Form error analysis of lattice structures struts via VIC methods: towards geometry compensation strategies in Additive Manufacturing

Virtual Image Correlation is a boundary detection process based on creating a virtual image defined by a mathematical parametric equation that will then be deformed to best represent its real counterpart. The deformation is guided by minimizing a score function that compares the virtual and the actual images. This study aims to implement and fine-tune two VIC methods previously developed at LURPA in practical applications, comparing their performance with the current state of the art methods. The focus is on compensating form errors in Additively Manufactured Lattice Structure struts. The goal is to reconstruct the surface of a batch of struts with the same nominal geometry and manufacturing parameters, identifying and isolating stochastic errors to reveal the general ‘mean’ shape of the strut bench. The optimal manufactured geometry will be identified as the printed nominal geometry corresponding to the manufactured feature with the lowest form error. Comparative analysis will be conducted between the two VIC methods, featuring parametric contour, and discrete contour identification methodologies (e.g., ISO50%). The evaluation includes both quantitative assessments of form error results and a metrological evaluation of the pros and cons of the methodologies. Different case studies will be investigated considering different process parameters (e.g., scanning speed, laser power, layer thickness) and/or nominal geometry (e.g., diameter, position, orientation), to establish a correlation between form errors and manufacturing process parameters, thus enabling efficient geometry compensation strategies.