OUSD (R&E) MODERNIZATION PRIORITY: Artificial Intelligence/ Machine Learning
TECHNOLOGY AREA(S): Materials; Information Systems
OBJECTIVE: To significantly accelerate the laser metal powder bed additive manufacturing (LPBAM) process and improve the microstructural quality of LPBAM parts by developing a scalable, intelligent multi-laser based system. In this AM system, some of the lasers query the powder bed quality prior to melting it, while other lasers use that information to intelligently process the powder bed, yet other lasers monitor and control the solidification rates for microstructural control and part quality.
DESCRIPTION: In spite of the significant advantages of the AM process to enable a broader design envelope, there are still significant technological challenges that limit the full utilization of AM. Some of these challenges include, but are not limited to: the slow production rates requiring several hours or days to build parts; post-processing treatments and machining to optimize the build; defects and residual stresses to meet dimensional tolerances. Depending on the materials, AM lacks process stability that can produce consistent, defect-free parts on a first fabrication.
At the root of all these challenges are three physical limitations: the small size of the laser processing volume (typically a 50-micron diameter by 50-micron tall cylinder), the wide powder size distribution, and finally the fast heating and cooling rates. The first two limitations contribute to a large point-to-point static variability of the powder bed physical properties (process mass, absorptivity, heat capacity, thermal conductivity, and density), while the last limitation adds dynamic variability to the powder bed from spatter, splatter and powder denudation effects. Another limitation of current AM systems is that they typically process the powder bed with a single laser leading to large thermal gradients across parts and the associated large residual stresses.
Intelligent multi-laser control and processing can enable unique processing regimes that could alleviate or even eliminate those challenges. For example, in one of many possible conceptions to this topic, a system could be scalable by modules. Each module could include three lasers, the process monitoring sensors and one intelligent controller. One of the lasers, moving at constant speed, would preheat the powder with constant power (without melting it) to produce a unique thermal pattern that would be captured by an intelligent controller. The second laser in the module, moving closely behind the first and operated by the same intelligent controller, could melt-process the powder to a precise final temperature based on the previous thermal history at and around that location. Finally, the third laser, moving closely behind the second, could control the cooling rate while the intelligent controller captures the thermal profile that could be used for processing the next tracks or layers.
In a different intelligent multi-laser conception, each laser could remain stationary while processing a single point on the powder bed. The thermal processing of each point could have three phases. First, the lasers could query the powder by preheating it with constant power for a precise amount of time (without melting) while an intelligent controller captures the unique thermal profile at and around each point. Next, based on the information gained, the same intelligent controller could melt-process the powder to a precise final temperature and let it cool following a precise cooling rate all the while capturing the thermal profiles for later use. Finally, after the processing phase, all the lasers could intelligently shift to new positions and the process would repeat.
The above are just two possible approaches out of many others that address the requirements of this topic. Any innovative approach that is scalable and that integrates multiple lasers (or multiple laser beams) with process monitoring sensors via an intelligent controller to probe the powder bed prior to processing it with the goal of making AM parts faster and with better microstructure will be considered for funding. The proposal author should clearly indicate which research advances the university partner will contribute to the project and how the small business plans to incorporate them into the AM system.
PHASE I: During the Phase I effort, the contractor will conduct a feasibility study of their proposed scalable, intelligent multi-laser AM system for metal powder bed fusion that includes validation tests that are consistent with a Phase I STTR budget by addressing only the most critical items of their approach and/or making simple coupons. By the end of the Phase I effort the PI will develop a detailed design of the AM system that includes at a minimum: 1- The approach for querying the powder bed and the anticipated information and resolution (spatial and temporal) to be gained from the approach; 2- The intelligent processing approach to be used (machine learning, deep neural networks, neuromorphic processing or others) including the input query parameters and the output laser system control parameters and time constants; 3- The intelligent processor training approach which could include: material process modeling and simulation tools; training coupons and auxiliary data, such as mechanical tests result, x-ray tomography and micrographic analysis for training and ground truth data generation. The contractor will continue building the scalable, intelligent multi-laser during the Option phase by acquiring system components and refining the design of the system based on the validation test results.
