CNN-based preform design: effect of training data configuration on strain distribution in forged products
This study investigates the effect of convolutional neural network (CNN)–based preform design on uniform strain distribution and microstructure in the forging of harmonic drive flex splines. Two CNN-based strategies were employed to achieve uniform strain distribution: one using NURBS (non-uniform r...
Gespeichert in:
| Veröffentlicht in: | International journal of advanced manufacturing technology 2024-12, Vol.135 (9-10), p.4837-4854 |
|---|---|
| Hauptverfasser: | , , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 4854 |
|---|---|
| container_issue | 9-10 |
| container_start_page | 4837 |
| container_title | International journal of advanced manufacturing technology |
| container_volume | 135 |
| creator | Park, Joonhee Han, Byeongchan Choi, Jaegu Shin, Sangyun Kim, Naksoo |
| description | This study investigates the effect of convolutional neural network (CNN)–based preform design on uniform strain distribution and microstructure in the forging of harmonic drive flex splines. Two CNN-based strategies were employed to achieve uniform strain distribution: one using NURBS (non-uniform rational B-splines) curves to generate a comprehensive training dataset and the other iteratively refining the training dataset by incorporating preform shapes from previous iterations to enhance learning. The CNN-designed preform reduced mean strain by 4.81%, standard deviation by 25.95%, and maximum strain by 19.22%. It eliminated folding defects and reduced forging load by 70.7% during preform forging and 7.40% during final forging. Electron backscatter diffraction (EBSD) confirmed the process’s effectiveness by comparing the microstructure of the forged flex spline. After isothermal normalizing heat treatment, the kernel average misorientation (KAM) value decreased by an average of 0.194° in forgings using existing preforms and by 0.249° with CNN-designed preforms. |
| doi_str_mv | 10.1007/s00170-024-14768-1 |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_3130563527</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3130563527</sourcerecordid><originalsourceid>FETCH-LOGICAL-c1511-71d21defbddb534a428f2784df23e9d8e31988761c3df4fbd0020abc668623b13</originalsourceid><addsrcrecordid>eNp9kE1LxDAQhoMouK7-AU8Bz9FM0iZZb7L4Bct60XNIm6RkcZs1aQ_-e7Ot4E0YyHw88w55EboGeguUyrtMKUhKKKsIVFIoAidoARXnhFOoT9GCstLkZXKOLnLeFVyAUAsU1tstaUx2Fh-S8zHtsXU5dP09dt67dsDR4yGZ0Ie-w9YMBrex96EbkxlC7HGJPM2xDSUJzTi1S13Eukk22rEd8iU68-Yzu6vfd4k-nh7f1y9k8_b8un7YkBZqACLBMrDON9Y2Na9MxZRnUlXWM-5WVjkOK6WkgJZbXxWMUkZN0wqhBOMN8CW6mXXL4a_R5UHv4pj6clJz4LQWvGayUGym2hRzLj_XhxT2Jn1roPpoqZ4t1cVSPVmqj9J8XsoF7juX_qT_2foBnRB6NA</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3130563527</pqid></control><display><type>article</type><title>CNN-based preform design: effect of training data configuration on strain distribution in forged products</title><source>SpringerNature Journals</source><creator>Park, Joonhee ; Han, Byeongchan ; Choi, Jaegu ; Shin, Sangyun ; Kim, Naksoo</creator><creatorcontrib>Park, Joonhee ; Han, Byeongchan ; Choi, Jaegu ; Shin, Sangyun ; Kim, Naksoo</creatorcontrib><description>This study investigates the effect of convolutional neural network (CNN)–based preform design on uniform strain distribution and microstructure in the forging of harmonic drive flex splines. Two CNN-based strategies were employed to achieve uniform strain distribution: one using NURBS (non-uniform rational B-splines) curves to generate a comprehensive training dataset and the other iteratively refining the training dataset by incorporating preform shapes from previous iterations to enhance learning. The CNN-designed preform reduced mean strain by 4.81%, standard deviation by 25.95%, and maximum strain by 19.22%. It eliminated folding defects and reduced forging load by 70.7% during preform forging and 7.40% during final forging. Electron backscatter diffraction (EBSD) confirmed the process’s effectiveness by comparing the microstructure of the forged flex spline. After isothermal normalizing heat treatment, the kernel average misorientation (KAM) value decreased by an average of 0.194° in forgings using existing preforms and by 0.249° with CNN-designed preforms.