Process‐Informed Multiscale Prediction of Nonlinear Compressive Behavior in Direct Compounded Compression Molded Short Fiber Composite Lattice Structures

ABSTRACT Direct compounded compression molding (DCCM) offers an efficient route for producing SFRTP lattice structures, but predicting their nonlinear compressive behavior is challenging due to spatially heterogeneous fiber orientations from the molding process. This study presents a process‐informed multiscale framework for such prediction. At the process scale, Moldflow simulation predicts fiber orientation tensors across lattice regions, validated by metallographic image analysis with a 7.89% relative error. At the material scale, a unit cell model with fiber‐matrix interfacial cohesive behavior captures the SFRTP's nonlinear response. A transversely isotropic elastoplastic surrogate model and orientation‐averaging method yield region‐specific constitutive properties. At the structural scale, these properties are assigned separately to struts and four joint types in a refined finite element model. The predicted maximum stress error in the nonlinear stage is 3.8%, a significant improvement over the previous homogeneous assignment strategy, which had a 12.28% error. This framework efficiently incorporates process‐induced local anisotropy into structural analysis of compression‐molded short‐fiber composite lattices, reducing the need for exhaustive region‐by‐region mechanical characterization.