While I don't have a ton of experience in the simulation side I would say just by the nature of the pultrusion process it would be hard to replicate the cross strand sections as the fibers are pulled from the loom through the die itself. The matting would be consistent but the individual strands would need to be moving as they are coming out of the looms and crossing over each other to form the material. That would be one tricky simulation I would think.
If I were to put an educated guess on it I would say you need a material that has a non linear body structure, which I would not have a clue as to what that might be.
Maybe someone has such a material in their data base.
Good luck on this one, hope you find what you are looking for.
Standard composite analysis techniques apply - the process doesn't matter much - there is nothing not understood about pultrusion - it is a composite. Levels of sophistication apply. First level would be to get a laminate calculator and calculation the laminate stiffness and use orthotropic stiffness approximation. Next would be a laminated shell element analysis (far as you can go in SWX Sim), then there are continuum shell and solid element approximations to individual lamina analysis, then there is the meso scale analysis (I think they call it) when fibres and matrix are explicitly analyzed (never done it but know it exists - more for a materials scientist approach than a structures approach). You should be able to do a decent job of analyzing a pultrusion with the tool. The proper analysis may also include non linear analysis as pultrusions tend to be able to have large elastic deformations that is to say nonlinear.