Abstract

Process parameter optimization in Additive Manufacturing is a resource intensive operation when done using trial-and-error experimentation. For metal powder bed fusion processes, parameter optimization typically involves 3 or more stages, months of time, and tens of thousands of dollars. This is true not only for introduction of a new material but also when changing powder suppliers for an existing material or just seeking to improve machine productivity by changing laser scan parameters and/or layer thickness. The first stage of process parameter optimization uses single bead experiments to determine laser scanning parameters which create good deposits for a desired layer thickness. These experiments are done by spreading a layer of powder over a baseplate and then creating a series of single laser scan tracks at various laser powers and scan speeds. In most cases hundreds of scan tracks are created. Scan tracks are visually inspected and those scan tracks which result in contiguous, smooth beads are cross-sectioned and measured using microscopy. Cross-sections which reveal porosity-free deposits with meltpool depth to width ratios between 0.3 and 0.6 are considered candidate laser scan parameters. The second stage of process parameter optimization is the production of porosity cubes. For the candidate laser power and scan speed combinations identified in stage 1, scan strategies are produced by varying laser scan spacing and scan pattern (such as a stripe or chessboard pattern). In most cases dozens, and in some cases hundreds, of laser power, speed, scan spacing and scan pattern combinations are investigated by building up at least one small cube of material using each combination. In some cases 3 replicates of each combination are produced to enable statistical analysis. For each cube that completes successfully, they are cross-sectioned, polished, photographed and analyzed to determine whether the cube has a low enough porosity level to be useful. In most cases users are looking for fully dense porosity cubes produced using process parameters which result in the fastest build speed. In the third stage of process parameter optimization, the microstructure for each porosity cube parameter combination of interest is investigated. Depending upon the alloy of interest, a material scientist may desire a columnar microstructure, an equiaxed microstructure, a certain phase, and/or a certain amount of precipitates. In some cases people re-use the porosity cubes for microstructure analysis if the cubes are large enough for such a use. In other cases, new microstructure samples are created. Once the microstructures of interest are identified, additional stages may involve tensile, fatigue, corrosion, and other testing to down-select the best process parameters. The use of simulation tools can dramatically reduce the time to achieve process parameter optimization. A combination of single bead, porosity and microstructure simulations, along with confirmatory experiments, result in an order of magnitude reduction in the time and cost needed for process parameter optimization. As a result, simulation-driven process parameter optimization enables additive manufacturing practitioners to explore a much broader process parameter space in a shorter time and lower cost than trial-and-error experiments. In this talk, an overview of simulation approaches for process parameter optimization will be presented, along with case studies and examples of comparisons between simulation results and experimental results. The use of new additive simulation tools for process parameter optimization clearly show that simulation driven process parameter optimization is a better way to develop new process parameters than the traditional, experimental approach.