S&T design & mean metal temps: short cut?

[BigTank]

Without going into too much detail, the problem is that I have several replacement S&T exchangers to model. In order to verify the previous design, I must input the correct design parameters, obviously. Of course, having only a U1 and the design spec. information, there is some input data missing.

Specifically, I do not have the mean metal temperatures for the tubes, shell or the tube sheets. Calculating these values require knowledge of the fluid film coefficients. Calculating those goes fairly deep into the design of these exchangers. Being replacement units, I do not want to spend too much time reverse-engineering them. Is there a 'quick and dirty' method that is applicable and acceptable?

I am using the Compress exchanger module for modeling and calcs.

--------------------------------
Fitter, happier, more productive

 

[thermmech]

Hi,

the short answer is NO.

For Div 1 exchangers you would need mean metal temperatures only for:

1. fixed tubesheet units (Temp's for shell and tubes)
2. units other than FXDTS HEx where there is a significant
inlet-outlet process temperature differential on either process side which could lead to detrimental radial thermal expansion and girth flanged joint leaks.

So, if you need those temperatures you should know:

1. how the exchanger will be operated (turndown / upset conditions)
2. whether clean/fouled on either side
in order to come up with the most onerous mode of operation and subsequently, most onerous mean metal temperatures. You will need a heat transfer program capable of simulating process conditions.

Cheers,
S.

 

[carthago]

..And undrstood from Bigtaljk that he needs the means metal temperature for stress , strength and thickness calculstions as per ASME VIII, div.1 , UHX chapter..

a mechanical would have seldom a process data sheet with mean temperature,..most of the time he has the operating temperature and design temperatures...

the question = is there a way or a rule of thumb to estimate the average temperature of: tubesheet , tubes and shell.. I believe talking the average is not correct !!

 

[BigTank]

carthago:

correct, taking the average will get you into trouble. serious trouble.

i'm thinking the best estimate would be to estimate these means as best as your engineering intuition will allow with the data you have at hand (average film coefficients perhaps?), and then calculate for the 'worst' or 'most extreme' design conditions in addition to this estimate (i.e. startup & shutdown, tube-side flow stoppage, shell side flow stoppage, etc.)

--------------------------------
Fitter, happier, more productive

 

[gr2vessels]

The mean metal temperature is input to the Compress from the HTRI calculation;- and you are correct, the rule of thumb in estimating the MMT is receipe for trouble. Sorry, but the only way out is the HTRI evaluation...
gr2vessels

 

[carthago]

..Guys..
Could you elaborate a bit more about this rule of thumb

an exemple or something from the basic litterature are welcome.

thanks

 

[thermmech]

In most of the cases these temperatures cannot be estimated.

To calculate them you need not only a process data sheet, but knowledge of all the possible modes of operation (temperatures, flows, and fouling factors not indicated on the process data sheet) which will determine the actual governing mean metal temperatures for mechanical design.

Doing mechanical design on the basis of process data sheet which usually doesn't properly specify the upset cases, clean case operation, turndowns, etc. can be costly. Again, this is only pertinent to the cases where mean metal temperatures DO matter.
Regards,
S.

 

[gr2vessels]

carthago,
Nobody in sane status of mind would put in writing a rule of thumb for this estimate, exposing himself to the obvious ridicule. I pulled myself out of trouble when I have used inlet / outlet temperatures recorded in the last six months of an exchanger operation, I have added some 15% margin on the differential temperature between the shell and tubes, then I used the PV Elite to calculate the thermal expansion stresses. The result was very good, so I have increased the temperature differential until reached the yield. The temperatures used were exaggerated to the extent that they were exceeding any possible process scenario. Then I said I can safely use this rule of thumb, whilst covering my back side. I do not however, recommend this to anyone, unless serious research is done, to estimate also the consequences of anything going wrong...
Cheers,
gr2vessels