Doctor of Philosophy (PhD)
Grating-based interferometry/tomography is being rapidly developed for non-destructive evaluation of additive manufacturing test articles. An application requiring an eﬃcient workﬂow is extremely necessary for stress and fatigue testing samples.
At present, scientiﬁc workﬂows play an important role for computational experiments in additive manufacturing 3D printing and interferometry/tomography imaging analysis. A clear workﬂow template allows scientists to process experiments easier and faster. Work ﬂow library grows, but to ﬁnd an appropriate workﬂow for their task is challenging. In our research, there are mainly three portions in the workﬂow, interferometry analysis, image reconstruction and 3D visualization. Currently, the hierarchy of workﬂows in interferom etry/tomography projects is Mathematica, TomoPy/ASTRA/Jupyter notebook, VisTrails and Dragonﬂy. In general, two methods of interferometry analysis have been used in the ﬁrst portion of workﬂow, single-shot interferometry and stepped-grating interferometry. As for the second portion, with a Jupyter notebook, the reconstruction method ’Gridrec’ in TomoPy and ’SIRT’ (Simultaneous Iterative Reconstruction Technique) in ASTRA gener ated a powerful reconstruction volume for absorption projections and dark-ﬁeld projections separately. For the last portion, Dragonﬂy developed by ORS (Object Research System) company is a 3D visualization software with powerful scripting capabilities implemented in Python macros. Meanwhile, the VisTrails workﬂow incorporated both interferometry anal ysis and image reconstruction portions into VisTrails modules. Workﬂows in VisTrails hide much of the complexity of Mathematica or Python programming from users. Instead, with a simple GUI, it is possible for users to make their interferometry/tomography workﬂows through VisTrails modules.
Especially, for DPC (diﬀerential phase contrast) images in grating-based interferome try/tomography, we addressed the phase unwrapping issue with the method of 2D integra tion through generating phase images. With the algorithm, we have demonstrated the 2D integrated phase images denote a clearer contrast than DPC images.
Yuan, Jumao, "Workflows For X-ray And Neutron Interferometry/Tomography As Applied To Additive Manufacturing" (2018). LSU Doctoral Dissertations. 4569.