Geração de caminhos de estruturas de suporte em manufatura aditiva que prioriza o desvio da peça a ser impressa

Abstract

Additive manufacturing (AM) is an alternative way of making parts that does not require specific molds or tooling. Building a piece in additive manufacturing is specific and involves successively adding material in layers, starting from the bottom to the top of the model. During the planning stage of the model via AM, regions that need support structures (suspended regions) are identified só that they can be manufactured properly. A simple way of generating the support structures is to have them oriented perpendicularly to the suspended regions, but this can result in support structures resting on the model of the part being printed. This situation causes problems in the post-processing of the produced part, since there is a need to sand, polish and remove imperfections caused on the part by support structures resting on the model. This work proposes ways of calculating support structure paths, with the aim of obtaining paths which deviate from the model, in order to reduce contact between the supports and the model, resulting in a reduction in post-processing time and possible imperfections in the production of the part. The calculation is based on a discretization of the model, separating it into layers that are discretized into regular grids. The model and the suspended regions are mapped onto the squares of the grids generated after discretization. From each square mapped as a suspended region, a search algorithm is used to find a path that deviates from the model and reaches the base. The search algorithms used are: depth-first search (DFS), breadth-first search (BFS), minimum path and simple deviation (proposed by the authors). The DFS algorithm proved to be the most cost-effective in terms of resource consumption (processing and memory) and efficiency in producing paths that deviate from the model. In addition, the minimum path algorithm resulted in the lowest material expenditure, but the difference was not significant. Finally, even though the simple deviation algorithm did not have the same capacity to generate paths that deviate from the model, it was the algorithm that showed the lowest processing and memory costs.