PHASE II: Complete the scalable, intelligent multi-laser AM system by developing the control software and GUI. Perform full system validation tests after completing all the training exercises of the intelligent component of the system. For the full system validation, identify a challenge part between the performer and the Navy/DoD team. Fabricate two challenge parts, one for destructive microstructural analysis and another for mechanical testing. The success criteria consists in making the challenge parts in a shorter amount of time (consequence of having a multi-laser system); with less defects and distortions (consequence of probing the powder bed prior to processing it); and with better control of the microstructure and mechanical performance (consequence of the intelligent process control) than the same parts made by other state of the art AM platform but without multi-lasers, powder bed probing and Intelligent control.
PHASE III DUAL USE APPLICATIONS: Military and Commercial sectors that could benefit from this AM system include: aerospace, shipping, space, transportation, rail, automobile and medical. Applications include almost all technology areas such as: engine parts, structural parts, mechanical or electrical parts, medical prosthetics, and tooth implants. The contractor shall support the Navy/DoD to help transitioning the system to a DoD facility in support of various programs. The contractor should also identify potential commercialization opportunities
- Alhart, Todd. “3D Printing: Just Press Print: GE Is Building A 3D-Printing Vending Machine For The Jetsons.” GE Reports, Apr 10, 2017. https://www.ge.com/reports/just-press-print-ge-building-3d-printing-vending-machine-jetsons/.
- Zhua, Z., Anwer, N., Huang, Q., and Mathieu, L. “Machine learning in tolerancing for additive manufacturing.” CIRP Annals, Vol 67, Issue 1, 2018, pp. 157-160. https://www.sciencedirect.com/science/article/pii/S0007850618301434#!.
- Scime, L. and Beuth, J. “Anomaly detection and classification in a laser powder bed additive manufacturing process using a trained computer vision algorithm.” Additive Manufacturing, Vol 19, Jan 2018, pp. 114-126. https://www.sciencedirect.com/science/article/pii/S221486041730180X.
- Scime, L. and Beuth, J. “Using machine learning to identify in-situ melt pool signatures indicative of flaw formation in a laser powder bed fusion additive manufacturing process.” Additive Manufacturing, Vol 25, Jan 2019, pp. 151-165. https://www.sciencedirect.com/science/article/pii/S2214860418306869.
- J Petrich, Z Snow, D Corbin, EW Reutzel , “Multi-modal sensor fusion with machine learning for data-driven process monitoring for additive manufacturing” - Additive Manufacturing, 2021.
- “Applying neural-network-based machine learning to additive manufacturing: current applications, challenges, and future perspectives.” X Qi, G Chen, Y Li, X Cheng, C Li - Engineering, 2019 https://www.sciencedirect.com/science/article/pii/S2214860421001305.
- Zielinski, Peter (ed). How Machine Learning Is Moving AM Beyond Trial and Error (Originally titled 'Where AM Meets AI). Additive Manufacturing. https://www.additivemanufacturing.media/articles/how-machine-learning-is-moving-am-beyond-trial-and-error.
- Everton, S.K., Hirsch, M. Stravroulakis, P, Leach, R.K., and Clare, A.T. “Review of in-situ process monitoring and in-situ metrology for metal additive manufacturing.” Materials and Design, 95 (2016), pp. 431–445. http://www.sciencedirect.com/science/article/pii/S0264127516300995.
- Spears, T.G. and Gold, S.A. “In-process sensing in selective laser melting (SLM) additive manufacturing.” Integrating Materials and Manufacturing Innovation 2016 (a Springer Open Journal). DOI 10.1186/s40192-016-0045-4. https://link.springer.com/article/10.1186/s40192-016-0045-4.
KEYWORDS: Metal Additive Manufacturing; Artificial Intelligence; AI; Machine Learning; Neural Networks; Laser Based Powder Bed Fusion; Process Monitoring Sensors; Lasers