Improving a tool-path optimisation method in material extrusion additive manufacturing by data clustering and collapsing

The material extrusion (MEX) additive manufacturing uses an extrusion head to deposit material following a tool path. Many jumps, motion without material deposition, are usually required to cover all regions, which can significantly impact the building time. In previous works, tool-path optimisation algorithms were proposed to reduce building time in MEX by minimising jump distances. One of them explored the combination of basic heuristics known as the nearest insertion (NI) and the 2-opt (named NI2OPT) with a high computing load. A second one presented an optimisation framework that decomposes and simplifies the problem but had a limited implementation. This work presents a better formal description of the tool-path optimisation problem, refines a data clustering and collapsing pre-processing approach, and implements a new method combining the NI and the 2-opt heuristics. Compared with the previous NI2OPT, the results showed that the new approach was much more efficient due to a considerable reduction in computing time, which could reach up to 88% for cases with a high amount of data (part or build geometric complexity). This result was achieved while keeping or slightly reducing the jump distance (i.e . building time).