Abstract

The production of novel types of complex shapes is nowadays enabled by new manufacturing paradigms such as additive manufacturing, also known as 3D printing. The continuous increase of shape complexity imposes new challenges in terms of inspection, product qualification and process monitoring methodologies. Previously proposed methods for 2.5D free-form surfaces are no longer applicable in the presence of this kind of new full 3D geometries. This paper aims to tackle this challenge by presenting a statistical quality monitoring approach for structures that cannot be described in terms of parametric models. The goal consists of identifying out-of-control geometrical distortions by analyzing either local variations within the part or changes from part to part. The proposed approach involves an innovative solution for modeling the deviation between the nominal geometry (the originating 3D model) and the real geometry (measured via x-ray computed tomography) by slicing the shapes and estimating the deviation slice by slice. 3D deviation maps are then transformed into 1D deviation profiles enabling the use of a profile monitoring scheme for local defect detection. The feasibility and potential of this method are demonstrated by focusing on a category of complex shapes where an elemental geometry regularly repeats in space. These shapes are known as lattice structures, or metamaterials, and their trabecular shape is thought to provide innovative mechanical and functional performance. The performance of the proposed method is shown in real and simulated case studies.

摘要

如今，新型复杂形状的生产是通过新的制造模式实现的，例如增材制造，也称为 3D 打印。形状复杂性的不断增加对检验、产品鉴定和过程监控方法提出了新的挑战。先前提出的 2.5D 自由曲面方法不再适用于这种新的全 3D 几何形状。本文旨在通过为无法用参数模型描述的结构提出一种统计质量监测方法来应对这一挑战。目标包括通过分析零件内的局部变化或零件之间的变化来识别失控的几何变形。所提出的方法涉及一种创新的解决方案，用于通过对形状进行切片并逐层估计偏差来对标称几何形状（原始 3D 模型）和真实几何形状（通过 X 射线计算机断层扫描测量）之间的偏差进行建模。然后将 3D 偏差图转换为 1D 偏差轮廓，从而能够使用轮廓监控方案进行局部缺陷检测。这种方法的可行性和潜力通过关注一类复杂形状来证明，其中元素几何在空间中有规律地重复。这些形状被称为 然后将 3D 偏差图转换为 1D 偏差轮廓，从而能够使用轮廓监控方案进行局部缺陷检测。这种方法的可行性和潜力通过关注一类复杂形状来证明，其中元素几何在空间中有规律地重复。这些形状被称为 然后将 3D 偏差图转换为 1D 偏差轮廓，从而能够使用轮廓监控方案进行局部缺陷检测。这种方法的可行性和潜力通过关注一类复杂形状来证明，其中元素几何在空间中有规律地重复。这些形状被称为晶格结构或超材料及其小梁形状被认为可提供创新的机械和功能性能。所提出的方法的性能显示在真实和模拟案例研究中。