Improvement of Machine-to-Machine Reproducibility of Stereolithography (SLA) Printers Using Gaussian Processes and Bayesian Optimization




Zilevicius, Danielius

Journal Title

Journal ISSN

Volume Title



SLA has gained popularity due to its high-resolution and high degree of design freedom, however, variability in the quality of the product remains an issue. The lack of reproducibility from printer to printer is a barrier to the qualification and certification of machines in the AM industry. Uncertainty Quantification (UQ) has gained increasing attention within the AM industry to combat reproducibility issues. The most common method for applying UQ in AM is physical experiments, however, the number of experiments required to relate process parameters to process outputs makes the current method expensive and time consuming. This study focuses on the transfer of optimal printing parameters to different printers to optimize dimensional accuracy of printed parts and reduce the number of experiments required to certify an AM machine. The printing parameters used in the study were X-, Y-, Z-orientation, layer thickness, support density, support touchpoint size, cure temperature, and cure time. A Gaussian Process was used to fit the data and Bayesian Optimization was used determine the sets of printing parameters to be used during the DOE. Once the optimal parameters were found for printer 1, the same parameters were used on printer 2 and the Gaussian Process was used to determine additional printing parameters. While printer 2 performed better than printer 1, the algorithm showed improved dimensional accuracy from the initial prints to the optimal prints on both printers. The approach presented in this study is effective for reducing the number of experiments required to certify SLA machines and can be adopted for other AM processes.


This item is available only to currently enrolled UTSA students, faculty or staff. To download, navigate to Log In in the top right-hand corner of this screen, then select Log in with my UTSA ID.


Bayesian Optimization, Gaussian Process, reproducibility, stereolithography



Mechanical Engineering