Turbine blades have complex shapes, including internal structures, making them difficult to inspect with most NDE techniques, including ultrasound. Radiographic CT is a very powerful tool for inspection, enabling a 2D cross section of the turbine blade to be produced. However, speed is a significant issue, with thousands of projections, each taking 0.1s, often needed to generate an image using traditional approaches. The quantity of blades produced by Rolls-Royce (one facility can produce over 2000 a week) also presents a challenge, because this acquisition approach must be repeated for each blade.

Limited data imaging provides an opportunity to significantly reduce acquisition time for CT inspection of blades, enabling much faster inspection and reducing the pressure to purchase multiple CT systems to increase throughput. More efficient use of CT systems can significantly reduce costs.

A variety of techniques are proposed, incorporating image processing algorithms, regularisation techniques, and the use of external information, to enable high quality reconstructions with less data.

Sponsors

EPSRC, RCNDE