Details
| Originalsprache | Englisch |
|---|---|
| Titel des Sammelwerks | Lecture Notes in Production Engineering |
| Herausgeber (Verlag) | Springer Nature |
| Seiten | 94-103 |
| Seitenumfang | 10 |
| ISBN (elektronisch) | 978-3-031-47394-4 |
| ISBN (Print) | 978-3-031-47393-7 |
| Publikationsstatus | Veröffentlicht - 18 Nov. 2023 |
Publikationsreihe
| Name | Lecture Notes in Production Engineering |
|---|---|
| Band | Part F1764 |
| ISSN (Print) | 2194-0525 |
| ISSN (elektronisch) | 2194-0533 |
Abstract
The prediction of workpiece quality in process planning, using machine learning models, is a common-researched topic. Until now, trained models were static and could not update themselves with new data. However, this aspect is crucial when considering the continuously changing manufacturing circumstances in regards to new process parameters, materials, and workpiece geometries. In addition, repeatedly training process models with an extended mixed dataset decreases the prediction quality due to the increased data divergence. This paper presents an approach to automatically generate sub-models, which maintain the prediction quality even if novel data is considered. The challenge is to define the amount and content of these sub-models through clustering. Tool grinding experiments will be conducted with different process parameters, materials, and workpiece geometries in order to obtain a divergent dataset. Subsequently, cluster approaches are compared to obtain dynamic growing models, which enable optimized planning for a more resource efficient process. Finally, the method will be generalized in order to ensure a process-independent usage.
ASJC Scopus Sachgebiete
- Ingenieurwesen (insg.)
- Wirtschaftsingenieurwesen und Fertigungstechnik
- Volkswirtschaftslehre, Ökonometrie und Finanzen (insg.)
- Volkswirtschaftslehre, Ökonometrie und Finanzen (sonstige)
- Ingenieurwesen (insg.)
- Sicherheit, Risiko, Zuverlässigkeit und Qualität
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
Lecture Notes in Production Engineering. Springer Nature, 2023. S. 94-103 (Lecture Notes in Production Engineering; Band Part F1764).
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Beitrag in Buch/Sammelwerk › Forschung › Peer-Review
}
TY - CHAP
T1 - Clustering of Learning Sub-models for Quality Prediction in a Resource-Efficient Tool Grinding Process
AU - Denkena, Berend
AU - Wichmann, Marcel
AU - Wulf, Michael
N1 - Funding Information: Acknowledgement. The authors would like to thank the German Research Foundation (DFG) for funding the project LearnWZS - Learning process adaptation for tool grinding (number 445811009), which enables this investigation. Furthermore, the authors thank the Sieglinde-Vollmer Foundation for supporting this research.
PY - 2023/11/18
Y1 - 2023/11/18
N2 - The prediction of workpiece quality in process planning, using machine learning models, is a common-researched topic. Until now, trained models were static and could not update themselves with new data. However, this aspect is crucial when considering the continuously changing manufacturing circumstances in regards to new process parameters, materials, and workpiece geometries. In addition, repeatedly training process models with an extended mixed dataset decreases the prediction quality due to the increased data divergence. This paper presents an approach to automatically generate sub-models, which maintain the prediction quality even if novel data is considered. The challenge is to define the amount and content of these sub-models through clustering. Tool grinding experiments will be conducted with different process parameters, materials, and workpiece geometries in order to obtain a divergent dataset. Subsequently, cluster approaches are compared to obtain dynamic growing models, which enable optimized planning for a more resource efficient process. Finally, the method will be generalized in order to ensure a process-independent usage.
AB - The prediction of workpiece quality in process planning, using machine learning models, is a common-researched topic. Until now, trained models were static and could not update themselves with new data. However, this aspect is crucial when considering the continuously changing manufacturing circumstances in regards to new process parameters, materials, and workpiece geometries. In addition, repeatedly training process models with an extended mixed dataset decreases the prediction quality due to the increased data divergence. This paper presents an approach to automatically generate sub-models, which maintain the prediction quality even if novel data is considered. The challenge is to define the amount and content of these sub-models through clustering. Tool grinding experiments will be conducted with different process parameters, materials, and workpiece geometries in order to obtain a divergent dataset. Subsequently, cluster approaches are compared to obtain dynamic growing models, which enable optimized planning for a more resource efficient process. Finally, the method will be generalized in order to ensure a process-independent usage.
KW - Machine Learning
KW - Process Models
KW - Tool Grinding
UR - http://www.scopus.com/inward/record.url?scp=85178357560&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-47394-4_10
DO - 10.1007/978-3-031-47394-4_10
M3 - Contribution to book/anthology
AN - SCOPUS:85178357560
SN - 978-3-031-47393-7
T3 - Lecture Notes in Production Engineering
SP - 94
EP - 103
BT - Lecture Notes in Production Engineering
PB - Springer Nature
ER -