We are sure that we set the boundary condition (4 coolant Mass-flow inlets' Total Temperature ) as 363.6459 K. But when the case converged, we look for the surface integrals(Area weighted or Mass weighted average), we found the temperatures in the 4 inlets are much lower than we set before. temperatures in the inlets are supposed to be constant(363.6459 K),but there are about 2K decrease in each inlets (361K) What factor may case such condition?

You are specifying a "total" temperature of 363. K on the inlets. The CFD solver is solving for static temperature, which will be lower by the Mach relationships. Total temperature is a post-processed quantity. On the inlet boundary, it is possible that not all flow field variables are being updated with results from the solver iterations. This is not necessarily an error, as the boundary probably does not need all flow variables updated. I would suggest that you create a post-processing plane near, but not exactly on, the inlet boundaries in question and see if the results are as expected there. I suspect they are.

Thank you , that seems make sense. I cut some surface near the inlet, but the Total Temperature is still a little bit lower(1K), So I think that's also because the solver is solving the static temperature.