Optimization of L-PBF Process Parameters: A Multi-Objective Framework for Variance Reduction and Mechanical Strength Enhancement Using Linear Programming and Multi-Objective Methods

A new approach to the multi-criteria optimization of L-PBF process parameters in the face of high variability is proposed, based on the application of the reward function (RF), which is essentially the inverse of the Taguchi quality loss function (QLF). The RF is defined as the logarithm of the ratio of the design parameter value to the minimum target parameter value. For values below 1, the RF becomes sharply negative. Logarithmic additivity enables the optimization of the target function to be defined as a sum of reward functions for the optimization criteria, each with a corresponding weight coefficient. This makes it possible to formulate the optimization problem in terms of linear programming. This approach was tested in the optimization of the L-PBF technological modes of AISI 321 powder, leading to a 6.7% increase in productivity. The mechanical properties of the samples obtained were not inferior to those of the alloyed samples using the previously recommended process parameters in terms of ultimate strength, and they exceeded the latter by 5% in terms of conditional yield strength. Process variability significantly decreased; the RMS of ultimate strength decreased nine-fold, conditional yield strength decreased ten-fold, and relative elongation decreased two-fold.