Thesis-2000-Jaeckel.pdf (41.57 MB)
Download file

An optical sensor for joint recognition in multi-axis welding

Download (41.57 MB)
posted on 08.04.2014, 12:37 by Torsten Jaeckel
In the automated welding technology on three dimensional constructions, the welding tools will be proceed along the components of programmed handling systems in order to weld the components with each other. Tolerances while manufacturing result in the joints being able to be found at different positions. The deviations from the correct way are registered with the aid of sensors and transformed in control data for regulation of deviation. For this purpose, different models were developed by sensors which are mounted firmly in front of the welding tool and in such a way that it can recognize the deviations. In this case, the freedom of movement of the handling systems is limited at one degree of freedom. Abrupt direction changes lead to measurement errors so that the sensor leaves the measuring range. For this reason a new optical sensor was developed in this that is able to detect all necessary work parameters while welding. The sensor is based on the purposeful evaluation of structured illuminated projections around the welding point. The structured light occurs with a holographic projection system. An adapted preprocessing of data as well as a optimised rule-based recognition of the joint direction make also possible to detect complex welding joints in real time. The seam tracking system determines the position of the joint to be welded in the working area, the deviation of the welding head from the normal at the surface as well as the width of the gap between the components with a solution of 50 !Jm. The optical, mechanical, electronical and data processing components were developed for the high orders while welding with the laser. The use of this new seam tracking sensor allows the use of handling systems with only five axes at three dimensional joint geometries.



  • Mechanical, Electrical and Manufacturing Engineering


© Torsten Jaeckel

Publication date



A Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University.

EThOS Persistent ID



Usage metrics