Hello Stressers!!!'

I know this is not a Caesar II question. But I'm doing a transient thermal analysis to calculate some thermal stresses in a pad. Basically my question is how fast (in sec.) does the temperature rises say from ambient (21 deg. C) to say 427 deg. C operating at a 24" STD wall pipe. I will use this data in the Load Curve for a Transient Thermal Analysis. What factors affect this data? Do I need to run CFD analysis to get this value (rise time of temp.) and are dependent on fluid velocity, etc.

Any expert opinion is highly appreciated. Basically, if my rise time is very fast the thermal stresses are very high obviously.

Many Thanks,

This is a loaded question. When you say your rise time being fast yielding high thermal stresses, that's not entirely true.

You have two thermal gradients to consider:

1) From inner diameter of pipe to outer diameter of pipe. You'll actually want a tremendously fast heat transfer rate in order to minimize these stresses.

2) From pipe origin to pipe destination. Having a high heat transfer rate will get you to your highest stress point in a simple, linear system. This is not necessarily true for non-linear systems.

Example:
• Think of a general pipe configuration where one half "A" the other half "B." Put an anchor at the end of "A" and "B," and put an axial stop at the A/B interface with 0 gap.
• If you grow A and B at the same rate, the force on the directional stop will always be 0. If you don't, your directional stop will take up some fraction of whatever the anchors normally would. While your anchors might have been designed to compress that pipe for the total growth of the pipe, your directional stop will just snap right off. And if you needed that stop for some other reason, say wind, well, it's not there, anymore.

To that end, you'd want your heat transfer rate very high to reduce your pipe stress headache.

I'll see if I can dig up some temperature gradient calcs tonight.

Thanks Michael for your input. Looking forward to your temperature gradient calcs. I understand your concept of high heat transfer rate so that the pipe will reach steady state at fast rate.


Many Thanks,

Hi Michael and fellow stressers!!

To clearly explain the study I'm doing, please see the following:

Given Data:

24" Std. Wall Pipe with 10mm thick pad (16" diameter) at top of 24" pipe.

Design Temperature = 427 deg. C
Ambient Temperature= 21 deg. C
Line is not insulated (for example purposes only)
Pipe material is Carbon steel.

Objective: To obtain the thermal stresses at the junction of pipe and pad.

The transient thermal stress I'm performing is described below:

Step 1: At time t=0 (both the outside wall of pipe/pad and the inside wall of pipe is ambient/21 deg. C means no fluid inside pipe).

Step 2: At time=? (the fluid carrying the 427 deg. C temperature suddenly flows inside the pipe at velocity V).

Basically, my question is on step 2. At what time t=? the inside wall of pipe will reach 427 deg. C. Is it abrupt, say after a 1 sec. upon the 427 deg. C fluid touches the inside wall of pipe (the inside wall of pipe will become 427 deg. C also?).

The boundary condition I set-up on the inside wall of pipe is the controlled temperature where the temperature will rise from 21 deg. C to 427 deg. C at time t=?. The outside wall of pipe and pad I put a air convection at ambient conditions (for example purposes only since we usually put hot insulation at this temperature).

I have done some sensitivity studies for this time t=?. I have copied one of the example in youtube video where rise time t=60 sec (for inside wall of pipe, and I've got an approximate Von Mises stress of 260 MPa at junction of pipe and pad). Then I've tried rise time t=30 sec and got an approximate Von Mises stress of 320MPa. I've tried time t=10 sec but the software cannot converge on a solution maybe due to some tolerances in the settings. The steady state stress, where the temperature is equilibrium for both pipe and pad is approximately 25 MPa.

I've noticed, that as the rise time is abrupt the transient Von mises stress gets higher. So I want to have some firm basis on this rise time t=? to get a realistic transient Von Mises stress. Maybe this time t=? is dependent on fluid velocity,etc. and maybe some CFD is required.

Please correct me fellow stressers if I've made some wrong statement.



Cheers!!!

Unfortunately, I could not find any definitive assistance in my literature.

A coupled CFD/structural analysis will surely deliver to you the temperature profile within the fluid and pipe, though you'll run into some limitations for how much pipe you can run a detailed CFD within.

However, are you sure you want to assume that your hot fluid is dropped into cold pipe? Your pipe will have a temperature profile already due to conduction if your system is already hot. Assume that your discharge piping is infinitely long, and at the end, ambient temperature. The temperature profile from the header to the end will be some inverse exponential function. This will reduce the maximum temperature differential between fluid and pipe to some degree.

I don't believe we've been introduced to the fluid in question. Liquid or gas? If it's gas, it may not have a great deal of mass behind it, and the temperature rise in your pipe will therefore be relatively slow, and your gas will drop in temperature relatively quickly along the length of the pipe.

If it's liquid, is the line already liquid-packed? You'll want to consider how well the liquid is mixing and what the temperature is.

To answer your original question, the only factors that matter is how much matter you're discharging to the pipe, how much that matter wants to give up heat to the pipe and to itself, how much the pipe wants to accept the heat. Velocity will indeed affect how much hot matter is being dumped into the system, but will also define how well heat can be dumped to the pipe.

Hi Michael,

Thanks for the input. You're correct a CFD analysis is required for this type of problem.

Cheers!!!

Calculating the Heat Transfer Coefficient from Conjugate Heat Transfer Calculation in CFD software is the key to this problem.

Any other opinion is greatly appreciated.

Cheers!!!