Unsupervised Deep Learning for Advanced Forming Limit Analysis in Sheet Metal: A Tensile Test-Based Approach
An accurate description of the formability and failure behavior of sheet metal materials is essential for an optimal forming process design. In this respect, the forming limit curve (FLC) based on the Nakajima test, which is determined in accordance with DIN EN ISO 12004-2, is a wide-spread procedur...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-11-01
|
Series: | Materials |
Subjects: | |
Online Access: | https://www.mdpi.com/1996-1944/16/21/7001 |
_version_ | 1797631670134767616 |
---|---|
author | Aleksandra Thamm Florian Thamm Annette Sawodny Sally Zeitler Marion Merklein Andreas Maier |
author_facet | Aleksandra Thamm Florian Thamm Annette Sawodny Sally Zeitler Marion Merklein Andreas Maier |
author_sort | Aleksandra Thamm |
collection | DOAJ |
description | An accurate description of the formability and failure behavior of sheet metal materials is essential for an optimal forming process design. In this respect, the forming limit curve (FLC) based on the Nakajima test, which is determined in accordance with DIN EN ISO 12004-2, is a wide-spread procedure for evaluating the formability of sheet metal materials. Thereby the FLC is affected by influences originating from intrinsic factors of the Nakajima test-setup, such as friction, which leads to deviations from the linear strain path, biaxial prestress and bending superposition. These disadvantages can be circumvented by an alternative test combination of uniaxial tensile test and hydraulic bulge test. In addition, the forming limit capacity of many lightweight materials is underestimated using the cross-section method according to DIN EN ISO 12004-2, due to the material-dependent occurrence of multiple strain maxima during forming or sudden cracking without prior necking. In this regard, machine learning approaches have a high potential for a more accurate determination of the forming limit curve due to the inclusion of other parameters influencing formability. This work presents a machine learning approach focused on uniaxial tensile tests to define the forming limit of lightweight materials and high-strength steels. The transferability of an existing weakly supervised convolutional neural network (CNN) approach was examined, originally designed for Nakajima tests, to uniaxial tensile tests. Additionally, a stereo camera-based method for this purpose was developed. In our evaluation, we train and test materials, including AA6016, DX54D, and DP800, through iterative data composition, using cross-validation. In the context of our stereo camera-based approach, strains for different materials and thicknesses were predicted. In this cases, our method successfully predicted the major strains with close agreement to ISO standards. For DX54D, with a thickness of 0.8 mm, the prediction was <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.659</mn></mrow></semantics></math></inline-formula> (compared to ISO’s <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.664</mn></mrow></semantics></math></inline-formula>). Similarly, for DX54D, 2.0 mm thickness, the predicted major strain was <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.780</mn></mrow></semantics></math></inline-formula> (compared to ISO <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.705</mn></mrow></semantics></math></inline-formula>), and for AA6016, at 1.0 mm thickness, a major strain of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.314</mn></mrow></semantics></math></inline-formula> (in line with ISO <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.309</mn></mrow></semantics></math></inline-formula>) was estimated. However, for DP800 with a thickness of 1.0 mm, the prediction yielded a major strain of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.478</mn></mrow></semantics></math></inline-formula> (as compared to ISO <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.289</mn></mrow></semantics></math></inline-formula>), indicating a divergence from the ISO standard in this particular case. These results in general, generated with the CNN stereo camera-based approach, underline the quantitative alignment of the approach with the cross-section method. |
first_indexed | 2024-03-11T11:26:33Z |
format | Article |
id | doaj.art-a5c9bb190b6049c68f60bc6315d8c57d |
institution | Directory Open Access Journal |
issn | 1996-1944 |
language | English |
last_indexed | 2024-03-11T11:26:33Z |
publishDate | 2023-11-01 |
publisher | MDPI AG |
record_format | Article |
series | Materials |
spelling | doaj.art-a5c9bb190b6049c68f60bc6315d8c57d2023-11-10T15:07:35ZengMDPI AGMaterials1996-19442023-11-011621700110.3390/ma16217001Unsupervised Deep Learning for Advanced Forming Limit Analysis in Sheet Metal: A Tensile Test-Based ApproachAleksandra Thamm0Florian Thamm1Annette Sawodny2Sally Zeitler3Marion Merklein4Andreas Maier5Pattern Recognition Lab, Friedrich-Alexander-Universität Erlangen-Nürnberg, Martensstr. 3, 91058 Erlangen, GermanyPattern Recognition Lab, Friedrich-Alexander-Universität Erlangen-Nürnberg, Martensstr. 3, 91058 Erlangen, GermanyInstitute of Manufacturing Technology, Friedrich-Alexander-Universität Erlangen-Nürnberg, Egerlandstr. 13, 91058 Erlangen, GermanyPattern Recognition Lab, Friedrich-Alexander-Universität Erlangen-Nürnberg, Martensstr. 3, 91058 Erlangen, GermanyInstitute of Manufacturing Technology, Friedrich-Alexander-Universität Erlangen-Nürnberg, Egerlandstr. 