The L2 norm of the deviation between the measured and computed transient displacement field in a test weld

Abstract The transient displacement field caused by an arc weld was measured experimentally with an Aramis stereo-camera. The transient displacement field was also predicted using a computational weld mechanics program called VrWeld for ten different sets of model parameters. The objective was to identify the effect of model parameters on the deviation between experimental data and the data computed by computational weld mechanics. For each set of model parameters, the L2 norm of the difference between the experimental and computed displacement fields was computed for each of the 800 time steps. The structure being welded was a mm stringer fillet welded to a mm low alloy steel plate. The displacement field was measured on the ‘back’ surface of the plate. A robot made the metal-inert-gas weld. The parameters varied in the ten models included welding current, width of the power density distribution, convection coefficient and yield stress of austenite at high temperatures.