A solution to modeling the gas tungsten arc welding process using a nonconventional technique is presented. This approach, a nonlinear modeling technique using neural networks, has exhibited the potential to learn to model the time responses of a nonlinear, multivariable system. The feasibility of this approach as an alternative to existing techniques is examined. Potential problems with this approach are discussed. A control architecture using a second neural network is suggested.
|Number of pages||6|
|Publication status||Published - 1989|
|Event||Proceedings of the 1989 American Control Conference - Pittsburgh, PA, USA|
Duration: 21 Jun 1989 → 23 Jun 1989
|Conference||Proceedings of the 1989 American Control Conference|
|City||Pittsburgh, PA, USA|
|Period||21/06/89 → 23/06/89|