End effector pose optimization based on an integrated normal vector for improved compaction pressure uniformity in AFP on complex curved molds

Abstract

During automated fiber placement (AFP) on complex curved molds, non-uniform compaction pressure can arise due to the inconsistency of surface normals within the finite roller–tool contact patch, which may induce local bridging and wrinkling and ultimately degrade the mechanical performance of the cured part. To address this issue, this study proposes an integrated-normal-vector-based end-effector pose optimization method. Unlike conventional approaches that orient the end-effector using only a point-wise surface normal along the toolpath, the proposed method explicitly accounts for curvature-induced normal variation over the contact patch. First, a measurable description of “insufficient contact” is introduced, and the minimum compaction force required to maintain continuous roller–tool conformity is estimated using three-dimensional Hertzian contact theory and verified through finite element analysis of the polyurethane-coated roller. Next, the local tool surface within the measured contact footprint is discretized into triangular facets, and an area-weighted integrated normal vector is computed to determine the optimized end-effector orientation. Pressure-sensitive film measurements and curved-mold layup trials demonstrate that the proposed method improves pressure uniformity and reduces surface defects, validating its practicality and effectiveness for AFP on complex curved surfaces.