Optimization of time-variant laser power in a cladding process

This work presents an optimization algorithm for time-variant additive manufacturing process parameters. The algorithm uses a finite-element model of the AM process to determine a sequence of laser powers (i.e., laser power as a function of time) that will result in the least variation in molten pool temperature. The algorithm was applied to a single-layer DED process made up of 36 unidirectional passes. The build material was Inconel® 625. A 15.1 °K (23.5%) reduction in standard deviation of molten pool temperature was achieved, and the upward trend of molten pool temperatures toward the later stages of the process was considerably reduced. Cooling rate was reduced by up to 67 °K/s (10.8%) at certain locations, although these improvements appeared largely at the later stages of the process. Another model was developed to facilitate comparison between the optimization algorithm and in situ feedback control. In this model, laser power was allowed to change in 5% increments whenever molten pool temperature dropped below 105% or rose above 120% of Inconel® 625’s melting temperature. This simulated control model yielded a 26.05 °K (44.1%) reduction in standard deviation of molten pool temperature and was able to reduce cooling rates across the entire print geometry by up to 86 °K/s (11.4%). These results indicate that the optimization algorithm can be used to achieve improved results in DED processes without the need for in situ interference. The algorithm may yield greater benefits during longer build times. In situ feedback control, if available, is still expected to perform best.