Posted on 4 October 2022 by David Harman
Jet engine turbine blades are required to operate in some of the most extreme conditions, where operating temperatures may exceed the melting point of the alloys used. As a result, complex cooling geometries, coatings, and other properties are used to keep blades within operational temperature, with failure not being an option. However, manufactured parts may contain small embedded and hidden defects, creating a need to use a methodology such as CT inspection to inspect the inside of these parts.
CT inspection is a nondestructive industrial x-ray process, where a part is rotated 360 degrees while hundreds or thousands of individual 2D radiographs are captured at specific degree intervals. This collection of 2D radiographs are then reconstructed into a 3D CT volume which can be digitally sliced through at any angle, allowing users to inspect both the external and internal features of the product without needing to physically cut or open their product.
However, one challenge with this workflow is that the ability to resolve defects and dimensions within the part is dictated by overall image quality and image quality can be directly correlated with inspection cycle time. High density and complex parts, like turbine blades, typically require a significant amount of scan and technician interpretation time. This can create a potential bottleneck for high-throughput applications.
Synopsys Simpleware Custom Modeler offers a solution to the challenge of bottlenecks in manual image review through case-specific analysis with AI-enabled Machine Learning methods. With this approach, time-consuming manual segmentation time is eliminated from the work, and inspection time is reduced by the pre-determination of defect locations. Further scaling up of inspection is therefore only dependent on the power of the hardware being used for the inspection. In collaboration with Avonix Imaging, a project was developed to detect low-frequency critical defects in high-pressure turbine blades, with the goal of creating a fully automated inspection assistant to highlight possible defect regions, and to speed up critical inspection workflows.