Implementation of modified void descriptor function to predict fracture location in additively manufactured AlSi10Mg

The interplay of many process variables associated with metal additive manufacturing causes uncertainties in the final printed part, resulting in limited adoption of metal AM in the industry. These uncertainties are typically due to defects causing unpredictable mechanical responses. Porosity, when exceeding 1%, significantly contributes to sample failure, one of the critical properties checked for real-world applications. The previously proposed Modified Void Descriptor Function (MVDF) aims to predict failure by incorporating three crucial factors - pore-pore interaction, stress concentration, and pore distance from surfaces, all weighted by pore volume. This approach showed promising results in predicting failures within 120 computationally generated pore networks. The proposed equation was further validated on six IN718 samples, accurately predicting failure in four of them. However, the data used is statistically insignificant to make conclusions, and validation across other materials is also required. The current study used a broad range of processing parameters to evaluate the MVDF function for 27 of 60 AlSi10Mg samples fabricated with laser powder bed fusion (PBF-LB). Additionally, the current study proposes a modified MVDF implementation to enhance the function’s predictive accuracy. The findings reveal that the updated MVDF implementation achieves a 37% better accuracy in predicting the failure location within a margin of ±2 mm. The new implementation demonstrates superior performance, accurately predicting failure in 48.14% of cases (13 out of 27 samples), compared to the original MVDF, which succeeded in only 11.11% (3 out of 27 samples). Upon analyzing three distinct factors individually, the study finds that at least one factor precisely forecasts failure locations in 66.6% (18 out of 27) samples within ±2 mm and 81.4% (22 out of 27) within ±4 mm of the actual failure location. This study thereby presents an improved MVDF approach, showcasing enhanced prediction accuracy.

Files

Metadata

Work Title Implementation of modified void descriptor function to predict fracture location in additively manufactured AlSi10Mg
Access
Open Access
Creators
  1. Yasham Mundada
License In Copyright (Rights Reserved)
Work Type Masters Culminating Experience
Sub Work Type Scholarly Paper/Essay (MA/MS)
Program Additive Manufacturing and Design
Degree Master of Science
Publication Date April 14, 2024
DOI doi:10.26207/7jkp-2481
Deposited June 10, 2024

Versions

Analytics

Collections

This resource is currently not in any collection.

Work History

Version 1
published

  • Created
  • Updated
  • Updated Description, Publication Date Show Changes
    Description
    • The interplay of many process variables associated with metal additive manufacturing causes uncertainties in the final printed part, resulting in limited adoption of metal AM in the industry. These uncertainties are typically due to defects causing unpredictable mechanical responses. Porosity, when exceeding 1%, significantly contributes to sample failure, one of the critical properties checked for real-world applications. The previously proposed Modified Void Descriptor Function (MVDF) aims to predict failure by incorporating three crucial factors - pore-pore interaction, stress concentration, and pore distance from surfaces, all weighted by pore volume. This approach showed promising results in predicting failures within 120 computationally generated pore networks. The proposed equation was further validated on six IN718 samples, accurately predicting failure in four of them. However, the data used is statistically insignificant to make conclusions, and validation across other materials is also required. The current study used a broad range of processing parameters to evaluate the MVDF function for 27 of 60 AlSi10Mg samples fabricated with laser powder bed fusion (PBF-LB). Additionally, the current study proposes a modified MVDF implementation to enhance the function’s predictive accuracy. The findings reveal that the updated MVDF implementation achieves a 37% better accuracy in predicting the failure location within a margin of ±2 mm. The new implementation demonstrates superior performance, accurately predicting failure in 48.14% of cases (13 out of 27 samples), compared to the original MVDF, which succeeded in only 11.11% (3 out of 27 samples). Upon analyzing three distinct factors individually, the study finds that at least one factor precisely forecasts failure locations in 66.6% (18 out of 27) samples within ±2 mm and 81.4% (22 out of 27) within ±4 mm of the actual failure location. This study thereby presents an improved MVDF approach, showcasing enhanced prediction accuracy.
    Publication Date
    • 2024-04-14
  • Added Creator Yasham Mundada
  • Added Mundada,Yasham-AMD Scholarly Paper.docx
  • Updated License Show Changes
    License
    • https://rightsstatements.org/page/InC/1.0/
  • Published

Version 2
published

  • Created
  • Deleted Mundada,Yasham-AMD Scholarly Paper.docx
  • Added Mundada,Yasham_AMD_Paper.pdf
  • Published
  • Updated
  • Updated Degree, Program, Sub Work Type Show Changes
    Degree
    • Master of Science
    Program
    • Additive Manufacturing and Design
    Sub Work Type
    • Scholarly Paper/Essay (MA/MS)