A Human Robot Collaborative Cell for automating NDT inspection processes

Bakopoulou, Katerina; Michalos, George; Mparis, Konstantinos; Gkournelos, Christos; Dimitropoulos, Nikos; Makris, Sotiris

doi:10.1016/j.procir.2022.10.076

Cited by 1 publication

(1 citation statement)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such systems are becoming increasingly feasible for NDT application due to recent innovations in collaborative robotics [2]. Katerina recently introduced a robotic human-machine collaborative system designed to fulfil the automation requirements of NDT measurement processes within the steel production industry [3]. Canzhi, through studying dual-robot synchronous motion systems, proposed a trajectory-planning -based solution to tackle NDT challenges specific to semi-enclosed regions of complex curved composite material components.…”

Section: Introductionmentioning

confidence: 99%

Efficient Autonomous Path Planning for Ultrasonic Non-Destructive Testing: A Graph Theory and K-Dimensional Tree Optimisation Approach

Zhang,

Sutcliffe,

Nicholson

et al. 2023

Machines

View full text Add to dashboard Cite

Within the domain of robotic non-destructive testing (NDT) of complex structures, the existing methods typically utilise an offline robot-path-planning strategy. Commonly, for robotic inspection, this will involve full coverage of the component. An NDT probe oriented normal to the component surface is deployed in a raster scan pattern. Here, digital models are used, with the user decomposing complex structures into manageable scan path segments, while carefully avoiding obstacles and other geometric features. This is a manual process that requires a highly skilled robotic operator, often taking several hours or days to refine. This introduces several challenges to NDT, including the need for an accurate model of the component (which, for NDT inspection, is often not available), the requirement of skilled personnel, and careful consideration of both the NDT inspection method and the geometric structure of the component. This paper addresses the specific challenge of scanning complex surfaces by using an automated approach. An algorithm is presented, which is able to learn an efficient scan path by taking into account the dimensional constraints of the footprint of an ultrasonic phased-array probe (a common inspection method for NDT) and the surface geometry. The proposed solution harnesses a digital model of the component, which is decomposed into a series of connected nodes representing the NDT inspection points within the NDT process—this step utilises graph theory. The connections to other nodes are determined using nearest neighbour with KD-Tree optimisation to improve the efficiency of node traversal. This enables a trade-off between simplicity and efficiency. Next, movement restrictions are introduced to allow the robot to navigate the surface of a component in a three-dimensional space, defining obstacles as prohibited areas, explicitly. Our solution entails a two-stage planning process, as follows: a modified three-dimensional flood fill is combined with Dijkstra’s shortest path algorithm. The process is repeated iteratively until the entire surface is covered. The efficiency of this proposed approach is evaluated through simulations. The technique presented in this paper provides an improved and automated method for NDT robotic inspection, reducing the requirement of skilled robotic path-planning personnel while ensuring full component coverage.

show abstract

Section: Introductionmentioning

confidence: 99%