I would set the CAESAR II ambient temperature to the install temperature and then run analyses for (T1=)45 and (T2=)0.6

For expansion stress evaluation I would create a new combination composed of the difference between these two states (like T1=T2).

I say 0.6 rather than 0.5 for T2 because, by default, values for T less than 0.5 are considered as thermal strain rather than temperature. If you use 0.5, make sure that this is used as temperature.

Others would do it differently.