| Autor: |
Muir, Peter J., Richards, Mark, Gray, Jaslyn, Self, Hamish W. A., Rosalie, Cedric, Rajic, Nik |
| Předmět: |
|
| Zdroj: |
EA National Conference Publications; 2023, p156-162, 7p |
| Abstrakt: |
Line scan thermography (LST) is a promising non-destructive inspection method that, when applied as part of a robotic scanning capability, can inspect aerospace platforms at relatively high speed compared to other methods. LST involves sweeping a line heat-source over an object and using an infrared imager to observe the surface temperature in the wake of the source. Subsurface defects are identified by the thermal contrast they produce. While laboratory implementations of robotic inspection allow for relatively tight control of scan motions, field applications, particularly those involving aerial robot deployment, are likely to be subject to significant uncontrolled perturbations due to external factors including wind gusts. To quantify the effect of positional instability on the detection performance of LST, this study compares the signal-to-noise ratio of defect signatures obtained for a range of path perturbations. Carbon fibre composite laminates containing synthetic flat-bottom hole defects were scanned using a 3-axis fixed Cartesian robot with the scan motion perturbed by inducing controlled periodic meandering. The results provide important design guidance with respect to the stability requirements for successful aerial drone-based deployment of LST. A machine learning algorithm developed to automate defect detection and localisation from LST data is also outlined. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
