Development of a fast inspection system for complex aerospace structure

To build lighter, more fuel efficient aircraft, industry has had to adopt new technologies at an increasing rate. Lighter, stronger, corrosion resistant structures are possible through the use of advanced composite materials, while the shortage of metals such as titanium have driven the development of additive manufacturing methods to reduce buy to fly ratios. These technologies enable the manufacture of intricate shapes in one operation, while at the same time varying material properties throughout the part to meet local loading conditions. This presents huge challenges for Non-Destructive Testing (NDT). Intricate geometries require time consuming manual inspection. Varying properties invariably mean that components are highly anisotropic making ultrasonic testing (UT) difficult due to changing acoustic velocities and high levels of structural noise. Couple this with the need to inspect every part and the
NDT process can become a bottleneck to production throughput. This paper describes the development of a robotic inspection system aimed at reducing the time to inspect components with complex geometry. The heart of the system is a cell comprising two 6-axis robotic arms each capable of working independently and cooperatively. By reading in CAD data the system is able to manipulate transducers around highly complex shapes at high speed and with great accuracy. Phased array ultrasonic testing (PAUT) or full matrix capture (FMC) algorithms have been developed that allow wide swathes of data to be acquired in a single pass, while simultaneously coping with surface curvature and features such as radii. Custom software allows the inspection of parts with highly variant thickness by referencing each A-scan to the CAD data where front and rear interfaces cannot be reliably detected. An algorithm combining aspects of FMC and the synthetic aperture focusing technique (SAFT) reduces the effects of coherent noise and acoustic velocities that vary with angle. Analysis and sentencing is made easier by displaying 3D data wrapped around a CAD wireframe. In addition to the usual A and B-scans, a rotatable and zoomable image shows features, not on the surface, but mapped in 3D within the part.