Objective of the proposal is to develop alloy compositions that enable additive manufacturing (AM) processes to produce properties that are not currently achievable, such as materials with a preferred crystallographic orientation, dispersion-forming alloys, which can either form the dispersion during AM or after AM through heat treatment. The alloy compositions shall reduce defects in components, thus promoting them to be more resistant to fatigue, with potential increases in strength. To address the Objective, we propose an Integrated Computational Materials Engineering (ICME) toolset capable of predicting the fatigue life of AM metallic components, along with an AM parameter optimization approach, utilizing multi-dimensional functional reconstruction and high-throughput testing. Existing toolsets have limited ability to predict how AM affects material properties of additively manufactured parts. To account for the complex underlying processes, we propose a machine learning (ML) framework for predicting the fatigue properties. The ICME framework is in part a generalization of Statistical Fatigue Life model by one of the authors. The innovation proposed relates AM processability with alloy chemistry, for purpose of predicting alloy chemistry that minimizes defects while maintaining base alloy properties. We achieve this using a ML framework linking materials data sets, modeling, and powder AM process variables. In Phase I, we will explore the literature to determine relationship of wrought alloys chemical compositions and the chemistries of its cast alloy corollaries, understand the underlining reasons for the different chemistries to enable an alloy to be similar by each process. We will focus first on IN 718, and then on Haynes 230 or IN 738, with the intent of producing properties equivalent to or greater than achieved by the wrought alloy. The proposing company will work with University of Wisconsin in Madison (UWM) on AM and process optimization, and with University of Tennessee in Knoxville (UTK) on characterization of AM metallic components. The team includes a renowned expert from UWM in powder AM, Prof. Dan Thoma, also with significant experience in super-alloys, plus accomplished another materials science expert from UTK on mechanical characterization and fatigue studies, Prof. Peter Liaw. The scheme proposed for controlling the AM process parameters will inherently find the sweet spots of the superalloys of interest and may not require redesign the alloy to avoid cracking or other defects. The high-throughput technique coupled with ML offers a nice way of quickly identifying the processing parameters for superalloys still not fully developed, such as the IN 738, Haynes 230 or Haynes 282. To our knowledge, nobody has implemented the high-throughput testing approach for superalloys, pioneered by UWM. For support towards product validation and commercialization, refer to the enclosed Letters of Support from Lockheed Martin, Boeing and Siemens.
Benefit: The AM process and alloy chemistries, that are suited specifically for AM processes, offer the opportunity of conformal, and unique design not possible with more conventional fabrication processes. Proven AM process optimization leading to a minimization of process - and materials - derived defects will improve acceptance of AM for producing component for the Navy and for the private industry. The use of AM can lead to more innovative designs capable of more efficiently removing heat, and meeting ever-increasing demands on components for the Navy, as such designs can eliminate or severely reduce joints. AM processing of components, that have already been qualified for Navy use, can also be applied to commercial use more quickly. We expect potential commercial applications to include, and be facilitated through, the Naval Warfare Research Center (NWRC) in Keyport WA, Lockheed Martin, Boeing and Siemens. The Principal Investigator (PI) gave a back on presentation to the NWRC in Keyport back on January 19 2021. We received great feedback on the enhanced versions of the statistical fatigue life model, originally applied to Titanium (Ti) alloys, esp. from their AM lead, Mark T. Sorna. The feedback involved benchmarking of the enhanced versions of the statistical fatigue life model against the original version, extensions from Ti alloys to superalloys, and relative contributions of the primary sources contributing to degradation in fatigue life. We expect communications with NWRC in Keyport to be further facilitated through liaisons (advocates ) at Lockheed Martin. With this proposal, we present a Letter of Support cosigned by representatives from Lockheed Martin Missiles and Fire Control as well as Lockheed Martin Aeronautics. The proposing company already has a relationship with Lockheed through an Air Force Phase I project with Oregon State University titled Use of Artificial Intelligence (Joint Optimization) to Accelerate Development of New Energetic Materials . On the commercial side, the offeror has a relationship in place with Siemens, through Dr. Anand Kulkarni. We have patents pending, had additional patents already granted and completed a textbook chapter with Dr. Kulkarni. In addition, we have submitted disclosures for upcoming intellectual property to Siemens. Our intent is, in summary, to develop the core AM process optimization technology in collaboration with UWM and UTK. In the event of a Phase II award, we intend to work with NUWC in Keyport and Lockheed Martin on application of the AM optimization towards qualification of parts for Navy use and get our first success story . We would present our first success story to Boeing and work with Boeing towards a second success story . We then intend to present our first two success stories to Siemens and work with Siemens towards a third success story . We expect much of subsequent commercialization activity to be facilitated through our advocates at Lockheed Martin, Boeing and Siemens.
Keywords: Data Analytics, Data Analytics, wrought alloy, statistical fatigue life model, additive manufacturing, integrated computational material engineering, Machine Learning, Fatigue Life Prediction, Superalloy
