Skip to main content

Analysis weld pool on-line adaptive weld process control hazardous environments & manufacturing applications using neural networks

Deliverables

The objective of the NEUROWELD project is to develop a fully automatic welding system based on the analysis of the weld pool for on-line adaptive weld process control using a Neural Network. It consists of developing an auto-adaptive system for integration on automatic welding systems, which will allow variations in groove form, dimensions and position, and parent metal variations to be taken into account in real time by the weld control system such that the torch position, trajectory and welding parameters can be automatically modified to avoid weld defects being produced, the variations of parameters having been determined as a function of a prior series of optimised tests and thus limited extents. The main benefit of NEUROWELD comes from the reduction of the number of weld defects and thus improving the security and reducing the repair cost as well as the costs of welding procedures development in laboratory by optimising the number of tests to be performed.
Fronius has developed a TIME-Twin power source for high speed welding applications. This equipment has been validated by ALSTOM-LHB in its production site at Salzgitter in order to produce the roof sections of the ICE high-speed train. The welding speeds achievable by the T.I.M.E-Twin process (100-200cm/min) are approximately two or three times higher than those of the so far common MIG single-wire welding process (40-80cm/min).