Laser-powder bed fusion process optimisation of AlSi10Mg using extra trees regression

Aluminium alloys have a wide range of applications, especially in the fields of aviation and aerospace. Near net-shape parts with a great structural complexity can be produced using additive manufacturing techniques. However, understanding the Laser-Powder Bed Fusion (L-PBF) process itself and the correlations between different process parameters and the mechanical properties is often very challenging. In order to achieve insights into the whole printing process, statistical tools such as Designs Of Experiment (DoE) are usually applied. In this study, we examined the additive manufacturing of AlSi10Mg, a well-studied aluminium alloy used for L-PBF, and the modelling of its printing process by applying machine learning. The influences of different printing parameters (i.e. laser power, laser spot size, hatching distance, layer height and scanning speed) on the mechanical properties (i.e. density, tensile strength and hardness) were examined by generating different machine learning models based on data obtained with DoE and additional experiments. The best performing models were evaluated regarding the printing process and the respective testing procedures used to measure the mechanical properties. Mean coefficients of determination ranging from 56.44% to 98.54% were achieved. Finally, a processing window for producing dense samples with high tensile strengths and high hardness values was found.