Project Code: 32850
Start date and planned duration: January 2019, 36 months
- Perform a state-of-the-art literature review to determine current capabilities and possible avenues for future development of technologies for manual adjustment of robotic welding processes.
- Conduct interviews and human factors analysis with skilled welders in the workplace.
- Develop a vision monitoring feedback system that will monitor variations in weld pool geometry.
- Develop an integrated alternative control method for in-process modification of robot weld path.
- Integrate and calibrate the vision and control systems.
- Validate welding-capable modification of tool path, benchmark weld quality, and verify welding outputs against skilled welders.
Skilled manual welders make continuous adjustments based on what they see and hear as the weld progresses. The aim of this project is to allow manual intervention for robotic welding processes, by relatively unskilled staff, in order to carry out continuous adjustments during welding, for example manual adjustment of a rough welding path to accommodate part tolerances.
The project envisages an alternative control method for in-process modification of robotic tool paths, to account for variations in part geometry and to minimise the amount of programming necessary for low-volume customised batch manufacture.
This will be tied to a vision quality system intended to act as an alerting system for weld quality, based on measurements of the weld pool using brightness thresholding.
Benefits to Industry
The increasing volume of work being performed by robotic welding systems requires multiple weld trials using an iterative process, consuming an example component for each trial. If the tool path can be modified in-process, this will reduce development time as well as wastage. Additionally, improved remote control, allowing manual adjustments, of robotic and mechanised systems, will be of value in providing welding solutions, such as repair, which must be carried out in environments that are unsafe for human operators.