and mathematical modeling. The first computational study (Ref. 13) of the effects of surface-active elements in welding was published by a team at MIT in 1983 that investigated convection in arc weld pools. They showed that, in many cases, the surface tension driven flow dominates the convection in the arc weld pool. But more important, they also proved that a small amount of sulfur or selenium influenced the direction and magnitude of the liquid metal flow, a behavior the Rocky Flats team observed experimentally. These calculations, with assumed weld pool shape and size, were a giant step forward because they provided a world of insight that could not be obtained by any other means. Only the Numbers Reveal the Whole Truth The influence of selenium or sulfur depends on the mechanism of heat transfer which, in turn, is determined by the magnitude of the velocities and the thermal conductivity of the liquid metal. If the velocities are small for the conditions of welding, the direction of liquid metal circulation does not affect the shape of the weld pool. As a result, sulfur and other surfaceactive elements do not always affect the weld pool shape. Only comprehensive computer simulations can reveal the velocity fields and the mechanism of heat transfer. Such calculations show when surfaceactive elements affect weld pool geometry and when they do not. Two welding experiments and their computer simulations are presented here to show how the same level of sulfur may or may not affect the weld pool geometry depending on the welding conditions. A side by side comparison of experimentally determined and numerically computed weld pool cross sections (Ref. 11) of two high-power laser spot welds of steels is shown in Fig. 9. For a sulfur content of 20 ppm, calculations show a strong convection current transports liquid metal from the middle of the weld pool sideways. The surface velocities are fairly large, on the order of about 20 cm/s, and at this velocity, convection is the main mechanism of heat transfer. In comparison, heat transfer by conduction is negligible (Ref. 11). The molten metal flows 62 WELDING JOURNAL / APRIL 2015 Fig. 9 — Experimentally determined and theoretically calculated weld pool geometries in a 15-mm-thick, high-speed steel plate spot laser welded for 5 s. The welds had 20 and 150 ppm sulfur on the left and right sides, respectively (Ref. 11). sideways from the middle, forming a shallow weld pool, as shown in Fig. 9A. When the sulfur content is 150 ppm, the circulation pattern is opposite to what was observed for the 20 ppm sulfur steel weld shown in Fig. 9B. The surface velocities are fairly large, higher than 20 cm/s. So the heat is carried mostly by convection, and conduction heat transfer is unimportant (Ref. 11). Hot weld metal flows downward under the heat source, the base metal melts near the root, and a deep weld pool forms. The computed weld pool geometry agrees well with the experimentally determined geometries in both cases. But sulfur does not always change the weld pool geometry (Ref. 11). Figure 10 shows no perceptible difference in the cross sections of low-power laser welds in steel containing 20 and 150 ppm of sulfur. The numerical simulation of heat transfer and fluid flow reveals why. The computed results show fairly low peak temperatures and lower velocities in the weld pool for these small welds. Convection did not carry much heat since the velocities in both cases were weak. As a result, conduction was the main mechanism of heat transfer. The direction of spin of the weld metal was opposite in the two cases as expected, but since conduction was the mechanism of heat transfer, the opposing spin did not result in any difference in geometry (Ref. 11). The mechanism of heat transfer was the most important factor, not the concentration of sulfur or the direction of weld metal spin. Dramatic effects of sulfur, selenium, and other surface-active elements known for many decades led many to believe these elements always affected weld pool geometry. In fact, only when convection is the dominant mechanism of heat transfer can the surfaceactive elements play an important role in affecting weld geometry. Solving a compelling problem of lack of reproducibility of the weld geometry has made the world a safer place for all people. But does the innovation and discovery stop once a longstanding mystery is solved? Not at all, because new welding problems that affect life and property arise frequently. The following example shows that the Fig. 10 — Comparison of the computed and experimental weld pool geometries at a laser power of 1900 W for steels containing A — 20 ppm; B — 150 ppm sulfur (Ref. 11).
Welding Journal | April 2015
To see the actual publication please follow the link above