Offline correction of tool path deviations for robot-assisted ultrasonic nondestructive testing

Introducing industrial robots into ultrasonic nondestructive testing systems makes it possible to test more complex structures. However, the tool path planned in offline programming software may not fit with the actual path in practice, reducing the detection resolution of the system. This paper describes a robot-assisted ultrasonic nondestructive testing system and presents an approach to correct tool path deviations. A feature contained in normal vector angles is used to describe three-dimensional point clouds, in which a matching algorithm is used to calculate the deviations between two point clouds. In this system, the point cloud is measured by an ultrasonic probe, based on the time-of-flight principle. A metal part is used in experiment to demonstrate the effectiveness of the approach. Distance deviations of 0.3-3.0 mm can be corrected. With a detection resolution of approximately 0.3 mm, artificial defects 1.0 mm in diameter and 15.0 mm in depth can be seen more clearly in C-scan images after correction.