View on GitHub

Air Force Research Laboratory (AFRL) Additive Manufacturing (AM) Modeling Challenge Series

GETTING STARTED

CHALLENGE DATA

Please check back for updates.

For each challenge listed below, there are links to the Problem Statement, Dataset, and Answer Template. The Dataset must be accessed through Globus where individual files from the dataset can be transferred to a location of your choice. If the Dataset link below takes you to ‘Error Fetching Search Record’ page, make sure you are signed into Globus AND click the Login link in the upper right hand corner to sign in using your Globus account information.

To access the datasets, click the “Globus” link under the “Get the Data” section, which should redirect you to Globus’ web application. Data may be transferred from Globus to your computer by setting up an endpoint. Documentation on establishing a Globus endpoint for each operating system is found here: MacOs, Linux, Windows. Once an enpoint is established, the Globus web application is used to conduct the download. Information on transferring data from Globus is found here.

CHALLENGE 1: Macro-scale Process-to-Structure Predictions
Given processing details, predict macroscale residual strain.
[Problem Statement] [Dataset] [Answer Template]

CHALLENGE 2: Micro-scale Process-to-Structure Predictions
Given processing details, predict deposit geometry and microstructural details at the single bead, single layer scale.
[Problem Statement] [Dataset] [Answer Template]

CHALLENGE 3: Macro-scale Structure-to-Properties Predictions
Given representative microstructure characteristics, predict aggregate stress-strain behavior.
[Problem Statement] [Dataset] [Answer Template]

CHALLENGE 4: Micro-scale Structure-to-Properties Predictions
Given an explicit microstructure representation, predict aggregate stress-strain behavior and strain in particular grains.
[Problem Statement] [Dataset] [Answer Template]

QUESTIONS & ANSWERS

Answers that AFRL provides to any questions about the challenges will be posted here.

What is the preheat temperature of the build plates?
The programmed build plate preheat temperature was set to 80°C for all builds used in the challenge series.

What is the temperature inside the chamber (i.e., above the powder) during the build?
No temperature data inside the build chamber was collected.

Is it correct that the 10 layers in the B21.cli represent a multi-layer formation? This would be in contrast to B27.cli which would represent 30 tracks of one layer formation.
Yes, B21 is a single track wall that is 10 layers tall. B27 is a 2D pad composed of multiple tracks with in a single layer. Slides 10 and 21 from the Challenge 2 Problem Statement are good references.

What was the laser beam diameter?
We did not directly measure the laser spot size during these experiments/builds. From the manufacturer’s data sheet: Gaussian laser spot diameter (4σ) = 0.1 mm (Challenge 2 Problem Statement slide 13). Note: Challenge 2 builds were fabricated without making adjustments to the laser spot size value, therefore the value is internally consistente whithin the data sets even if it deviates from the manufacturer’s stated nominal value.

What is the atmosphere used to print the challenge samples?
Argon

Regarding the multi-layer, thin-wall cases for Challenge 2, what is the dwell time between the deposition of each layer, if any?
Please refer to the HomeIn_Build B.csv file located in the Input data folder in the Challenge 2 data set for the layer times. Refer to the Challenge 2 Problem Statement slide 9 (substrate and timing) for a description of processing to help you interpret the HomeIn file.

What is the expected temperature of the deposited powder once leaving the hopper?
This value was not measured. There was no intentional heating of the powder in the hopper.

Is detailed information (i.e. location, size, deposition path) of the substrate block available?

ELIGIBILITY

Challenge problems are open to academia, small and large businesses, national laboratories, both in the U.S. and internationally. Data packages will be publicly released by AFRL and made available to participants. Contact information will be required for accessing data packages to allow for challenge updates to be communicated but does not obligate participation. Submissions will be kept private during grading but will be shared anonymously as an aggregate after grading. Only AFRL will know the identity associated with each submission at any time. If a participant is selected as a Top Performer in one of the challenges, then that participant must agree to openly associate their identity with the submission. Participants will be informed of their submission’s grade/quality and can see other submissions in an anonymous form in a debrief document/journal article. Participants interested in having their submission graded, but not considered for an award, may discuss with AFRL regarding inclusion in the anonymous aggregate. Any questions regarding eligibility to participate or accept Top Performer recognitions should be addressed to AFRL through the challenge participation site. Participants may submit multiple times but must detail the differences in their modeling approach in the form of a brief abstract, as detailed in the grading section below. Participants should limit themselves to a reasonable number of submissions and are subject to AFRL discretion on how many will be reviewed if more are submitted.

