What Is the Ideal Approach Temperature for Shell and Tube Heat Exchangers?

[jim hardy]

I would appreciate some advice from "old hand" heat exchanger guys regarding "good design practice" on approach temperature for tube and shell .

Here's what's up:

A friend works in a solar plant.
Oil is heated in parabolic reflectors and used to preheat feedwater for a traditional combined cycle boiler.

The large heat exchanger was initially tubed backward so it operated parallel flow instead of counterflow. Thermal performance aside, it experienced lots of mechanical tube failures attributed to the significant temperature difference across the tubes.

They're replacing it now and correcting the plumbing error so it'll be counterflow.

The hot side is Dowtherm A* oil at low pressure, cold side is water at ~2700 psi.
Temperatures are as follows (Fahrenheit):

Oil Inlet 650 Water Outlet 645

Oil Outlet 595 Water Inlet 325

At the cold end the tubes still see 595-325 = 270 degF ΔT.

My friend's concern is that the mechanical failures will continue , for no basic change to the heat exchanger is planned just correct the plumbing error. We fear there might be "groupthink" afoot.

So my question is -

What are the effects of large delta-T across tubes in one end of a shell&tube heat exchanger?
Do above numbers seem high enough to warrant special design features?
What questions do we need to ask his design group?

I've read several design type papers on 'net and am okay with NTU, LMTD, etc
but have not stumbled across a nuts&bolts construction article yet that mentions high thermal stress across tubes. Will try to get more details about the exchanger - I'm not even sure yet if it's single pass or u-tube.
And being an electronics guy not a ME I'm short of vocabulary for search terms.

This is not homework, it's a real question from workingmen in industry.

* Dowtherm A datasheet:
http://msdssearch.dow.com/PublishedLiteratureDOWCOM/dh_08a5/0901b803808a5b98.pdf?filepath=/heattrans/pdfs/noreg/176-01463.pdf&fromPage=GetDoc Thanks for any help -

old jim

 

[Chestermiller]

The first question to ask is "what is the nature of the mechanical failures in the existing design?"

 

[AlephZero]

With those numbers it must have been designed to handle the thermal stresses.

As an order of magnitude estimate, thermal expansion of metals is about $10^{-5}$ per degree C, so a 100C temperature change creates a strain of about 0.1%. For bigger temperature changes you soon reach the elastic limit of the material, if it can't expand freely.

But if the thermal design was screwed by the plumbing error, the thermal stresses may have been more severe than were designed for, and the material properties may be degraded more as well.

Another possible issue is vibration and fatigue problems - that's a different (and more complicated) ballgame!

 

[jim hardy]

Thanks guys

i'm far away from it, have asked my friend to join PF and help out with the questions.

Will post more details as I get them.

Reversed plumbing placing thermal stress on wrong mechanical parts sounds really plausible - thanks.
still working, old jim

 

[rollingstein]

Most such exchangers will have a floating head (or a U-tube design) so that the tube-sheets can expand. But that accounts for overall high running Temps. and the associated expansion on startup.

How deltaT across a tube affects things is a slightly different issue.