Here's my problem. I want to figure the stress in a casting
as it cools. How I've set it up is I first run a thermal study I
give it two loads the first is my initial temp as the melting
temperature of 2800 F. The next load is I'm saying it cools to 80
F. I set this up as a transient study and I'm letting it cool for 4
hours and setting it to solve in 4 steps. I used the variation with
time option for my 80 F temp and it varies from the 2800 to 80. My
results for the temp seem reasonable.
The next step is to find stress in the casting from it cooling. Under the thermal options of my static study I import temperatures from my thermal study at time step 4 and use my melting point for iron as the reference temp at zero strain. I get reasonable displacement in the .005" range but I'm seeing 300-400 ksi over a lot of the part when I have a yield of 35 ksi
If it makes a difference in the case I'm using the faces the core of the casting would touch as a fixed restraint. Any ideas of what's happening?
The next step is to find stress in the casting from it cooling. Under the thermal options of my static study I import temperatures from my thermal study at time step 4 and use my melting point for iron as the reference temp at zero strain. I get reasonable displacement in the .005" range but I'm seeing 300-400 ksi over a lot of the part when I have a yield of 35 ksi
If it makes a difference in the case I'm using the faces the core of the casting would touch as a fixed restraint. Any ideas of what's happening?
The conclusion we came to is that... well this is a tough problem... If you cool a body uniformly throughout the volume & it is unrestrained, there is no stress. Any restraint that inhibits shrinkage will generate unreasonable stress as thermally induced expansion/contraction is a force of nature that is hard to stop. Only mathematically perfect restraints in FEA can do it.
In this case, the sinks sat on a table, unrestrained. However, thinner sections cooled more quickly and would shrink at different rates than thicker sections did. This caused stress. Also, we identified that convection conditions varied throughout the product. Some areas were exposed to a measureable airflow in the plant while others were blocked from moving air due to the geometry of the product itself. This again caused unequal cooling.
Finally, and this might not be as applicable to your problem, the enamal contracted at different rates than the cast iron but was subjected to the same cooling variations.
You certainly need to watch your restraints as Wayne suggested but also consider all the thermal effects in your problem before signing off on the results.
-- Vince
Changing the restraint swung my results in the other direction to where I'm at about 300psi now. In case anybody is wondering.