I just asked to prompt a third post.
Air vs water involves a lot of design trade-offs. But at 1/24th to 1/10th scale water has the advantage for duplicating the correct Reynolds numbers. But you don't have as much problems with air leaks instead of water leaks. Having the water tunnel upside down means the load cells and other pieces don't have to be at the bottom of the water.
You have the posts by graysgarage, but he doesn't seem to come around much lately. I think generating hydrogen bubbles in the water to replace smoke trails is genius.
Here's a design I made for a water tunnel; it's rough.
I think two paddlewheels at the rear would be sufficient. The idea is to move water without touching it as much as is possible; to suppress turbulence.
If you call it a PC it's probably under-powered for the task. Computing power limits the granularity of the simulation and the volume it can encompass. And the speed can mean a big calculation is in a race with the heat death of Universe. I don't even try. But I will get an X-15 when it ships as a general replacement, with the expectation that the software will fall in place over time.