It's darn near impossible to accurately tell when the air will separate other than with direct flight testing or in a wind tunnel. The problem is that the angle of attack where this happens is so strongly tied to the reynolds number and the surface finish. Having said this a lot of the work being done in new airfoil designs by Selig, Drela and many others is geared to achieving shapes that show smooth and not overly steep pressure distribution curves when you run the shape through something like Xfoil or the Eppler predictive software. My favourite way of accessing Xfoil is through Profili2 which comes with it embedded and uses a far nicer interface than the command line version. To unlock the Xfoil and other features in Profili2 costs something like $20US paid to the writer. Or if you are stubborn you can download Xfoil for free and learn to use it from the command line. There's also an online forum for Xfoil to discuss all the ins and outs.

For shapes that have some serious discontinuities in the pressure distribution curves you can often figure on the airflow separating at or near the location where there's a radical bend in the curves. But even then it's dependent to a big extent on the Reynolds number and surface finish. Rougher finishes promote a turbulent boundry layer which tends to hold the airflow tighter to the wing while smooth and accurate shapes promote laminar flow for a longer portion of the airflow. The deal is that laminar flow can detach from the surface moer easily.

One place where you can be assured that you'll get a separation bubble is on a D tube style wing with the sheeting ending at the high point. When you cover this design the covering shrinks and pulls tight forming a really nice sharp discontinuity in the curve right at one of the worst possible points At any sort of semi seroius lift task for the wing you'll find at least a little sepration bubble between each rib and just downstream from this ridge.

As for the front fairing you'd want it to be shaped like the front end of a big thick symetrical airfoil where the curve just comes to the point where the last bits are parallel and in line with the sides of the block. The fairing block being airfoiled like this in a squarish shape and in two directions. Ideally it would use the back half of the airfoil behind the block to ease the air back together.

As for a more accurate modeling of the airflow it comes down to a few very serious computational flow dynamics programs. Us mere mortals can't even begin to think of what this costs.