During plasma melting, pre-alloyed rods are gradually fed beneath a plasma arc. The molten metal flows into a cooled mould, where it solidifies into an ingot for further processing. Stable production requires precise monitoring of the melt level and corresponding adjustments to the movement of the material. However, the intense light emitted by the plasma overwhelms conventional camera systems, so the process has until now required the constant attention of an experienced operator.
Advanced Metal Powders and the Technical University of Liberec have developed a system combining a cooled HD camera, optical filters and artificial intelligence. The filters suppress the light emitted by the plasma column, while AI determines the melt level from the camera image in real time. The data is then sent to a control system, which automatically adjusts the speed at which the material is withdrawn from the mould as the level changes.
The technology was tested during ten melts of round rods with precisely known cross-sections and a further four melts of rods produced using HPAM technology. In every case, the system responded consistently and predictably. Before it can be used in fully unattended operation, testing of critical operating scenarios must be completed and backup safety mechanisms added.
The result opens the way to automating demanding metallurgical and foundry processes that still depend on the experience and constant attention of operators. It could be particularly useful in processing specialised metal materials, where consistent and repeatable production is essential.
The technology was developed as part of the Sensing for the 21st Century project, within the National Centre of Competence for Materials, Advanced Technologies, Coatings and Their Applications. The centre is supported by the Technology Agency of the Czech Republic.
The NCK for MATCA is supported by the