I think there is going to be many opinions on how to solve this by going to 2D. I would approach this in stages. The first would be to determine what is an appropriate heat transfer coefficient inside (water side) of the pipe. We will use this value later on. On to your model, I image your boundary conditions will be like this: (Note I always get tripped up by this, but for axisymmetric models in Fluent put the axis on x-direction.)
Notice the tubes do not have any flow going through them. What we are modeling here is simply the ethanol side. The inlet and outlet for the ethanol are pretty straightforward. What I want to draw your attention to is the tubes. Here, I would set the boundary condtions using the aforementioned heat transfer coefficient. Here you can even include the thermal resistance for the thickness of the tube. A Named Expression can be used to define what is the water temperature.
The next stage would be to go back and adjust some of the assumptions and improve them. For example, the water temperature that is give above is linear, but this is probably not the case. It may be useful to even have a 3D model of just the water side that you can iterate back and forth to get to a converged solution.
You are navigating away from the AIS Discovery experience