N252-107 TITLE: A Hybrid Digital Twin Framework for Modeling Machine Process Parameters for Automated Fiber Placement-based Manufacturing of Composite Structures

OUSD (R&E) CRITICAL TECHNOLOGY AREA(S): Advanced Materials;Sustainment

OBJECTIVE: Develop a physics-based and artificial intelligence (AI) driven digital twin framework to tailor automated fiber placement (AFP) process to manufacture complex composite structures with improved manufacture rates and quality.

DESCRIPTION: Low-cost composite airframes are strongly demanded by DoD and commercial industries to reduce total life cycle cost through the use of integration of automation and digital manufacturing of complex composite structures. AFP has been used extensively by major aerospace industries such as Boeing and Spirit AeroSystems for the manufacturing of large-scale composite fuselage structures. However, the poor process planning and time-consuming operator-dependent manual inspection can significantly interrupt the manufacturing process and increase the manual rework rate due to the intolerable level of fabrication-induced defects. These interruptions and need for artisan intervention to maintain part quality diminish the benefits of AFP-based composite manufacturing. An increasing amount of research and technology development has been devoted to implementing a data-driven planning, in-process inspection and defects identification, and a machine learning (ML)-based defects ranking and decision-making framework on the corrective actions to improve production rates and quality. However, key technology gaps still exist including the lack of physics models for the process simulation that is driven by the interaction of moving heat source, roller compaction, and strip tension during fiber placement. Thus, the process still requires operator intervention for corrective actions.

The Navy requires an integrated Digital Twin (DT)-based system that can construct a smart AFP (S-AFP) system by using combined sensor and synthetic data to reduce or eliminate false actions. The primary goal is to demonstrate the use of the developed hybrid DT-based framework to enhance the in-process defects monitoring, process control, and rational metric for rework. The secondary goal is to use the system to improve part performance and quality. The delivery of the project will be a software tool, which is validated by producing a representative aircraft component.

PHASE I: Develop a physics-based and AI-driven DT modeling system. Establish the feasibility of the concept. Prepare a Phase II plan.

PHASE II: Build a prototype system and demonstrate it on a composite part with at least one curvature. To demonstrate the fidelity of the algorithm, use it to predict temperature of the composite part during the AFP process. After demonstrating the accuracy of the temperature prediction using collected sensor data, develop an approach towards future implementation of a closed loop to control the heating element for achieving a desired temperature profile required for the high-quality manufacturing. The Phase II effort will include the demonstration of manufacturability and mechanical performance of the part compared with traditional AFP including metrics on rework time and machine layup time as a percentage of total working time. Apply the same modeling approach to the other process parameters including compaction, speed, and tow tension to generate a combined DT model, including a synthetic environment for simulating and analyzing possible part layups and with the ability to communicate in a closed loop fashion with the smart AFP machine to adjust process parameters to improve layup quality. Analyze the effects these parameters have on process-induced defects through successive manufacturing trials. Perform the final demonstration of the effectiveness of the DT on defect reduction by manufacturing a complex panel with at least two forms of curvature. Analyze and report the same metrics used in the initial phase. Deliver a complete software that is machine agnostic and capable of being applied to a variety of AFP machines.

PHASE III DUAL USE APPLICATIONS: Support the Navy in transitioning the system to Navy use.

The expected transition targets are the original equipment manufacturers (OEMs) for aircraft.

REFERENCES:

1. Brasington, A., Francis, B., Godbold, M., & Harik, R. (2023). A review and framework for modeling methodologies to advance automated fiber placement. Composites Part C: Open Access, 100347. https://doi.org/10.1016/j.jcomc.2023.100347
2. Sun, S., Han, Z., Fu, H., Jin, H., Dhupia, J. S., & Wang, Y. (2020). Defect characteristics and online detection techniques during manufacturing of FRPs using automated fiber placement: A review. Polymers, 12(6), 1337. https://www.mdpi.com/2073-4360/12/6/1337/pdf
3. M. Kirkpatrick, A. Brasington, AD Anderson, R. Harik (2020) Creation of a digital twin for automated fiber placement. https://nextusc.com/wp-content/uploads/2020/11/2020Kirkpatrick.pdf
4. BJ Fancis, MB Kirkpatrick, DB Sander, AR Brasington, R Harik, (2023) Towards a Virtual Manufacturing Architecture to Capture the Automated Fiber Placement Lifecycle https://www.researchgate.net/profile/Alex_Brasington/publication/370895192_Towards_a_Virtual_Manufacturing_Architecture_to_Capture_the_Automated_Fiber_Placement_Lifecycle/links/64679dfe66b4cb4f73c19982/Towards-a-Virtual-Manufacturing-Architecture-to-Capture-the-Automated-Fiber-Placement-Lifecycle.pdf

KEYWORDS: Composites; fabrication; AFP; Automated Fiber Placement; lightweight composites; virtual manufacturing

TPOC 1: Neil Graf
[email protected]

TPOC 2: Anisur Rahman
[email protected]

---