Depending on your model. If you are going to model the pipe Wall in a complex model, it doesnot matter, but the surface will be the inner surface. If the model will deal with the pipe as a contour you can take the average diameter.
My query is whether thermal conductivity of the pipe material plays a role or not? If the thermal conductivity of the pipe material is high - heat of the fluid passing through the pipe makes the outside diameter of the pipe to reach a higher temperature soon - in this case we can consider r2. The reverse is applicable for a low thermal conductivity pipe material. Am I right?
The wall thickness is proportional to the thermal resistance for the same material, so yes, it does have an effect.
Dobson & Meyer published some relevant papers, using the thermal resistance model to solve thermosyphon/heat pipe problems. The method offers a simple solution, with accurate results, for heat pipe / thermosyphon problems. I've used the method in a couple of designs with really good results.
Here are links to some of their papers:
Flow and heat transfer in a closed loop thermosyphon. Part I – theoretical simulation - http://www.erc.uct.ac.za/jesa/volume18/18-4jesa-dobson.pdf
and,
Thermal performance characterization of R134a and Buthane charged two phase thermosyphons.
It is necessary to to calculate the thermal (logarithmic) resistance of the pipeline. If it gives very small it must suppress. To calculate the global coefficient of heat transference, there takes for convenience the outside diameter of the pipeline, this due to the fact that it is more easily to measure it.