CORDIS provides links to public deliverables and publications of HORIZON projects.
Links to deliverables and publications from FP7 projects, as well as links to some specific result types such as dataset and software, are dynamically retrieved from OpenAIRE .
Exploitable results
Laser spot welding is widely used in production with currently over 200,000 laser welding cells in use. This number is increasing by 10-20% per year. A typical application area for micro-spot welding is the assembly of micro-mechanical components and semiconductors. There are a number of advantages to be had from using this technology as opposed to for example resistance, ultrasonic or friction welding, namely: no special tools are needed, the heat applied is strictly local, it is suitable for miniature welds and the processing speed is very fast.
The basic problem associated with this technology is how to attain the part per million defect level needed for the wider industrial application of laser spot welding in micro-mechanics and electronics. This project addresses the two most important defect sources; namely those due to position tolerances of the products and those resulting from variations in the surface conditions. The MAIL project was started to improve the reliability of micro-spot laser welding through the development of a self-learning, real-time feedback control system. The basic innovations in the project include:
-An "aim and shoot" system which addresses the problem of position tolerances, by recognizing the joint positions and adjusting the actual laser spot position accordingly.
-A generic multi-sensor real-time monitoring system takes care of the defects which result from the surface variations. This system contains modules for advanced pattern recognition used for joint classification, adaptive control based on process modelling and a self-learning neural network and real-time closed loop control of the beam energy.
Phase one of the project was concerned with the investigation into relevant parameters for joint assessment and classification. Several sensing techniques have been tested and evaluated. A subset of sensing techniques passed the evaluation phase and is used for generating a broad process related signal database for each of the four different carriers used in the project. Automatic classification has been based on the condensed data of these databases using neural networks. The results from these efforts are very promising.
In phase two, adaptive feedback has been introduced in the project. This control loop is expected to compensate for slowly changing process parameters (focal distance, laser efficiency, pollution etc.) by adapting the laser parameters from pulse to pulse in case of necessary changes. Large volume experiments have been done to determine the relevant signals for process control and to find the significant process relations. Neural networks have been a valuable tool here to locate.
Searching for OpenAIRE data...
There was an error trying to search data from OpenAIRE
No results available