Modelling of GMAW process using an abductive network

Abstract

An abductive polynomial network model of the gas metal arc welding (GMAW) process was established in this paper. This network model is capable of learning the relationship between GMAW process parameters (including wire diameter, gas flow rate, welding speed, arc current and welding voltage) and weld performance (i.e., weld bead penetration). The estimated value of weld bead penetration derived from network training was compared with the measured value. The comparison proves that the abductive polynomial network model established in this study possesses a certain degree of feasibility in GMAW process application.