Help with Nonlinear Transient Heat Transfer Analysis

[WillL84]

Hi guys,

 After following this tutorial I'm trying to get my simulation to work properly. Basically there are 300W heaters in the selected holes and I want to see how the parts heat up over time. I set the time scale but the animation seems to only show like 5 steps, none of which actually show any heat progression. After some searching I came across this thread and my time steps is set to constant not adaptive so it's not that.

Also while it's solving I notice the temp shoots up over 2700ºF. Is there a way to limit the temp? Basically these heaters are connected to a standard PID controller, the temp will be set to 700ºF

The files are too large to attach so I've uploaded them to my Google Drive.

I'm a complete Nastran newbie so any help/input/direction would be appreciated. The material for these parts is Hastelloy C276 which I've set up in my PC with the required physical properties. Not sure if that'll all translate in the pack n go or not.

[John_Holtz]

Hi @WillL84 

One problem is you did not apply 300 W. You applied a heat generation (units of energy per volume) to a face. Who knows what that equates to in the analysis. That is, what is the volume of a face? 

Instead, you should be calculating and applying a heat flux (units of energy per area) and apply that to the faces of the holes. 300W = 0.2843 BTU/s, and the area of 1 hole = pi*0.375 inch*2 inch = 2.3562 in^2, so the heat flux is 0.1207 BTU/s/in^2.

About the animation, we would need more detail about what steps you used, input, and so on. But there is a good chance that this article explains what you should be using. See Animation does not match the individual results of a Nastran analysis | Inventor Nastran | Autodesk ....

There is no equivalent to a thermostat in Nastran. My suggestion is to run one analysis with the heat flux to determine how long it takes to heat the surface to 700 F. (Also, the heaters may ramp up to 300 W instead of instantaneously starting at 300 W. Ramping the heaters would also help with numerical stability and calculating the temperatures.) Once you have calculated how the temperature of the hole changes with time, you can apply a temperature load (or convection with a large convection coefficient, whichever type of load allows a time dependent multiplier) and ramp the temperature until it reaches the thermostat temperature of 700 F.

I noticed a few other things that I thought were odd.

1. You have two parts that appear to be in contact, but you did not define bonded contact between the parts.
2. If the two parts are mirror images (geometry and load), you only need to analyze one of the two parts. Reduce the model size so that it runs faster!
3. It's hard to see the convection load, but it appears to be applied to the same holes as the heaters. I expect the heaters to plug the holes so that the convection does not also cool the holes.
4. The convection appears to be applied to the faces that are in contact. I would expect there to be no fluid flow on those flat faces where the parts cover each other.

[WillL84]

As far as the animation goes I fixed that. I had multiple time steps so the animation was for just a single step. I've since changed it to a single step that's longer.

The controllers turn on the heaters to 100% until they get close to the set temp and then PWM the outputs based on the calibrated PID values until it gets to temp and then to maintain the set temp so there's no ramp-up upon start unfortunately.

Basically the issue we're having is the compound that runs through the inner channel is burning in spots so I'm trying to see how the heat flows from the heater cartridges thought he assembly. The channel is closer to the heaters than the TC ports so it would seem the channel is overheating before the unit gets up to temp so I'm trying to see how hot the channel is getting near the heaters. We've added a few more TC ports that are right on the channel and have noticed up to a 50ºF difference between TC's depending on location. I'm trying to see what would change if we moved the heater ports in relation to the inner channel.

1. Yes the parts are together, it's a bolted connection. They're near mirror images - one side has threaded holes while the other has countersunk clearance holes with thermocouple port holes.

2. I was thinking that having the two parts together would properly show how the inner channel heats up vs having one half of it with the other open. Am I wrong on that? Or would only selecting certain faces (everything except the mating face) for the convection load make that not matter?

3. I did mistakenly apply the convection load to the holes as well (they will be plugged by the heaters) but have since changed to an initial temp for the only load with an 800F constraint applied to the inner hole surfaces (tried that based off of another tutorial video I watched). I'll remove that and go back to the heat flux setup.

4. Correct, there would be no fluid flow where the two faces meet. How much does a convection load (simple air convection) change the results? I'll try again with manually selecting the main faces that are exposed to air.

[WillL84]

Ok that didn't work. Within 15 seconds of calculating the hot spot was up over 50,000ºF lol. Also I'm having a strange behavior where the model disappears if I kill a solve using the stop button.