A Robust Direct Slicing Method for Strut-Based Lattice Structures

Lattice structures, renowned for their superior mechanical and functional performance, have gained increasing attention across various industries. However, the direct slicing of strut-based lattices—particularly those featuring variable radii, dense intersections, and complex layouts—remains a significant challenge in additive manufacturing. This article proposes a robust and GPU-accelerated direct slicing method specifically designed for digital light processing (DLP) fabrication of strut-based lattice structures. By leveraging a discrete data preprocessing strategy, together with dedicated node and strut processing schemes, the proposed method mitigates slicing failures caused by intersecting struts and enables designers to adjust strut radii without explicitly resolving local intersections. In addition, a CUDA-based pixel-domain generation strategy is introduced to exploit the independent and idempotent nature of layer-wise pixel updates, thereby accelerating the conversion from geometric contours to projection-ready binary images. The algorithmic principles, including layer preprocessing, strut and node contour generation, pixel–domain integration, and GPU parallelization, are presented in detail. Experimental verification through DLP printing confirms the robustness and effectiveness of the proposed approach, and sensitivity tests further characterize the influence of frustum discretization and shallow plane–strut intersections on contour generation. Timing benchmarks on five representative lattice models further show that the CUDA implementation achieves speedups ranging from 1.45× to 8.50×, with an average speedup of 4.79×, compared with the CPU multiprocessing baseline. The proposed method therefore improves both the robustness and computational efficiency of direct slicing and provides a practical foundation for multiscale lattice fabrication and printing support generation.