The problem is of course the fact that the subject material is not homogenous.
Isn't this usually resolved by buying/fabricating a load spreading part that would be bonded/assembled to the back side with resin/mat? This part, I guess would be threaded st. stl. or w.h.y. "The ideal client is one possessed of great good sense and perfect judgement;
that is to say, one who agrees entirely with the designer at all times and in every respect."
Fenwick Williams, Naval Architect.
Equations to calculate the recommended thread engagement length (and load to failure given a fixed engagement length) can be found in Machinery's Handbook, under Fasteners, Torque and Tension in Fasteners, Formulas for Stress Areas and Lengths of Engagement of Screw Threads. In the 25th Edition, it is on pp. 1415-1416.
Buy some fiberglass nuts from either Harrington Plastics or McMaster-Carr, and do some tests on stripping. Then, compare your fiberglass to that of the nuts, do some testing on your material & allow a big safety factor.
The following info is from the McMaster-Carr online catalog:
Fiberglass—Made of urethane resin. It’s nonconductive,
flame retardant, and is resistant to mild acids, solvents, and saltwater. Working temp.: up to 212° F. Minimum Rockwell
hardness: R119. Minimum tensile strength: 27,000 psi. Strength is greater than Nylon 6/6.
I think the Harrington Plastics Engineering book (not the catalog) gives torques for various sizes of fiberglass nut & bolt combinations, but of course, your material may differ.