Temperature Gradients

Typically, we model entire lengths of pipe as one temperature, making the assumption that at steady state, the entire pipe achieves a single temperature, or simply to be conservative, we model it at maximum temperature.

Question 1) Does anyone know if any the codes specifically prohibit analyzing pipe a multiple simultaneous temperatures along the length of a pipe, i.e. if one part of the pipe sees the extreme temperature case, do you have to analyze the whole thing at the extreme temperature?

Question 2) As far as I know, CAESAR II really only allows you to change temperatures at nodes, and in order to model a temperature gradient, you have to split up a length of pipe into a series of nodes and discretize the temperature at each node. While splitting up a section of pipe into a hundred nodes isn't the problem, I don't relish manually entering the temperature points manually. Is there any other way to do this?

And presuming the answer to both questions is "No," I would like to request this feature in CAESAR II.

Suggestions: Leave it to the user to determine by which method the software interpolates a section of pipe. Require the user to input two or more temperature points, and require the user to tell CAESAR II how to interpolate between those individual points (e.g. linearly, polynomially, exponentially, step function - although, technically we can already [only] do step function - I think).

Thanks for reading.

I doubt the usefulness of such a feature ... how could you be sure about temperature at a given point ? Can you give an example of where this sort of analysis would provide more realistic or practical results than the usual assumption of a constant temperature ?

If you are going to make bold temperature assumptions, consider this ... I have witnessed a thick wall discharge pipe from a gas compressor with a flowing temp at the nozzle of 175 deg C, reported on 4 radial temp probes. Measuring the external surface temperature of the pipe locally .. and accurately through a piece of insulation ... it was never over 75 deg C. Should this quite severe temperature gradient be considered in addition ?

For new designs, it would behoove the designer to incorporate this kind of conservatism.

For existing systems with well known operating temperatures that were never modeled in CAESAR II? For informational purposes only, this can be quite useful.

To answer your question about a temperature gradient from the inside surface of the pipe to the outside surface of the pipe:

Without having looked in the code to verify if this is otherwise addressed, my gut instinct would be "yes," even though CAESAR II doesn't have this feature (that I'm aware of). It doesn't seem like it would be too hard to do, except I'd suggest doing it with the data exported from C2 in Excel.

For long Flare systems it is not uncommon to use a gradual cooldown from the the main sources. But this temperature gradient is calculated or simulated by the process group or specialists. A stress engineer should not estimate this. Request and let the experts provide you the data, else run the entire line at full temp.

The Nuclear Class 1 piping must consider "thru the wall" stress gradients. Time consuming fatgue evaluations are part of the every day work for those lucky enough to have lived through these piping systems.

For petrochemical piping, these evaluations are seldom performed but they do exist for startup, spray conditions, etc. B31 type piping systems typically use proven rules of thumb to avoid stress problems associated with these "thru the wall" stress gradients such as avoiding repads and certain welded attachments.