Thermal stress in fin (boiler wall)

[Pawel27]

I would like to discuss about thermal stress in the fin of the boiler wall. Along fin width there is temperature distribution more or less like on picture "temp_distribution.jpg". Some fin has bigger elongation and some smaller one. There will be something like shear.

For our considerations we can omit(?) temperature difference across fin thickness. At the beginning at least.

My current experience with thermal stress concerns Stress Categorization Method acc. to ASME or EN standard. In this method thermal expansion difference is considered as fatigue.

I am thinking how to consider thermal expansion difference (thermal stress). Is it important beyond fatige phenomena or not? What calculation method can be used?

In some company there can be general rules about max temperature difference between pipe-tube weld and center of fin. I am interested in more precious method to check such boiler parts.

http://files.engineering.com/getfil...-8fd6-068a29779483&file=temp_distribution.bmp

http://files.engineering.com/getfil...1da0-4c82-a042-4bc89b3bfb5d&file=fin_tube.bmp

 

[davefitz]

You should use a finite difference or finite element model of the tube + fin . The same program can predict the temperature distribution due to conduction heat transfer as well as the thermal stress distribution. The key issue is correct input of the boundary conditions.

In a radiant furnace, the heat from the flame to the tube is via thermal radiaton , with the source temperature at about 2400 F ( 1588 K)and the reciever temp about 900 F ( 755 K). Therefore the source temperature dominates and the heat input is a fixed value that does not need iterative improvement- some other means of determining the radiant flux tothe tube surface is used and that value is maintained when solving for the "average tube"( at average flux) and the "worst tube" ( at the spot flux, assuming slag is removed). The back side of the membrane wall is adiabatic, as may be the end of the fins. The tube inside heat trasnfer coefficient is as per other calculations- very tricky if it is a supercritical fluid whose bulk fluid temp is slightly below the "psuedocritical temperature".


The stress boundary conditions are tricky- the membrane wall is constrained by buckstays and prevented from flexure out-of- plane by the other tubes in the wall, so a null rotation may be one boundary condition. If the tube being studied is assumed to be the "worst tube", then it may be operating at a higher temperature then the adjacent "average tubes", so it is constrained from thermal growth at the fin edges to the growth of the rest of the average wall. These constraints lead to very high longitudinal thermal stresses at the OD of the face of the worst tube facing the flame ; that section of the tube will yield in compression during load operations, and the resulting residual stress will lead to a horizontal circumferential crack when the unit is shutdown to ambient temperature.

 

[Pawel27]

Thank You for reply. I was very busy lately. What I am interested in: what are criteria considering tube pitch (i.e. tube distance). Looking from thermal stress of course.
I attached some screenshots. Please look. It seems that criteria are yield stress and ultimate tensile. Plasticity redistribution seems to be similar to that in static structural analysis (not thermal analysis). Any other things should be considered?

http://files.engineering.com/getfile.aspx?folder=a0796f2e-65bb-4a25-a824-9ec18263985d&file=1.pdf

http://files.engineering.com/getfile.aspx?folder=8c775bab-6b59-4a01-9e7f-81e171cca0b6&file=2.pdf

 

[Pawel27]

I am sorry, there is error in above attachments. Please use below:

Your file's link is:

http://files.engineering.com/getfile.aspx?folder=5de36253-88cd-4418-a850-3dece8921de5&file=1.jpg

Your file's link is:

http://files.engineering.com/getfile.aspx?folder=36785a18-086e-46fb-8892-9e7ae55b8fb8&file=2.jpg