</description><identifier>ISSN: 0268-3768</identifier><identifier>EISSN: 1433-3015</identifier><identifier>DOI: 10.1007/s00170-024-14768-1</identifier><language>eng</language><publisher>London: Springer London</publisher><subject>Artificial neural networks ; B spline functions ; CAE) and Design ; Computer-Aided Engineering (CAD ; Configuration management ; Datasets ; Electron back scatter ; Engineering ; Forging ; Forgings ; Heat treatment ; Industrial and Production Engineering ; Mechanical Engineering ; Media Management ; Microstructure ; Misalignment ; Normalizing (heat treatment) ; Original Article ; Preforms ; Strain distribution</subject><ispartof>International journal of advanced manufacturing technology, 2024-12, Vol.135 (9-10), p.4837-4854</ispartof><rights>The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c1511-71d21defbddb534a428f2784df23e9d8e31988761c3df4fbd0020abc668623b13</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://link.springer.com/content/pdf/10.1007/s00170-024-14768-1$$EPDF$$P50$$Gspringer$$H</linktopdf><linktohtml>$$Uhttps://link.springer.com/10.1007/s00170-024-14768-1$$EHTML$$P50$$Gspringer$$H</linktohtml><link.rule.ids>315,782,786,27931,27932,41495,42564,51326</link.rule.ids></links><search><creatorcontrib>Park, Joonhee</creatorcontrib><creatorcontrib>Han, Byeongchan</creatorcontrib><creatorcontrib>Choi, Jaegu</creatorcontrib><creatorcontrib>Shin, Sangyun</creatorcontrib><creatorcontrib>Kim, Naksoo</creatorcontrib><title>CNN-based preform design: effect of training data configuration on strain distribution in forged products</title><title>International journal of advanced manufacturing technology</title><addtitle>Int J Adv Manuf Technol</addtitle><description>This study investigates the effect of convolutional neural network (CNN)–based preform design on uniform strain distribution and microstructure in the forging of harmonic drive flex splines. Two CNN-based strategies were employed to achieve uniform strain distribution: one using NURBS (non-uniform rational B-splines) curves to generate a comprehensive training dataset and the other iteratively refining the training dataset by incorporating preform shapes from previous iterations to enhance learning. The CNN-designed preform reduced mean strain by 4.81%, standard deviation by 25.95%, and maximum strain by 19.22%. It eliminated folding defects and reduced forging load by 70.7% during preform forging and 7.40% during final forging. Electron backscatter diffraction (EBSD) confirmed the process’s effectiveness by comparing the microstructure of the forged flex spline. After isothermal normalizing heat treatment, the kernel average misorientation (KAM) value decreased by an average of 0.194° in forgings using existing preforms and by 0.249° with CNN-designed preforms.</description><subject>Artificial neural networks</subject><subject>B spline functions</subject><subject>CAE) and Design</subject><subject>Computer-Aided Engineering (CAD</subject><subject>Configuration management</subject><subject>Datasets</subject><subject>Electron back scatter</subject><subject>Engineering</subject><subject>Forging</subject><subject>Forgings</subject><subject>Heat treatment</subject><subject>Industrial and Production Engineering</subject><subject>Mechanical Engineering</subject><subject>Media Management</subject><subject>Microstructure</subject><subject>Misalignment</subject><subject>Normalizing (heat treatment)</subject><subject>Original Article</subject><subject>Preforms</subject><subject>Strain distribution</subject><issn>0268-3768</issn><issn>1433-3015</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><recordid>eNp9kE1LxDAQhoMouK7-AU8Bz9FM0iZZb7L4Bct60XNIm6RkcZs1aQ_-e7Ot4E0YyHw88w55EboGeguUyrtMKUhKKKsIVFIoAidoARXnhFOoT9GCstLkZXKOLnLeFVyAUAsU1tstaUx2Fh-S8zHtsXU5dP09dt67dsDR4yGZ0Ie-w9YMBrex96EbkxlC7HGJPM2xDSUJzTi1S13Eukk22rEd8iU68-Yzu6vfd4k-nh7f1y9k8_b8un7YkBZqACLBMrDON9Y2Na9MxZRnUlXWM-5WVjkOK6WkgJZbXxWMUkZN0wqhBOMN8CW6mXXL4a_R5UHv4pj6clJz4LQWvGayUGym2hRzLj_XhxT2Jn1roPpoqZ4t1cVSPVmqj9J8XsoF7juX_qT_2foBnRB6NA</recordid><startdate>20241201</startdate><enddate>20241201</enddate><creator>Park, Joonhee</creator><creator>Han, Byeongchan</creator><creator>Choi, Jaegu</creator><creator>Shin, Sangyun</creator><creator>Kim, Naksoo</creator><general>Springer London</general><general>Springer Nature B.V</general><scope>AAYXX</scope><scope>CITATION</scope></search><sort><creationdate>20241201</creationdate><title>CNN-based preform design: effect of training data configuration on strain distribution in forged products</title><author>Park, Joonhee ; Han, Byeongchan ; Choi, Jaegu ; Shin, Sangyun ; Kim, Naksoo</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c1511-71d21defbddb534a428f2784df23e9d8e31988761c3df4fbd0020abc668623b13</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Artificial neural networks</topic><topic>B spline functions</topic><topic>CAE) and Design</topic><topic>Computer-Aided Engineering (CAD</topic><topic>Configuration management</topic><topic>Datasets</topic><topic>Electron back scatter</topic><topic>Engineering</topic><topic>Forging</topic><topic>Forgings</topic><topic>Heat treatment</topic><topic>Industrial and Production Engineering</topic><topic>Mechanical Engineering</topic><topic>Media Management</topic><topic>Microstructure</topic><topic>Misalignment</topic><topic>Normalizing (heat treatment)</topic><topic>Original Article</topic><topic>Preforms</topic><topic>Strain distribution</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Park, Joonhee</creatorcontrib><creatorcontrib>Han, Byeongchan</creatorcontrib><creatorcontrib>Choi, Jaegu</creatorcontrib><creatorcontrib>Shin, Sangyun</creatorcontrib><creatorcontrib>Kim, Naksoo</creatorcontrib><collection>CrossRef</collection><jtitle>International journal of advanced manufacturing technology</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Park, Joonhee</au><au>Han, Byeongchan</au><au>Choi, Jaegu</au><au>Shin, Sangyun</au><au>Kim, Naksoo</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>CNN-based preform design: effect of training data configuration on strain distribution in forged products</atitle><jtitle>International journal of advanced manufacturing technology</jtitle><stitle>Int J Adv Manuf Technol</stitle><date>2024-12-01</date><risdate>2024</risdate><volume>135</volume><issue>9-10</issue><spage>4837</spage><epage>4854</epage><pages>4837-4854</pages><issn>0268-3768</issn><eissn>1433-3015</eissn><abstract>This study investigates the effect of convolutional neural network (CNN)–based preform design on uniform strain distribution and microstructure in the forging of harmonic drive flex splines. Two CNN-based strategies were employed to achieve uniform strain distribution: one using NURBS (non-uniform rational B-splines) curves to generate a comprehensive training dataset and the other iteratively refining the training dataset by incorporating preform shapes from previous iterations to enhance learning. The CNN-designed preform reduced mean strain by 4.81%, standard deviation by 25.95%, and maximum strain by 19.22%. It eliminated folding defects and reduced forging load by 70.7% during preform forging and 7.40% during final forging. Electron backscatter diffraction (EBSD) confirmed the process’s effectiveness by comparing the microstructure of the forged flex spline. After isothermal normalizing heat treatment, the kernel average misorientation (KAM) value decreased by an average of 0.194° in forgings using existing preforms and by 0.249° with CNN-designed preforms.</abstract><cop>London</cop><pub>Springer London</pub><doi>10.1007/s00170-024-14768-1</doi><tpages>18</tpages></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 0268-3768 |
| ispartof | International journal of advanced manufacturing technology, 2024-12, Vol.135 (9-10), p.4837-4854 |
| issn | 0268-3768 1433-3015 |
| language | eng |
| recordid | cdi_proquest_journals_3130563527 |
| source | SpringerNature Journals |
| subjects | Artificial neural networks B spline functions CAE) and Design Computer-Aided Engineering (CAD Configuration management Datasets Electron back scatter Engineering Forging Forgings Heat treatment Industrial and Production Engineering Mechanical Engineering Media Management Microstructure Misalignment Normalizing (heat treatment) Original Article Preforms Strain distribution |
| title | CNN-based preform design: effect of training data configuration on strain distribution in forged products |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-05T01%3A17%3A36IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=CNN-based%20preform%20design:%20effect%20of%20training%20data%20configuration%20on%20strain%20distribution%20in%20forged%20products&rft.jtitle=International%20journal%20of%20advanced%20manufacturing%20technology&rft.au=Park,%20Joonhee&rft.date=2024-12-01&rft.volume=135&rft.issue=9-10&rft.spage=4837&rft.epage=4854&rft.pages=4837-4854&rft.issn=0268-3768&rft.eissn=1433-3015&rft_id=info:doi/10.1007/s00170-024-14768-1&rft_dat=%3Cproquest_cross%3E3130563527%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3130563527&rft_id=info:pmid/&rfr_iscdi=true |