Arc welding seam tracking system based on artificial neuralnetworks

摘要

A new approach towards controlling an arc welding seam trackingsystem based on process-oriented data is presented; an artificial neuralnetwork is the central controlling element. Special features of neuralnetworks, such as trainability, abstraction capability andfault-tolerance give this controlling system significant advantages. Theadaptation of the system to new welding tasks has been simplifiedsubstantially in comparison with conventional approaches: analysing thenew process parameters and programming the system are replaced byproducing reference data from the new welding process and subsequentautomatic learning of an artificial neural network. Results on the basisof a quality function are presented