Parallel tool-path generation for Additive Manufacturing: A GPU-based zigzag filling

Abstract

This paper presents a parallel zigzag (raster) tool-path generation method for Additive Manufacturing (AM). Based on the analysis of some ordinary serial algorithms, it was observed that some compute-intensive operations could be parallelized by using a Graphics Processing Unit (GPU) architecture. However, to achieve this, many challenges were faced and solved by designing a method to work concurrently with individual contour segments on multiple layers while keeping the data organized. The method’s ability to solve the zigzag generation problem was verified, and its performance was measured by running an exhaustive search for optimal raster angles to reduce manufacturing time. The results showed that the method was effective and generated relevant computational gain, being up to 9 times faster than its serial counterpart. In the tool-path optimization, the simulations found configurations yielding an average length of raster lines up to 38% longer, which, in turn, can reduce manufacturing time.