1) What is difference between Ts and Tmrt in your model/softeware?
2) I get from TRNSYS simulation Tmrt (but only long wave) and Total solar radiation (shortwave), how should I input these data to RayMan to calculate UTCI (is it possible)?
well, Ts is the surface temperature of the surroundings while Tmrt is the mean radiant temperature. These are different parameters.
What do you mean by Tmrt longwave? The mean radiant temperature is a temperature?
What exactly is solar radiation (shortwave)? If this is downward shortwave radiation, it may serve as global radiation to estimate Tmrt (and, thus, UTCI based on that).
Are you mean that Ts regards only to temperature of all surrounding surfaces (so no sky, no direct/diffuse radiation)?
In TRNSYS I can model only building (close structure). But if ceiling will be window with very thin glass then there is almost no obstacle for solar short radiation. This way I can model e.g. courtyard where temperature of building structure is influenced by solar radiation. On the other hand there is obstacle for sky longwave radiation. So, in such model Tmrt is calculated base on temperature of all inner surfaces (so walls and glass ceiling). Additionally I can get solar (short wave) radiation incoming to courtyard (ceiling glass).
Now I am considering how can I use these data in RayMan to calculate e.g. UTCI. It seems to me that TRNSYS Tmrt is almost like Ts in RayMan.
Ts is the average temperature of all surrounding surfaces. This is not a radiation, but strongly influenced by incident radiation and the basis for long-wave emissions.
I don't know TRNSYS, so I'm afraid I can't help you on that. Based on physics, I'd expect that (perfect) glass might only have a minor effect on shortwave radiation (depending on the angle), but it will definitely strongly modify the longwave radiation fluxes. If you calculate anything based on those (e.g. some surface temperature), I'd expect some considerable error.
Tmrt does include all short and longwave radiation. Why would you want to add shortwave radiation again? If you have Tmrt stick with that. However, if the model is an indoor model, the estimation for some courtyard might not be very precise.
For UTCI you need a considerable wind speed. It can't be calculated for indoors.
I'm not sure if I get a lot of what you wrote. However, the glass-roof solution does not sound very promising to me. Better use the RayMan obstacles editor to consider the surroundings and calculate Tmrt.