OUSD (R&E) CRITICAL TECHNOLOGY AREA(S): Sustainment; Trusted AI and Autonomy OBJECTIVE: Design and develop a data access tool that can determine if a part could be and should be produced via additive manufacturing (AM). These disciplines can include, but are not limited to the following: engineering design, manufacturability, producibility, testing, and machine learning to develop expert-guided algorithms to identify which readiness degraders, sustainment issues, and next generation components can be produced via AM. DESCRIPTION: AM has the potential to increase readiness and improve maintenance and sustainment operations by reducing long lead times and eliminating obsolescence related issues. Furthermore, the technology enables improvements to current systems (e.g., light-weighting, part count reduction, increased system performance) through designs that are not possible by conventional manufacturing techniques. However, for the technology to continue to transition from indirect uses to efficiently producing qualified end use parts several technology barriers need to be overcome. One of the primary needs is the development and integration of data access tools with analytical capability to optimize the selection of viable families of AM candidate parts without requiring the burden of manual item-by-item review. The solution also should include analytical capabilities to effectively manage product technical and logistics information and provide users with substantive assessments on an item's suitability to AM production. Knowledge of computer aided design (CAD), technical data packages (TDPs), and product lifecycle management (PLM) tools is required, as well as the ability to quantify the limitations of existing AM systems and processes. Innovative design concepts are being sought for the development of an AM candidate assessment tool with the ability to: (1) coarsely filter and screen for irrelevant parts, (2) identify candidate parts using criteria such as material, performance requirements and parts family types, (3) predict production estimates and delivery schedules by building/expanding upon a cost and time estimation tool, and (4) automatically search Navy databases for parts most suitable for AM and subsequently validate them using a machine learning model or algorithm. PHASE I: Develop, design, and demonstrate feasibility of a concept for an AM candidate assessment tool utilizing representative data. Develop a coarse filter or screening mechanism for candidate parts. The filter will use binary (yes/no) expert judgments, combined with active machine learning (ML) (e.g., adding expert judgements iteratively to understand the value of additional information), to filter parts unsuitable for AM. The tool will screen by critical dimensions (i.e., work envelope or bounding box) and known limitations of existing additive manufacturing systems of interest. Design should consider other criteria such as material, performance requirements, and parts family when determining the suitability of a part for AM. Refine existing cost and time estimation tools to predict production cost estimates and delivery schedules for representative AM part candidates. Production cost estimates should consider all post-processing operations (e.g., heat treatment, surface treatment, final machining, and inspection) required to meet the part's acceptance criteria. The Phase I effort will include prototype plans to be developed under Phase II. PHASE II: Extend the decision model(s) developed under Phase I to address Navy part characteristics and mission priorities to develop a mutually agreed upon prioritization schema. Produce a ML algorithm, seeded with the aforementioned models, to integrate and search Navy databases for parts most suitable for AM, and the value of potentially (costly) additional information. Demonstrate and validate the prototype by utilizing actual Navy data. PHASE III DUAL USE APPLICATIONS: Transition the tool under the guidance of PEO-CS Digital Thread team and/or NAWCAD LKE's Digital Enterprise Tools Branch. Commercialize the tool resulting from the Phase I/II R/R&D activities. This would likely involve further integration with existing, commercially-available CAD and PLM platforms. Military and Commercial sectors that could benefit from this AM part identification tool 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 dental implants. Support the Navy/DoD to help transitioning the system to a DoD SYSCOM in support of various programs. REFERENCES: 1. Parks, T. K., Kaplan, B. J., Pokorny, L. R., Simpson, T. W., & Williams, C. B. (2016). Additive manufacturing: Which DLA-managed legacy parts are potential AM candidates? LMI. https://apps.dtic.mil/sti/pdfs/AD1014552.pdf 2. Page, T. D., Yang, S., & Zhao, Y. F. (2019, July). Automated candidate detection for additive manufacturing: a framework proposal. In Proceedings of the design society: international conference on engineering design (Vol. 1, No. 1, pp. 679-688). Cambridge University Press. https://www.cambridge.org/core/services/aop-cambridge-core/content/view/08AD686E70255907AA0DBC9D6F9B6E09/S2220434219000726a.pdf/automated-candidate-detection-for-additive-manufacturing-a-framework-proposal.pdf 3. Yang, S., Page, T., Zhang, Y., & Zhao, Y. F. (2020). Towards an automated decision support system for the identification of additive manufacturing part candidates. Journal of Intelligent Manufacturing, 31(8), 1917-1933. https://link.springer.com/article/10.1007/s10845-020-01545-6 4. Lindemann, C., Reiher, T., Jahnke, U., & Koch, R. (2015). Towards a sustainable and economic selection of part candidates for additive manufacturing. Rapid prototyping journal. https://www.emerald.com/insight/content/doi/10.1108/RPJ-12-2014-0179/full/html KEYWORDS: Additive Manufacturing; AM; Artificial Intelligence; AI; Machine Learning; ML; ; Neural Networks; Laser-Based Powder Bed Fusion; Candidate Identification; Decision Making
