Non-Destructive imaging of HVDC Cable using sub-THz CSAR with Robotic Arm

Sammanfattning

HVDC cables are difficult to inspect internally, as defects in the insulation can lead to malfunction or reduced service life. This thesis investigates whether sub-terahertz circular synthetic aperture radar (CSAR) can be used to image the cross-sectional structure of HVDC cables, with particular focus on cable joints, which are manually assembled and therefore have an increased risk of internal defects. By combining a Frequency-Modulated Continuous Wave (FMCW) radar with an ABB GoFa CRB 15000 cobot, measurements are performed and used to reconstruct images of the cable structure. The work is based on the theory of FMCW radar, CSAR, backprojection-based image reconstruction, linear interpolation, and robot kinematics. This theoretical foundation is used to design a CSAR system for imaging cable samples. Furthermore, practical aspects affecting image quality are addressed, including alignment of the antenna phase center relative to the robot’s tool center point, as well as trajectory planning for circular scans around the cable. The results show that sub-THz CSAR is a promising method for non-destructive inspection of HVDC cables. The method combines the penetration capability of sub-THz radiation with the improved resolution provided by synthetic aperture processing, as well as the flexibility offered by industrial robotics. At the same time, the results indicate that image quality is highly dependent on accurate robot motion, stable radar positioning, and precise calibration of the antenna phase center. Two scanning methods were evaluated. In the first method, a continuous sweep was performed with a distance of approximately 60 cm between the radar phase center and the cable surface. A total of 61 measurements were acquired over an angular range of 60°. In the second method, discrete measurement points were used, where the radar was stationary at each position. This method resulted in 69 measurements over the same angular range and distance. The measured data were used to generate range profiles and reconstructed CSAR images, both of which indicate reflections from the XLPE insulation as well as the copper conductor. The reconstructed images exhibit asymmetry and artefacts, which can be explained by the limited aperture of 60° compared to a full 360° scan. Despite this limitation, both the insulation and the conductor are clearly visible in the images. In conclusion, this work demonstrates that the combination of an industrial robot and radar enables non-destructive inspection of HVDC cable insulation using CSAR. Future work should focus on enabling full 360° scanning, for example by elevating the robot from ground level, using a robot with a larger reach, or implementing a radar system with a shorter far-field distance between the radar and the cable.