Water table modeling of radiator air ducts

Hi, I am looking at different mounting options of for radiators and there are significant technical risks. If the inlet and divergent sections have too tight of a bend I risk flow separation and loss of efficiency. The radiator should be normal to the flow so having a uniform flow area is important. On the exhaust side, the convergent section needs to efficiently accelerate the return, hot-air stream. There are lots of 'rules of thumb' but math modeling soon gets pretty . . . intense. So I was wondering about using a water table model for two-dimensional modeling. I can build models of the different approaches and use a laminar water table to look for flow separation. But water is about 800 times denser and the Reynolds numbers are going to be . . . a challenge. Density and Reynolds number potentially change the geometry of the model requiring scaling a longer or shorter profile. So I thought I'd ask if anyone has some recommended references for water table, aerodynamic modeling beyond Mr. Google? Thanks, Bob Wilson

I would first visit the NASA technical report server at ntrs.nasa.gov and search thru the *many* reports related to fluid flow thru ducts and radiator ducts. I'll have best results filtering down to just the NACA reports. Examples are

NACA-WR-L-407 and NACA-WR-L-115.