Method of setting reference value using artificial neural network of arc welding device

摘要

The present invention relates to a method for tracking a welding line during arc welding, wherein the welding factors are used as input values of the artificial neural network, and the artificial neural network is trained using the weaving direction and depth direction current recognition amount during the welding line tracking as the output value of the artificial neural network. Making a first step; After performing the first step, the artificial neural network outputs a current recognition amount for welding seam tracking using the welding factors set by the operator as an input value, and sets the output current recognition amount as a reference reference value for welding seam tracking. Two steps; After performing the second step, the welding is performed, and the current data measured during welding is calculated as the current recognition amount for tracking the welding line, and the welding line is in a direction approaching the reference value according to the difference between the calculated current recognition amount and the reference reference value. And a fourth step of correcting tracking.;That is, according to the present invention, it is possible to reliably weld seam tracking by setting a reference reference value using an artificial neural network, and do not need to obtain a complicated experimental relationship, thereby simplifying the seam tracking system.