With the increasingly widespread use of Additive Manufacturing processes in industrial practice, the functional requirements for dimensional verification and tolerances have also changed. AM products are characterized by particularly complex geometries. Consider, for example, lattice structures; the basic idea is to fill the part's volume with a series of elementary cells formed by combining more or less simple geometries. Due to the radical difference between these types of products and those produced using more traditional technologies, the development of measurement processes suitable for functional requirements is necessary. The development of standards that meet the needs of AM processes is at the forefront of the work of various ISO and GPS committees. The focus of this research, carried out through an internship at the LURPA laboratory of the Ecole Normale Supérieur Paris-Saclay, is to compare the classical methods described by standard regulations with some promising alternative approaches. The methodology under examination is Virtual Image Correlation (VIC), which allows the reconstruction of geometry from images and subsequently derives measurements. The working principle involves comparing two images: the real physical image and a virtual image representing the nominal geometry, and then deforming the nominal geometry to better match the real geometry. The objective of this work is to assess how these methods perform when compared to current GPS standards, providing insights into new horizons in metrology applied to AM. In particular, three methods were examined during this research. The first method, representing current standards, is the ISO50% methodology followed by all the necessary data analysis steps to obtain the respective measurement results. The two proposed VIC approaches are based on contours defined by mathematical functions. The first one, called V2C and developed in other studies, decomposes the final contour into a finite series of modes representing specific geometric deviations. The second one, called DBACD and developed in the course of this research, defines the contour identified through a parametric B-spline curve. All three methodologies were tested on images obtained from computed tomography measurement machines of lattice structures with cylindrical geometries.
Con l'ampio utilizzo dei processi di Additive Manufacturing nell'industria, i requisiti funzionali per la verifica dimensionale e le tolleranze hanno subito cambiamenti significativi. I prodotti realizzati mediante AM presentano geometrie particolarmente complesse, come le strutture reticolari, che vengono create riempiendo il volume con celle elementari composte da geometrie semplici o complesse. A causa di queste differenze radicali rispetto alle tecnologie tradizionali, è necessario sviluppare metodi di misura adatti ai requisiti funzionali di tali prodotti. A tale scopo, diversi comitati ISO e GPS sono impegnati nello sviluppo di standard specifici per i processi di AM. Lo scopo di questa ricerca, condotta durante un tirocinio presso il laboratorio LURPA dell'Ecole Normale Supérieur Paris-Saclay, è confrontare i metodi tradizionali descritti dalle normative standard con approcci alternativi promettenti. In particolare, si è esaminata la metodologia della Virtual Image Correlation (VIC), che consente di ricostruire una geometria e ottenere misurazioni a partire da immagini. Questo metodo confronta due immagini: quella reale del componente e una virtuale che rappresenta la geometria nominale, cercando di deformare quest'ultima per adattarla alla geometria reale. L'obiettivo di questo lavoro è valutare le prestazioni di tali metodi rispetto alle normative GPS attuali e fornire spunti per nuovi orizzonti nella metrologia applicata all'AM. Durante la ricerca, sono stati esaminati tre metodi. Il primo metodo, rappresentativo degli standard attuali, è la metodologia ISO50%, seguita da tutti i passaggi di analisi dati necessari per ottenere i risultati di misura corrispondenti. I due approcci VIC proposti si basano invece su contorni definiti da funzioni matematiche. Il primo, chiamato V2C e sviluppato in studi precedenti, scompone il contorno finale in una serie finita di modi che rappresentano specifiche deviazioni geometriche. Il secondo, denominato DBACD, sviluppato durante questa ricerca, definisce il contorno tramite una curva parametrica di tipo B-spline. Tutti e tre i metodi sono stati testati su immagini ottenute da macchine di misura a tomografia computerizzata di strutture reticolari con geometrie cilindriche.
Analisi delle prestazioni e confronto di algoritmi di Virtual Image Correlation applicati a strutture reticolari fabbricate con processi di Additive Manufacturing
MIOLI, FILIPPO
2022/2023
Abstract
With the increasingly widespread use of Additive Manufacturing processes in industrial practice, the functional requirements for dimensional verification and tolerances have also changed. AM products are characterized by particularly complex geometries. Consider, for example, lattice structures; the basic idea is to fill the part's volume with a series of elementary cells formed by combining more or less simple geometries. Due to the radical difference between these types of products and those produced using more traditional technologies, the development of measurement processes suitable for functional requirements is necessary. The development of standards that meet the needs of AM processes is at the forefront of the work of various ISO and GPS committees. The focus of this research, carried out through an internship at the LURPA laboratory of the Ecole Normale Supérieur Paris-Saclay, is to compare the classical methods described by standard regulations with some promising alternative approaches. The methodology under examination is Virtual Image Correlation (VIC), which allows the reconstruction of geometry from images and subsequently derives measurements. The working principle involves comparing two images: the real physical image and a virtual image representing the nominal geometry, and then deforming the nominal geometry to better match the real geometry. The objective of this work is to assess how these methods perform when compared to current GPS standards, providing insights into new horizons in metrology applied to AM. In particular, three methods were examined during this research. The first method, representing current standards, is the ISO50% methodology followed by all the necessary data analysis steps to obtain the respective measurement results. The two proposed VIC approaches are based on contours defined by mathematical functions. The first one, called V2C and developed in other studies, decomposes the final contour into a finite series of modes representing specific geometric deviations. The second one, called DBACD and developed in the course of this research, defines the contour identified through a parametric B-spline curve. All three methodologies were tested on images obtained from computed tomography measurement machines of lattice structures with cylindrical geometries.File | Dimensione | Formato | |
---|---|---|---|
Mioli_Filippo.pdf
accesso aperto
Dimensione
7.68 MB
Formato
Adobe PDF
|
7.68 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/50801