Field-based partition path planning for automated fiber placement on complex surfaces via combinatorial optimization

Automated fiber placement (AFP) potentiates the efficient fabrication of fiber-reinforced composite. Laying paths in AFP must consider fiber orientation and parallelism to prevent quality loss. It is difficult to satisfy these strict constrains on complex surface by the overall method without partition. And the sector partition method lacks the ability of global analysis, hindering a better partition scheme. This paper develops a field-based partition strategy using combinatorial optimization with differential evolution (DE) for AFP path planning, which aims to minimize partitions while ensuring strict fiber alignment and parallelism of laying paths. The partition path planning process is mapped into a grid model, in which a grid node represents the partial laying paths on the surface meeting the constraints of fiber direction and path parallelism. Combining the nodes, a zigzag path in the grid model corresponds to a partition path planning solution of ply surface. And a DE-based method is proposed to optimize path result with the least partitions. Moreover, the vector field-based path smoothing helps to increase the steering radius of laying paths. Typical cases show that the proposed method outperforms the existing methods.