Towards defect monitoring for metallic additive manufacturing components using phased array ultrasonic testing

Additive manufacturing (AM) is a rising technology bringing new opportunities for design and cost of manufacturing, compared to standard processes like casting and machining. Among the AM techniques, direct energy deposition (DED) processes are dedicated to manufacture functional metallic parts. Despite of their promising perspectives, the industrial implementation of DED processes is inhibited by the lack of structural health control. Consequently, non-destructive testing (NDT) techniques can be investigated to inspect DED-manufactured parts, in an online or offline manner. To date, most ultrasonic NDT applications to metallic AM concerned the selective laser melting process; existing studies tackling DED processes mainly compare various ultrasonic techniques and do not propose a comprehensive control method for such processes. Current researches in the GeM laboratory focus on a multi-sensor monitoring method dedicated to DED processes, with a structural health control loop included, in order to track defect formation during manufacturing. In this way, this paper aims to be a proof of concept and proposes a comprehensive control method that opens the way to in situ ultrasonic control for DED. In this paper, a control method using the phased array ultrasonic testing (PAUT) technique is particularly illustrated on wire-arc additive manufacturing (WAAM) components, and its applicability to laser metal deposition (LMD) is also demonstrated. A specific attention is given to the calibration method, towards a quantitative prediction of the size of the detected flaws. PAUT predictions are cross-checked thanks to X-ray radiography, which demonstrates that the PAUT method enables to detect and dimension defects from 0.6 to 1 mm for WAAM aluminum alloy parts. Then, an applicable scenario of inspection of a WAAM industrial and large-scale part is presented. Finally, perspectives for in situ and real-time application of the chosen method are given. This paper shows that real-time monitoring of the WAAM process is possible, as the PAUT method can be integrated in the manufacturing environment, provides relevant in situ data, and runs with computing times compatible with real-time applications.