Problem 1 - This isn't an API 650 tank. API 650 starts a diameters much larger than this. While I've never attempted API 650 calcs on a small tank such as this, I guess now we know what happens.

Problem 2 - Even if this were an API 650 tank, the assumption that thermal growth of the perimeter directly translates into increase in radius is not reflective of reality. Thus, there shouldn't be a rigid element going to the center. To the bottom of the tank? Sure.

Observation - Rigid elements in CAESAR are not true rigid elements. They are as stiff as the pipe you specify them as, with the thickness multiplied by... 10? If I recall correctly. Check the help file.

Is there guidance for a tank such as this? Run FEA to obtain deflections.