SATPC0033006 Tab 04 4 SOW.pdf

Draft SOW Application of a DARWIN Framework for Fracture Risk Assessment in

Additively Manufactured Parts

James Sobotka, Michael Enright, and Craig McClung Southwest Research Institute

San Antonio, Texas

August 8, 2023

Additive manufacturing (AM) is a novel fabrication process with the potential to create unique parts relatively quickly and inexpensively. Propelled by this promise, AM applications are increasing rapidly for space applications, including safety-critical structural parts. However, the emergence of AM for these critical parts is also bringing significant certification challenges. The AM process can sometimes create anomalies such as voids that may lead to fatigue crack formation, growth, and fracture.

A zone-based probabilistic damage tolerance (PDT) methodology has been proposed as an appropriate framework for the assessment and certification of AM parts [1]. This approach calculates the probability of fracture due to the formation and growth of fatigue cracks at material or manufacturing anomalies.

The component is subdivided into different zones each having different properties, such as material properties, non-destructive inspection (NDI) probability of detection (POD), and anomaly distributions.

Zone-based PDT methods have been implemented in the DARWIN® software developed by Southwest Research Institute (SwRI®) [2]. Originally developed over twenty years ago to address rogue material anomalies in titanium rotors for aircraft engines, DARWIN has now expanded to address a wide range of integrity threats, including fracture of AM components.

NASA Marshall Space Flight Center (MSFC) has funded four previous projects at SwRI to investigate and improve the DARWIN framework for application to AM parts for space propulsion. The first project (Contract 80NSSC19P1439) focused on the anomaly distribution, which is one of the most significant input variables to the PDT analysis. SwRI documented the workflow for different methods of characterizing the anomaly distributions. They also performed a series of simple case studies to investigate and demonstrate the differences between probabilistic and traditional deterministic damage tolerance analysis based on these different anomaly distributions.

In a second project (Contract 80NSSC20P0737), SwRI investigated MSFC-supplied finite element (FE) models of actual AM components to determine the suitability of current DARWIN capabilities to analyze these models, to identify opportunities to improve DARWIN robustness, and to implement improvements where possible. One enhancement significantly reduced the amount of computer memory required for large FE models. Improvements in the computational efficiency of the DARWIN capability to calculate the critical initial crack size (CICS) at every desired location within the component were also investigated and implemented. The CICS is the size that would just grow to failure during a desired service lifetime.

In a third project (Contract 80NSSC21P1796), SwRI developed a better understanding of the AM anomaly distributions that may be encountered in production settings. This included collecting multiple examples of actual AM anomaly data and developing improved modeling approaches for these data in order to use them effectively in DARWIN. The resulting anomaly distributions were used in case studies to identify the most significant issues for further work. These case studies included additional investigations with large

FE models to evaluate recent DARWIN enhancements and identified the most significant needs for additional process and software improvements.

In a fourth project (Contract 80NSSC22PA860), SwRI implemented enhancements to DARWIN to improve computation speed associated with fracture risk assessment of AM materials and GUI visualization of large finite element models that are typically required for these assessments. SwRI developed a recommended workflow for PDT analysis of AM materials that provided an overview of deterministic and probabilistic certification approaches for AM materials, a description of the essential elements of a recommended workflow for PDT assessments, and a step-by-step illustration of the how the PDT workflow can be applied using existing capabilities in DARWIN.

Based on this completed work, two critical needs have emerged for additional research and development activities that will further develop the DARWIN probabilistic damage tolerance framework for fracture risk assessment of additively manufactured parts in a NASA context.

The first need is for SwRI to help NASA to develop a better understanding of the PDT workflow for AM materials. In Task 1 of this proposed project, SwRI will develop an application example problem to demonstrate the PDT workflow developed on a previous (fourth) project to NASA personnel. We anticipate that a relatively smaller FE model will be used for the demonstration to help avoid the computational challenges that are currently associated with use of very large FE models for AM assessments in DARWIN. A representative anomaly distribution will be applied to the analysis based on the AM anomaly distributions that were developed in a previous (third) project. The influence of NDE inspection on fracture risk will be quantified using representative POD curves that are available in DARWIN. Deterministic damage tolerance (DDT) analysis will also be performed for comparison to the PDT results. SwRI will demonstrate to NASA the step-by-step process of preparing input data and executing DARWIN to complete both the DDT analysis and the PDT analysis according to the PDT workflow. Lessons learned and opportunities for DARWIN enhancements will be documented.

The second need is to understand better the influence of PDT input parameters on the risk of fracture. In Task 2 of this proposed project, SwRI will perform sensitivity analysis to demonstrate the influence of PDT input parameters on the fracture risk of AM materials. The DARWIN model that is developed for Task 1 of this project will serve as the baseline model for a series of DARWIN runs that will be executed on Task 2 of this project. Candidate input parameters may include (but are not limited to) anomaly-related variables (i.e., anomaly size, anomaly frequency), NDE inspection-related variables (i.e., NDE method and associated POD, inspection location), percentage of the part that experiences crack nucleation lives that are shorter than the design service life, and percentage of the part is above the threshold for fatigue crack growth (material properties, applied stresses, and temperature may all play a role). SwRI will collaborate with NASA to identify the list of parameters for the sensitivity studies over the course of the project.

Deliverables. The primary deliverable from this project will be a final report summarizing the investigations and results of the two tasks demonstrating the application of the DARWIN framework to support fracture control of AM parts. A draft work plan will be presented at the kickoff meeting. Short milestone progress reports will be provided after the fifth and tenth months of the project. With NASA permission, selected results may be presented at a suitable AM workshop or conference.

Quote. This proposed project is estimated to cost $60,000 and require thirteen months to complete. This estimate is for budgetary purposes only and does not constitute a formal bid or a quotation for competitive evaluation. This estimate does include proprietary pricing information, however, and we request that it not be released to any third party. Upon request, a formal offering can be prepared.

[1] M. Gorelik, “Additive Manufacturing in the Context of Structural Integrity,” International Journal of

Fatigue, Vol. 94, 2017, pp. 168–177.

[2] R. C. McClung, M. P. Enright, J. P. Moody, Y.-D. Lee, J. C. Sobotka, V. Bhamidipati, and J. W. McClure, “A Comprehensive Framework for Probabilistic Damage Tolerant Design of Aerospace Components,” Proc. 29th Symp. Int. Comm. Aeronautical Fatigue and Structural Integrity (ICAF), Nagoya, Japan, 2017.