Uncertainty assessment techniques for selective laser melting process control

Making decisions and inferring control actions in manufacturing environments requires considering many sources of uncertainty. The inherent ability of fuzzy logic to incorporate imperfect information into a decision model has made it suitable for production optimization by combining both productivity and final quality requirements. Indeed, in selective laser melting processes, these aspects must take into account the complex interaction of many process parameters. The proposed fuzzy-based decision-making approach overcomes this difficulty by taking into account heterogeneous error sources associated with both process variability and modelling assumptions, in general unable to accurately describe the process physics. The experimental part of the work concerned the creation of the starting data set on which develop the fuzzy model by evaluating the ultimate tensile strength of the AlSi10Mg printed samples by varying the building orientation and the volumetric energy density. Then, the fuzzy process map allowed to define the optimal process parameters' combination able to minimize the uncertainty due to the process variability and the simplification introduced in the model, together with the requested performing properties. It suggests that for orientations lower than 40 degrees and energy lower than 73 J/mm3 it is possible to guarantee a tensile strength within the range 300-366 MPa.