Bayesian inference of fiber orientation and polymer properties in short fiber-reinforced polymer composites

We present a Bayesian methodology to infer the elastic modulus of constituent polymer and fiber orientation state in short-fiber reinforced composite (SFRP). The properties are inversely determined using only few experimental tests. Developing manufacturing digital twins for SFRP processes, including injection molding extrusion deposition additive (EDAM) requires extensive material characterization. In particular, characterizing mechanical is time consuming therefore, micromechanics models used fully identify elasticity tensor. Hence, objective this paper effective tensor with minimal To that end, we develop hierarchical model coupled simultaneously which then use estimate motivate demonstrate EDAM process but development such it applicable other composites processed via methods. Our results approach provides reliable framework inference, as three tensile tests, while accounting epistemic aleatory uncertainty. Posterior predictive checks show able recreate data well. ability calibrate its associated uncertainties, make promising tool enabling probabilistic twins.