = flow through shellside)

(the tube metal temperature is another issue that may be extremely complex to determine)


My initial question was mainly about DP damaging mechanical components. But it can be extended to temperature differential across the tubes also causing damage due to higher mean metal temps due to fouling.

 

[georgeverghese]

So, since shellside flow cannnot be increased, you've increased shellside temp and reduced tubeside flow to counteract the increased tube side fouling to the extent you now have a concern with mean metal temp at the tubesheet and tubes.
Shouldnt take much effort to find out the current tubeside fouling coeff and the new tube metal wall temp at the hot end of this HX, since you have the flows and temps on both sides of the HX in the current fouled condition.