In planning to use automatic welding processes for remote weld repair applications, particularly if they are to replace more flexible manual welding techniques, use of sensing technology for seam tracking, processing monitoring and adaptive control should also be considered. Sensing techniques, in-process control and offline programming techniques can all be used for automatic or robotic welding.
Camera systems/teach and playback
Using this technique, the welding head is positioned adjacent to the joint and the motion of the welding head is 'taught' while viewing the joint area by remote camera. Once the motion of the head in relation to the joint-line is fixed, the weld can be made. The camera system can also be used, after welding, for visual inspection. This approach is suitable for short, repetitive weld runs.
Seam tracking systems rely on (laser) vision-based devices or other signal output (such as arc voltage or infrared emissions) to correct the position of the welding head relative to that taught earlier (e.g. by using a remote camera). This path modification, or seam tracking, takes place in real time, during the welding operation.
A seam-tracking device would commonly be used for longer operations (weld runs) where the joint might shift (during welding) or path teaching costs become prohibitive. This technique is commonly used for essentially straight or curvilinear weld runs which vary slightly in position and length. It is more difficult to apply where the joints involve many sharp changes in direction.
Seam tracking techniques require a well defined joint feature to give a significant and measurable signal which can then be processed and followed. Consequently, some systems are not easily applied to closed, square-edge butt joints.
Vision-based seam tracking sensors can also be used after welding for inspection, e.g. of weld profiles and detection of some types of surface breaking defects.
In-process monitoring and adaptive control
Welding process monitoring techniques (e.g. monitoring arc current, voltage, travel speed, beam power, surface temperature, or process emissions) can be used to provide quality assurance data to show that nothing untoward has happened during welding and hence infer the integrity of a weld, to highlight where process deviations have momentarily strayed out of limits, or even to halt the process altogether should deviations become permanently out of limits.
As an extension to this, seam tracking sensors can also be used to maintain weld quality during fabrication or weld repair, by monitoring joint variations in real-time during welding, and then adjusting welding parameters (e.g. travel speed) accordingly. This is the concept of adaptive control.
Offline programming is a fully automatic concept. Known data about the structure and the joint, typically CAD data, is used to determine the path to be followed by the welding head manipulator. This path is generated offline, in the computer, without the need to 'teach' using the actual processing machine.
The process operating system would then normally use one or more of the above tracking and monitoring techniques because of differences between theoretical and actual structures. In practice, it would be normal to carry out a dummy-run to confirm the motion of the welding head relative to the joint-line.
Offline programming allows torch to joint-line angles to be set precisely, which is difficult to do with other remote teaching approaches. Offline programming techniques have most benefit for larger, one-off, complicated structures.
Please note that the following items are only available to Industrial Members of TWI -
Adaptive control of multipass MAG welding
Adaptive control for hybrid welding 8mm thick steel plate
A system for integrated off-line programming of welding robots