you tried to simulate the PV incorrectly and that leads you to complicated case .
you can simulate the PV panel as it in actual case which consist of five layers ( glass , EVA , PV silicon , EVA and tedler) . Therefore the first layer is the glass and the top surface have two boundary conditions which are the convention and the radiation.
If you want , I can send you the thermostat-physical properties for each layer.
At the first stage, I would like to recommend you to check the "Theory- and User-manual" of your commercial Softwares. If you use the "Fluent", then please check it. If you do so, then you will find immediately which is possible and which is not possible by your selected solvers (for example, FLUENT, COMSOL Multiphysics, STAR-CD, OpenFOAM, Frontflow/red, and so on).
And at the second stage, for improving your basic knowledge and understanding about the associated physics related to your research, please take a look at any "Fundamental Text Book" on "Heat Transfer" , "Fluid Dynamics", OR, Heat Transfer in the Fluid Flow System, and PV-solar pannel system.
However, technically speaking, the heat flux (into the system or outside the system) is controlled by the "conditions of the surroundings" of your system of interest. Although I am not familiar with the "PV solar panel", I believe that the "basic theory of heat transfer" (heat gain /heat loss) will not be unaltered based on the applications. In this sense, the basic theory of heat transfer should come in the first place, and then how to implement the theory (i.e., BCs in your case) for a specific problem, which should come in the later stage.
And, this stage is the "technical stage", which can be found easily if you go through the "User guide" of your SOLVER, which is highlighted in the above section.
Anyway, I hope that you will find the suitable ways if you check the Books ( Heat transfer, PV-solar panel, etc), USER MANUALS, etc., as suggested in the above section.