If the section is symmetrical about both axes, it's pretty straightforward. I agree with ishvaag tha tmost engineers would inscribe a circle in the hexagon and use the circle for the design calcs. When you get into an L-shape, it's not as straightforward. This is similar to a single angle compression member for steel, and buckling is more of a consideration. I don't know that modeling a stick column of a steel angle (or an L-shaped concrete column) is going to do you much good other than getting out accurate second order moments. It won't provide you with buckling information, and that's something that should be investigated with a section like that. For the strength check of an L-shaped column, I would determine the principal axes. AISC pretty much does this for you with single angles, but to do it yourself you'll need to go through a little bit of math, including finding the product of intertia, Ixy. When you have that information, you can recompute your moments to be about these principal axes instead of the geometric axes and proceed with the strain compatibility calculations to get 4 or 5 critical points on the strength interaction diagram. You will likely end up with moments about both axes which makes the interaction "surface" a little more difficult to visualize since it will now be 3-D instead of 2-D. After the strength considerations are worked out I would do a buckling check with rz, or rmin, by developing a DE for your specific beam-column or checking to see if one that is in Timoshenko's "Theory of Elastic Stability" fits your case.