Today, we used IMC output of 10W for ghost beam hunting at OMC chamber, so I checked the temperature of HPBD during the work.
Figure 1 shows the temperature of several places around HPBD and IMC output.
Top, second and third signals show the temperature at beam dump, middle of the copper block, and top of the optical table, respectively.

Though IMC lost lock sometimes and beam on HPBD is not high for all time, temperature drift of HPBD itself seems fine with 10W IMC output.
Also, the temperature of optical table at IFI chamber did not change so much: that is good.
So, we can probably say it is fine to increase the IMC output if PRM is correctly misaligned (PRM is in MISALIGNED_BF state).