Using "snow" as a cooling agent virtually eliminates weld distortion
Diode laser welding with CO2 cooling
The high heat inputs associated with thermal welding processes produce distortion that has always been an unwanted by-product. This distortion is caused by a number of effects, including differential shrinking, phase transformations and residual stress effects. These effects have been tolerated by the vast majority of welders as being an inevitable consequence of using a thermal welding process. A small number of techniques have been developed to control or limit this distortion, including pre-heating, pre-stressing, post-weld annealing, weld sequencing, copper cooling bars, strong backs and mechanical tensioning. The development of certain Low Stress No Distortion (LSND) techniques is also well documented in academic journals. But, perhaps due to their complexity, they have yet to make a significant practical impact.
The advent of laser welding was hailed as a major breakthrough in terms of weld-related distortion, because it is a high energy density, low overall heat input process. Although laser welding can produce relatively high-quality, low-distortion structures, residual distortion remains a problem when welding high-accuracy 3D components, particularly for applications such as those encountered in the aerospace industry.
When working with thin-sheet aluminium and titanium alloys, one of the key obstacles to wider adoption of fusion welding is the induced distortion. Although thicker materials exhibit less distortion, the large residual stresses generated by the welding process may lead to premature failure of the welded structure. This has been a serious hurdle to the wider adoption of welding as a joining process in the aerospace industry, particularly in airframes.
Figure 1. Relative cooling potential.
An industrial collaborative project has produced a promising implementation of LSND using cryogenic gases as the applied cooling medium. The use of solid CO2 has shown to be a promising solution due to the relative ease with which it can be applied. A supply of high-pressure CO2 liquid on expansion through a delivery nozzle forms a mixture of gas and solid CO2 snow. The amount of snow and, hence, the available cooling is very much dependant on the supply conditions of the CO2, due to the unique thermodynamic properties of this molecule. This gas/solid mixture is simply blown on to a point immediately behind the weld zone to enhance post-weld cooling. It is the extraction of the heat, by the sublimation of this snow, that makes CO2 such an effective coolant (see Figure 1), particularly when compared with other cooling media. Compressed air jets and water mist sprays have also been used, but both techniques suffer from limited effectiveness and, in the case of water, residual liquid often needs to be removed.
The use of cryogenic coolant has been successfully applied in thermal spraying applications where the interpass temperature needs to be controlled in order to prevent coating spallation due to differential thermal expansion rates between the coating and the substrate materials. It required only a small innovative step to transfer this cooling technology from coating to welding.
The use of cryogenic cooling has now been successfully applied to a number of thermal welding processes, including arc, laser and friction stir welding (FSW). With arc processes care must be taken not to interfere with the welding arc plasma. A physical barrier may be required to stop the cooling jet from affecting the weld fusion zone or, if space allows, the cooling can be applied from the underside or backside of the weld with equal efficacy. In laser and FSW this is less of a problem so practical implementation of this solution is potentially far easier.
Figure 2. Effect of weld cooling.
Extensive results have been obtained using weld cooling with both diode and CO2 lasers.1 Much of the analytical work has been conducted on stainless-steel welds produced with a high-power diode laser, and the results have been extremely promising. Using 1.5 kW at 808 nm focused to a 3 X 1.5 mm spot there is insufficient energy density for keyhole welding, but enough to obtain high-quality conduction welds in a range of materials, including stainless steels. Using this arrangement to make autogenous bead-on-plate runs on 304L 1.6 X 50 X 275-mm samples generally resulted in considerable distortion with the samples bowing by as much as 15 mm from the flat over the 275 mm sample length. When CO2 cooling was applied trailing the weld point at a distance of 30 mm, the distortion was significantly reduced and at 20 mm it was virtually eliminated (see Figure 2). The weld quality of both the cooled and non-cooled samples was similar with slight widening of the weld and slight variability in penetration observed when the cooling was used.
To assess the cooling impact some thermal field measurements were made using a thermal imaging camera. The thermal profile of a weld is typified by a long "comet" like heat trail, which reflects the relatively slow cooling rate achieved from self-quench conduction and ambient radiation. However, the results revealed the effectiveness of the CO2 as a cooling medium. The truncation of the heat trail of the weld by the CO2 can be seen to be complete and almost instantaneous. Temperatures of more than 700°C are reduced to <100°C over a distance of about 6 mm, which corresponds to the width of the cooling jet (see Figure 3). Interestingly, the thermal profile and the temperature gradient in the actual molten weld zone remain unaltered, implying that the forced cooling does not significantly affect the actual welding process in terms of the solidification of the weld pool.
Figure 3. IR camera measurements.
The use of this technique to change the microstructural properties of the material in the weld zone has also been demonstrated. This is of significant potential benefit for a number of regularly welded materials, however, it can potentially be a two-edged sword. The additional cooling can be extremely beneficial in the case of welding ferritic stainless steels where grain growth in the post-weld cooling phase can have significantly detrimental effects on the ductility and corrosion resistance in the solidified weld and heat affected zones. On the other hand, using increased weld cooling rates in some alloys can potentially increase the hardness and reduce ductility. Hence, each material and application must be considered on its own merit with regards to potential benefits that can be achieved by implementing weld cooling.
Another interesting benefit that was experimentally observed was the effect weld cooling had on weld cleanliness. Splatter can often be a significant problem particularly in applications where the aesthetic qualities of the weld are of primary importance to the product. Cleaning up the weld can be a time-consuming additional step that does not add any value. Using CO2 cooling, it was observed that splatter was far less likely to stick to the plate, and this was attributed to two main reasons. The first is that airborne spatter is effectively quenched before it actually hits the plate and, hence, bounces off. The second is that the cooling effect on the plate surrounding the weld means that any spatter that does impact in a molten state solidifies before forming a strong bond to the plate (see Figure 4).
Figure 4. Cleaning effect of CO2 snow.
Industrial applications of this technique are very much limited to automated welding processes because of issues arising from operator safety associated with cryogenic gases. However, provided that there is sufficient ventilation and extraction, the technique could be safely incorporated into most industrial situations.
With all new innovations cost of implementation is always a critical factor. Extensive studies have shown that CO2 consumption for weld cooling can vary between 1-5 lb/min depending on the specific nozzle arrangement. Some fully costed industrial examples have shown that cryogenic cooling can be applied for as little as $0.04 per linear foot of weld in thin sheet welding applications and capital cost starting from around $3000 for a simple system. This represents a potentially acceptable cost compared to the costs of rectification of distortion-related defects.
An effective LSND welding technique has been developed based on an intense CO2 snow cooling jet being directed just beyond the solidified material at the rear of the weld pool. The cooling jet introduces a "thermal tensioning" effect on the solidifying and cooling weld metal counteracting the compressive forces that lead to residual stresses and component distortion. Using this technique, the virtual elimination of weld distortion has been demonstrated on a wide range of materials and welding processes. Beneficial metallurgical effects have also been obtained.
The commercialization of this process by the aerospace industry requires a more detailed understanding of the process and the metallurgical effects on component properties. It is the subject of an ongoing UK government-sponsored project whose partners include BOC, BAE Systems, Qinetiq and Airbus UK. However, the implementation of this technology to other industry sectors is expected to be far more rapid.
Dr. Jack Gabzdyl is market development manager for BOC Ltd UK. Contact him at firstname.lastname@example.org.
- S.W.Williams et al., "Direct Diode Laser Welding of Aerospace Alloys", Laser 2101 Munich.
Some of the Figures supplied courtesy of BAE Systems.