Abstract Scope |
Laser cladding is a modern process used to deposit high performance wear and corrosion resistant coatings. The bonding between overlay and substrate is produced by melting a thin layer of the base material with a highly concentrated laser heat source and added material typically in powder form. Low heat input, minimal mixing of dissimilar materials, high process speeds, and automated control systems are major advantage of this state-of-the-art process. With a push in industry for maximum productivity and the increasing trends towards additive manufacturing (AM) using direct energy deposition (DED) processes (laser cladding based AM), there is a need to control the process adaptively. Thermal monitoring of the molten clad pool is a promising avenue of process control for these applications.
Various pyrometric and camera systems are used to monitor the temperature distribution and the process parameters in laser cladding. The areas of interests include the melt pool geometry, the deposited clad bead geometry, and the powder flow. The maximum width of the melt pool defines the width of the resulting clad track. Systems based on cameras working in visible spectral range (VIS) and mid-wave infrared (MWIR) are two major groups that are used for monitoring the laser cladding process. The MWIR cameras are more sensitive to the thermal emission of the melt pool and the clad bead; however, they are also prone to be affected by any reflected thermal emissions. Also, they require special non-glass optics and the presence of no glass between the camera and the clad area. Visible light and near-infrared (NIR) cameras are less sensitive to the target thermal emission of the melt pool and the clad bead but are simultaneously less sensitive to the any unwanted IR emissions. Short-wave infrared (SWIR) camera systems are reported to be able to combine the benefits of the VIS and MWIR. They are sensitive to thermal emission of metals starting from around 250 C, i.e., they are less prone to interference created by the reflected IR light while being more sensitive to thermal emission than the visible light and near-infrared (NIR) systems. The SWIR cameras can also use less expensive, standard glass optics. The focus of this work is on the experimental comparison of the performance of the camera systems working in three spectral bands: visible, near-infrared and short-wave infrared applied to laser cladding.
The experimental set up included the ytterbium fiber laser with the working wavelength of 1070 nm. The cameras were set on a tripod off the laser beam axis. Laser cladding processes typical for industrial applications were used for the experiments including cladding of Inconel 625 powder onto the substrates of Inconel 718 and 4140 and 4330V alloy steels and spherical tungsten carbide in Ni-B-Si matrix onto 4330V. Imaging was performed both perpendicular and in-line to the clad bead. Flat and cylindrical substrate samples were used for the experiments.
In VIS, only the thermal emission of the Inconel 625 powder was distinguishable from the ambient. The melt pool and clad bead thermal emission was negligible in the VIS spectral range. In the NIR spectral range, the powder flow emission was slightly higher than in VIS; however, analogously to VIS, the melt pool and the clad bead thermal emissions were not distinguishable from the ambient emission.
In the SWIR band, the thermal emission of the Inconel 625 melt pool, the cooling clad bead and the powder was clearly visible. The illumination of the powder particle was more intense than in VIS and NIR. The powder reflected from the clad surface was also visible. The cooling clad bead SWIR illumination was distinguishable from the substrate for the length of around 2-3 melt pool lengths. The melt pool was clearly visible through the powder stream and overspray. The liquid-solid interface was also clearly observable, as the emissivity of the molten clad material at the same temperature around melting point appears lower than of the solidified material. Thermal emission of the tungsten carbide (WC) clad bead on 4330V was clearly visible as well as the Inconel 625. Liquid-solid interface for the solidifying Ni-B-Si matrix, however, was less detectable due to high percentage (60%) of non-melting spherical WC particles in the melt pool.
The effect of the variations in the laser power, travel speed, and powder feed rate on the Inconel 625 melt pool was investigated by imaging the process with the SWIR thermal camera. At high power settings, the melt pool was elongated forming a long slightly oscillating tail. The extended time that the surface of the molten material spent in the atmosphere resulted in a discoloration at the bead center line. Excessive powder feed rate led to an excessive elongation of the melt pool as well.
SWIR band thermal camera performance for the laser cladding application was found significantly better for monitoring the laser cladding melt pool area than the VIS and NIR band camera. The thermal emission of the melt pool, the clad bead and the powder flow was clearly detectable in SWIR, whereas only powder flow was illuminated enough in the VIS and NIR spectral ranges.
Keywords: SWIR, short-wave infrared, laser cladding, melt pool control, clad bead size control |