Characterization of micro-structural features in complex parts for use in digital technologies
Topic(s) : Special Sessions
Co-authors :
Vincent MAES (UNITED KINGDOM), James KRATZ (UNITED KINGDOM)Abstract :
While the commercial use of structural composites has grown in recent decades, the overall defect sensitivity and limitations on production rate are common problems that restrict the viability of composite components in many applications. Where composites are used, the weight savings achieved are generally much lower than the theoretical savings that could be achieved on the basis of the pristine properties. While some level of variation is innate to the raw materials, a significant reduction in performance is caused by the deviations in the as-manufactured part where a mixture of resin rich regions, voids, thickness changes, and fibre waviness and wrinkling occur during the forming and consolidation steps.
In order to certify parts, non-destructive inspection (NDI) techniques are needed to detect and characterise them where they occur. However, the variation in material systems and their physical properties (e.g. electric conductivity) as well as the convolution of the signatures of different defects complicates standardization and calibration for various tools that have been developed, such as ultra-sound and eddy current sensors. Combined with a limited understanding of the full range of possible defect morphologies and their relevance to performance, characterisation is historically limited and not sufficient for high fidelity recreations possible in simulations. [1]
In recent years the scaling up and proliferation of X-ray Computed Tomography (XCT) has offered a new window into the inner architecture of cured parts at resolutions not previously possible. However, the data generated is too large to be easily stored, shared and handled for further use. Hence a scheme for data reduction into a standardizable database is needed to get the true value of the XCT scans. To achieve this the high-resolution 3D volume is converted into discrete metrics capturing the location and morphology of the feature which can then be used for digital verification.
As part of the CerTest programme, which is developing various tools required to enable digital certification of aerospace structures, this work presents a case study of a C-spar with inclusions, wrinkles, and voids. The c-spars are about 0.5 m long with a recessed waist section in the middle. Multiple copies of the part were manufactured using hand lay-up and autoclave curing of IM7/8552 prepreg with a quasi-isotropic lay-up of 6 mm thick. The parts were scanned with X-ray CT globally at 80 μm resolution and locally at key regions at 40 μm resolution using a Diondo d5 system at the University of Southampton. From this CT data, metrics of the various defects are extracted and stored in a database structure that can then be used to feed a multiscale FE framework used to build digital twins of the parts.
In order to certify parts, non-destructive inspection (NDI) techniques are needed to detect and characterise them where they occur. However, the variation in material systems and their physical properties (e.g. electric conductivity) as well as the convolution of the signatures of different defects complicates standardization and calibration for various tools that have been developed, such as ultra-sound and eddy current sensors. Combined with a limited understanding of the full range of possible defect morphologies and their relevance to performance, characterisation is historically limited and not sufficient for high fidelity recreations possible in simulations. [1]
In recent years the scaling up and proliferation of X-ray Computed Tomography (XCT) has offered a new window into the inner architecture of cured parts at resolutions not previously possible. However, the data generated is too large to be easily stored, shared and handled for further use. Hence a scheme for data reduction into a standardizable database is needed to get the true value of the XCT scans. To achieve this the high-resolution 3D volume is converted into discrete metrics capturing the location and morphology of the feature which can then be used for digital verification.
As part of the CerTest programme, which is developing various tools required to enable digital certification of aerospace structures, this work presents a case study of a C-spar with inclusions, wrinkles, and voids. The c-spars are about 0.5 m long with a recessed waist section in the middle. Multiple copies of the part were manufactured using hand lay-up and autoclave curing of IM7/8552 prepreg with a quasi-isotropic lay-up of 6 mm thick. The parts were scanned with X-ray CT globally at 80 μm resolution and locally at key regions at 40 μm resolution using a Diondo d5 system at the University of Southampton. From this CT data, metrics of the various defects are extracted and stored in a database structure that can then be used to feed a multiscale FE framework used to build digital twins of the parts.