An official website of the United States government
Here’s how you know
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.
CHARACTERIZATION OF ELASTOPLASTIC PROPERTIES OF ADDITIVELY MANUFACTURED SPECIMENS FROM INDENTATION DATA USING STOCHASTIC INVERSE MODELING
Published
Author(s)
Ridwan Olabiyi, Jordan Weaver, Ashif Sikandar Iquebal
Abstract
Rapid characterization of the mechanical properties and material structure of additively manufactured components via nondestructive techniques is becoming the sine-qua-non for their wider adoption. In this research, we primarily focus on estimating the elastoplastic properties of AM components from instrumented indentation measurements, often referred as the inverse indentation problem. While several empirical correlations and machine learning models have been developed for the inverse indentation problem, they either generalize poorly beyond the training datasets or do not provide a framework to accurately estimate the variability in material properties. In this work, we focus on alternate formulation of the inverse indentation problem, referred to as the stochastic inverse problem (SIP). In particular, SIP aims to estimate a distribution over the elastoplastic properties which when propagated through a forward model matches the observed force-displacement, i.e., the indentation data distribution. We implement our methodology for characterizing additive manufactured components that are subject to different heat treatments. The results demonstrate that our methodology accurately predicts the average elastoplastic characteristics, the Young's modulus (E) to within 1% and yield strength () to within 5% of the actual average values. Additionally, the recovered property distributions closely match those from standard tensile tests, as evidenced by KL divergence values approaching zero for the respective distributions. Based on the results, we are not only able to accurately estimate the mean elastoplastic properties, but also the associated variability. This allows us to distinguish the elastoplastic properties of the two groups from high throughput indentation measurements.
Proceedings Title
Proceedings of the ASME 2024 Manufacturing Science and Engineering Conference MSEC2024
Conference Dates
July 17-21, 2024
Conference Location
Knoxville, TN, US
Conference Title
MSEC 2024 Manufacturing Science and Engineering Conference
Olabiyi, R.
, Weaver, J.
and Iquebal, A.
(2024),
CHARACTERIZATION OF ELASTOPLASTIC PROPERTIES OF ADDITIVELY MANUFACTURED SPECIMENS FROM INDENTATION DATA USING STOCHASTIC INVERSE MODELING, Proceedings of the ASME 2024 Manufacturing Science and Engineering Conference MSEC2024, Knoxville, TN, US, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=957004
(Accessed October 7, 2024)