We have a gear shrunk onto a shaft that we want to remove with minimal damage to the bores. The gear is ~26" in diameter with a ~16" bore with .009 to .012" diametral interference. The gear is carburized so we need to keep the temperature at the teeth less than 350°F to avoid drawing back the hardness. When we assembled the gear it needed to be 90 to 120°F hotter than the shaft for the opening to be a line to line fit. We could heat the gear separate from the shaft so assembly was easy. Now with the two items in contact I am wondering if we can enough of a thermal gradient to allow separation.

In the past we would put the assembly in a press apply a load to the gear and apply heat to the outside of the gear until the gear slipped and then apply more load to press the gear off. This often damages the bore that needs to be repaired before re-assembly. So I am thinking we could apply a little more science to our method using a transient thermal analysis that may tell us how long we can heat it before we lose our window of opportunity because the heat goes into the shaft and it starts expanding too.

I have not done a transient thermal analysis so am unsure how to apply boundary conditions to this problem and how I would determine how much the interference has been reduced at a given time in the analysis.

Has anyone done this type of analysis or know if it is even doable?