Identifying weld pool dynamics for gas metal arc fillet welds

AbstractArc welding processes are conventionally modelled under static (steady state) conditions to provide a relationship between the controllable process parameter values and the characteristic weld geometry parameters that result. While this is appropriate for open loop control of the process, should closed loop (feedback) control of the process be required, the dynamic (or transient) response must also be identified. In the present paper, the crosscorrelation approach for system identification is evaluated for developing a dynamic process model of the relationship between welding voltage and weld pool width. Based on the identified impulse response, a second order process model is assumed. The non-linearity of the model is demonstrated and the gain, damping, and natural frequency are estimated.