Thermal Gradient Stress Calculation

[Guest102023]

I would like to know a simple method for estimating the strain/stress resulting from a temperature gradient across a very thick-walled pipe. The low alloy steel part is essentially a donut shape, having 11" I.D. and wall thickness 5". Temperatures are 1000°F internal and ~300°F external.
I have zero expertise in and no access to computational methods and software . Thanks in advance.

 

[metengr]

http://books.google.com/books?id=1E...&resnum=8&ved=0CDUQ6AEwBw#v=onepage&q&f=false

 

[davefitz]

Another classic text that provides a wide range of thermal stress solutions is by Timoshenko.

Some other interesting approximations are associated with linear gradients thru the pipe wall, such as:
- As per background tehnical papers to the TRD 301 annex 1 ( and later EN codes) , a constant increase in the rate of change of mid-wall metal temperature will yield a linear temperature gradient thru the wall of the pipe. The thermal stress generated will be proportional to the (wall thickness ) squared * rate of change mid wall metal temperature.
- a simple "boiler stress monitor " can be based on the above approximation, by installing a thermocouple at the mid wall location and moitoring the rate of cahgne of mid wall metal temperature. Within the DCS, this rate is multiplied by the know ( wal lthickness ) squared and other material properties to generate an apparent therma stress. This can be added to the current stress caused by fluid pressure to yield total stress , and either the rate of fatigue life consumpti nis calculated ( using the rainflow cycle counting algorithm) or the rate of unit startup/shutdown is restrained based on limiting this total stress. Using such fatigue monitoring is one requirement of the new fast startup Siemens combined cycle plants.

 

[davefitz]

error-
should read " a constant rate of increase in midwall metal temperature"

 

[desertfox]

Hi brimstoner

A very crude approximation could be obtained as attached
file.

desertfox