Using Plasma Color to Predict Weld Quality

The Science

High-quality welding is critical to the strength and long-term stability of various structures, from bridges to nuclear reactors. However, monitoring weld quality during the welding process can be costly and may not take into consideration the chemistry of these welds. Researchers investigated an approach to correlate the light emitted by the weld flash with the composition of the metal being removed from the melt pool. This information can be used to assess the composition and quality of the weld. They also noted that the particle’s shape changed based on the sampling height, indicating that the laser beam remelts emitted particles.

The Impact

Welds with multicomponent alloys are becoming increasingly important as these complex materials are more broadly used. Finding ways to assess weld quality during the welding process can lead to better additive manufacturing processes. This work represents a necessary first step in developing real-time feedback on weld quality, essential for creating autonomous additive manufacturing systems. Future work will incorporate machine learning algorithms to analyze and rapidly make decisions based on the generated data.

Summary

Controlling multicomponent alloys during welding is challenging because scientists need a real-time understanding of their composition. The optical emissions of the plasma formed during laser-induced metal welding correlate with the composition of particles ejected from the melt pool. Studied plasma emissions contain large signatures of iron, manganese, chrome, and copper, matching the composition of emitted particles captured from the plasma. The particles closest to the melt pool have a core–shell structure composed of iron-manganese-chrome cores within copper shells. Particles collected farther from the melt pool do not exhibit a core–shell morphology but do have a similar elemental composition. The correlation between plasma optical emissions and particle composition can be used to predict the composition of the melt pool, allowing for real-time control over welding and sintering. Having feedback at low latencies is essential to creating autonomous systems that make decisions based on the output of highly trained artificial intelligence algorithms.

Contact

Matt Olszta, Pacific Northwest National Laboratory, Matthew.Olszta@pnnl.gov

Funding

This research was supported by the AT SCALE Initiative under the Laboratory Directed Research and Development program at Pacific Northwest National Laboratory, a multiprogram national laboratory operated for the Department of Energy by Battelle Memorial Institute under contract no. DE-AC05-76RL01830.

This work benefited from a synergistic collaboration with a DOE/NNSA/DNN R&D (NA-22) funded project developing a hot metal vapor condensation model.