13, 91058 Erlangen, GermanyPattern Recognition Lab, Friedrich-Alexander-Universität Erlangen-Nürnberg, Martensstr. 3, 91058 Erlangen, GermanyAn accurate description of the formability and failure behavior of sheet metal materials is essential for an optimal forming process design. In this respect, the forming limit curve (FLC) based on the Nakajima test, which is determined in accordance with DIN EN ISO 12004-2, is a wide-spread procedure for evaluating the formability of sheet metal materials. Thereby the FLC is affected by influences originating from intrinsic factors of the Nakajima test-setup, such as friction, which leads to deviations from the linear strain path, biaxial prestress and bending superposition. These disadvantages can be circumvented by an alternative test combination of uniaxial tensile test and hydraulic bulge test. In addition, the forming limit capacity of many lightweight materials is underestimated using the cross-section method according to DIN EN ISO 12004-2, due to the material-dependent occurrence of multiple strain maxima during forming or sudden cracking without prior necking. In this regard, machine learning approaches have a high potential for a more accurate determination of the forming limit curve due to the inclusion of other parameters influencing formability. This work presents a machine learning approach focused on uniaxial tensile tests to define the forming limit of lightweight materials and high-strength steels. The transferability of an existing weakly supervised convolutional neural network (CNN) approach was examined, originally designed for Nakajima tests, to uniaxial tensile tests. Additionally, a stereo camera-based method for this purpose was developed. In our evaluation, we train and test materials, including AA6016, DX54D, and DP800, through iterative data composition, using cross-validation. In the context of our stereo camera-based approach, strains for different materials and thicknesses were predicted. In this cases, our method successfully predicted the major strains with close agreement to ISO standards. For DX54D, with a thickness of 0.8 mm, the prediction was <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.659</mn></mrow></semantics></math></inline-formula> (compared to ISO’s <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.664</mn></mrow></semantics></math></inline-formula>). Similarly, for DX54D, 2.0 mm thickness, the predicted major strain was <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.780</mn></mrow></semantics></math></inline-formula> (compared to ISO <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.705</mn></mrow></semantics></math></inline-formula>), and for AA6016, at 1.0 mm thickness, a major strain of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.314</mn></mrow></semantics></math></inline-formula> (in line with ISO <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.309</mn></mrow></semantics></math></inline-formula>) was estimated. However, for DP800 with a thickness of 1.0 mm, the prediction yielded a major strain of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.478</mn></mrow></semantics></math></inline-formula> (as compared to ISO <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>0.289</mn></mrow></semantics></math></inline-formula>), indicating a divergence from the ISO standard in this particular case. These results in general, generated with the CNN stereo camera-based approach, underline the quantitative alignment of the approach with the cross-section method.https://www.mdpi.com/1996-1944/16/21/7001forming limit curvepattern recognitionsheet metal formingmachine learning |
spellingShingle | Aleksandra Thamm Florian Thamm Annette Sawodny Sally Zeitler Marion Merklein Andreas Maier Unsupervised Deep Learning for Advanced Forming Limit Analysis in Sheet Metal: A Tensile Test-Based Approach Materials forming limit curve pattern recognition sheet metal forming machine learning |
title | Unsupervised Deep Learning for Advanced Forming Limit Analysis in Sheet Metal: A Tensile Test-Based Approach |
title_full | Unsupervised Deep Learning for Advanced Forming Limit Analysis in Sheet Metal: A Tensile Test-Based Approach |
title_fullStr | Unsupervised Deep Learning for Advanced Forming Limit Analysis in Sheet Metal: A Tensile Test-Based Approach |
title_full_unstemmed | Unsupervised Deep Learning for Advanced Forming Limit Analysis in Sheet Metal: A Tensile Test-Based Approach |
title_short | Unsupervised Deep Learning for Advanced Forming Limit Analysis in Sheet Metal: A Tensile Test-Based Approach |
title_sort | unsupervised deep learning for advanced forming limit analysis in sheet metal a tensile test based approach |
topic | forming limit curve pattern recognition sheet metal forming machine learning |
url | https://www.mdpi.com/1996-1944/16/21/7001 |
work_keys_str_mv | AT aleksandrathamm unsuperviseddeeplearningforadvancedforminglimitanalysisinsheetmetalatensiletestbasedapproach AT florianthamm unsuperviseddeeplearningforadvancedforminglimitanalysisinsheetmetalatensiletestbasedapproach AT annettesawodny unsuperviseddeeplearningforadvancedforminglimitanalysisinsheetmetalatensiletestbasedapproach AT sallyzeitler unsuperviseddeeplearningforadvancedforminglimitanalysisinsheetmetalatensiletestbasedapproach AT marionmerklein unsuperviseddeeplearningforadvancedforminglimitanalysisinsheetmetalatensiletestbasedapproach AT andreasmaier unsuperviseddeeplearningforadvancedforminglimitanalysisinsheetmetalatensiletestbasedapproach |