Reactive coverage planning for robotic NDT of complex parts

Speaker:
Eitzinger, Christian; Profactor GmbH; Austria

Authors:
Eitzinger, C.; Profactor GmbH; Austria
Walch, A.; Profactor GmbH; Austria
Palfinger, W.; Profactor GmbH; Austria
Beyer, S.; Profactor GmbH; Austria
Meyer-Heye, P.; Profactor GmbH; Austria

ID: ECNDT-0048-2018
Download: PDF
Session: Robotic and automation 2
Room: H2
Date: 2018-06-14
Time: 09:20 - 09:40

Whenever parts of complex shape need to be tested using thermography, X-ray or other image-based inspection technologies, it is necessary to move the sensor system over the whole part to fully scan the surface. This is typically done using a robot that either moves the part or the sensor. In order to define the motions of the robot, a path needs to be calculated so that the part’s surface is fully covered by the inspection while avoiding collisions. Additionally, variations of the part in terms of shape or material properties may require an adjustment of a previously planned path while the inspection is being done. We propose a path planning method that uses a tree-based approach with local optimization to plan and update the path during the inspection process. A pre-defined vector field of preferred scanning directions is used to ensure that the local optimization does not deviate too much from the global scanning strategy. While the robot is moving continuously, the currently covered surface area is calculated and the path is updated. By locally optimizing the area covered through the next robot position(s) and by including the global scanning strategy defined by the vector field, a decision is made which path to follow from the current position.
Results of the path planning method are presented for experiments with thermography of forged steel parts and X-ray inspection of composite parts. Tests indicate that a locally planned path leads to inspection times that are reasonably close the theoretical lower limit.