TOP PERFORMERS - RECOGNITIONS

Challenge problems will have a plaques and recognitions for all Top Performers. All participants are eligible for Top Performer recognitions.

AWARDS

There is also an associated monetary and/or resource award that may be issued to Top Performers if they meet eligibility requirements and submit the required documentation. Exact details of the breakdown of prize money/resources is under development and award amounts will be communicated through the challenge participation site and may be determined by number of participants. Likely, winners and runners-up of each challenge will receive awards. An independent third-party will coordinate awards with selected Top Performers.

AWARD ELIGIBILITY

To be eligible for the monetary and/or resource award, participants must completely fill in and sign the Air Force Research Laboratory (AFRL) Additive Manufacturing (AM) Modeling Challenge Series Agreement with the National Center for Defense Manufacturing and Machining (NCDMM) who will act as the independent third-party administrator for the awards. Submit your completed agreement to challenge@americamakes.us within a week of challenge submission.

LOGISTICS

Data packages will carry unlimited data rights to the participants with proper citation. Submissions will be the property of the corresponding participant. AFRL is requesting rights to use the submissions in derivative work, but will not distribute submissions directly without approval of the corresponding participants. ARFL will use submissions as an anonymous aggregate unless participants make individual arrangements with AFRL. Participants will be asked to submit results and relevant details of their simulation procedure using a template provided with the challenge problem data package. The simulation details will be treated with the same rights and procedures as the results themselves. It should be noted that AFRL is not setting any constraints on the modeling approaches allowed by participants. Further, any additional information available in literature can be utilized by participants, but should be reported with the submission.

GRADING

Challenges will be graded by awarding points for each requested prediction, based on accuracy. For each requested prediction, levels generally deemed “Excellent” > “Good” > “Acceptable” have been established and points will be awarded in increasing value for obtaining each level. Each challenge problem consists of numerous (> 10) requested predictions. It is not required that any given submission make all requested predictions, but no points will be earned for unmade predictions. Each submission should have a brief abstract accompanying the provided answer template detailing the general modeling methodology used to make the predictions. Lack of an abstract and refusal to provide one may result in decremented points. Only answers submitted in the provided answer template (included in the data set) will be graded.

SUBMISSION

The answer templates included in each challenge data set must be used for all submissions. Each answer template includes an overview sheet that must be completely filled in. Answers submitted in any other form will not be graded.

Record the generated submission number for personal records and for the NCDMM award eligibility agreement. To be eligible for award through NCDMM, teams/participants must submit a completed agreement to NCDMM no later than 2 weeks after challenge submissions are due.

SCHEDULE

Schedule updates will be posted here and communicated to all registered participants via email. The deadline to submit challenge answers is January 15, 2020. AFRL will grade submissions and announce Top Performers approximately four (4) months after the submission deadline. A workshop will be hosted to debrief participants on the results and honor the Top Performers.

QUESTIONS

All questions related to the challenge and the data provided should be directed to marie.cox@us.af.mil.

DISCLAIMER

The Air Force Research Laboratory (AFRL) uses its best efforts to deliver a high-quality data and analysis. However, AFRL makes no warranties to that effect, and AFRL shall not be liable for any damage that may result from errors or omissions in the provided data.

The appearance of hyperlinks does not constitute endorsement by the U.S. Air Force of this Web site or the information, products, or services contained therein. For other than authorized activities such as military exchanges and morale, welfare and recreation sites, the U.S. Air Force does not exercise any editorial control over the information you may find at these locations. Such links are provided consistent with the stated purpose of this DoD Web site