CAD Model-based Robotic Pulse-echo Ultrasonic Propagation Imaging for 3D Printed Composites and Data Acquisition for AI-based Defect Diagnosis

Three-dimensional (3D) printed composites are increasingly being used in unmanned aerial vehicle (UAV) designs, particularly for secondary structures with complex geometries. Subsurface defects can occur during manufacturing, requiring post-production nondestructive testing (NDT) to ensure structural integrity. This study proposes a robotic pulse-echo ultrasonic propagation imaging (PUPI) system that uses laser ultrasonics for the effective inspection of 3D printed composites. The system controls a 6-axis robot arm to follow the scan paths generated from computer-aided design (CAD) geometric data, allowing pulse-echo mode laser ultrasonic testing (LUT) of specimens with complex, irregular shapes that defy traditional mathematical representations. A 3D printed UAV winglet prototype composed of honeycomb structures reinforced with continuous carbon fibers was inspected using this approach. The inspection results successfully visualized the internal composite features and demonstrated the capability of the system to perform a CAD-model-based scan path. To aid in AI-based defect diagnosis, training blocks were designed and 3D printed with varying thicknesses to represent the possible infill patterns. The training blocks were inspected using the robotic PUPI system to acquire the corresponding 1D ultrasonic signals across the thickness variations. The collected data allowed for the classification of ultrasonic signal propagation in the thickness direction, which supports the identification of debonding depths in 3D printed specimens. The AI model training based on these data is still ongoing.