2019
Schmidt, Carsten; Hocke, Tristan; Denkena, Berend
Deep learning-based classification of production defects in automated-fiber-placement processes Artikel
In: Production Engineering, Bd. 13, 3-4, S. 501-509, 2019.
Abstract | Links | BibTeX | Schlagwörter: Automated Fiber Placement, Industry 4.0, Thermal Imaging
@article{Schmidt2019,
title = {Deep learning-based classification of production defects in automated-fiber-placement processes},
author = {Carsten Schmidt and Tristan Hocke and Berend Denkena},
url = {http://link.springer.com/article/10.1007/s11740-019-00893-4},
doi = {https://doi.org/10.1007/s11740-019-00893-4},
year = {2019},
date = {2019-03-15},
journal = {Production Engineering},
volume = {13, 3-4},
pages = {501-509},
abstract = {This paper presents a deep learning-based approach for the detection and classification of production defects that comple-
ments an existing thermographic online monitoring system for Automated-Fiber-Placement (AFP) processes. The detection
and classification procedure is performed in two stages. In the first stage, the system monitors each tow individually and
classifies its process status. Furthermore, it detects and classifies production defects that affect individual tows such as a
tow-twist. In the second stage, the system monitors the total width of the faultless tows. In this stage, production defects
effecting multiple tows, for example gaps or overlaps, are detected and classified. Twelve different deep convolution neural
networks (CNN) with three various architectures are learned supervised relating to different data sets. The performance of
both identification stages is explored separately before the entire system will be set up. Therefore, the thermal images of the
data sets are superimposed by noise to test the performance of the selected CNN.},
keywords = {Automated Fiber Placement, Industry 4.0, Thermal Imaging},
pubstate = {published},
tppubtype = {article}
}
This paper presents a deep learning-based approach for the detection and classification of production defects that comple-
ments an existing thermographic online monitoring system for Automated-Fiber-Placement (AFP) processes. The detection
and classification procedure is performed in two stages. In the first stage, the system monitors each tow individually and
classifies its process status. Furthermore, it detects and classifies production defects that affect individual tows such as a
tow-twist. In the second stage, the system monitors the total width of the faultless tows. In this stage, production defects
effecting multiple tows, for example gaps or overlaps, are detected and classified. Twelve different deep convolution neural
networks (CNN) with three various architectures are learned supervised relating to different data sets. The performance of
both identification stages is explored separately before the entire system will be set up. Therefore, the thermal images of the
data sets are superimposed by noise to test the performance of the selected CNN.
ments an existing thermographic online monitoring system for Automated-Fiber-Placement (AFP) processes. The detection
and classification procedure is performed in two stages. In the first stage, the system monitors each tow individually and
classifies its process status. Furthermore, it detects and classifies production defects that affect individual tows such as a
tow-twist. In the second stage, the system monitors the total width of the faultless tows. In this stage, production defects
effecting multiple tows, for example gaps or overlaps, are detected and classified. Twelve different deep convolution neural
networks (CNN) with three various architectures are learned supervised relating to different data sets. The performance of
both identification stages is explored separately before the entire system will be set up. Therefore, the thermal images of the
data sets are superimposed by noise to test the performance of the selected CNN.
2017
Brüning, Jan; Denkena, Berend; Dittrich, Marc-Andre; Hocke, Tristan
Machine Learning Approach for Optimization of Automated Fiber Placement Processes Konferenz
2017.
Links | BibTeX | Schlagwörter: Automated Fiber Placement, Industry 4.0, Process Monitoring
@conference{Brüning2017b,
title = {Machine Learning Approach for Optimization of Automated Fiber Placement Processes},
author = {Jan Brüning and Berend Denkena and Marc-Andre Dittrich and Tristan Hocke},
editor = {Procedia CIRP 66},
doi = {10.1016/j.procir.2017.03.295},
year = {2017},
date = {2017-06-07},
pages = {74-78},
keywords = {Automated Fiber Placement, Industry 4.0, Process Monitoring},
pubstate = {published},
tppubtype = {conference}
}
Schmidt, Carsten; Denkena, Berend; Hocke, Tristan; Völtzer, Klaas
Thermal imaging as a solution for reliable monitoring of AFP processes Konferenz
3rd ACM Automated Composites Manufacturing, Montreal, Canada, 2017.
BibTeX | Schlagwörter: Automated Fiber Placement, Industry 4.0, Manufacturing Quality, Process Monitoring, Thermal Imaging
@conference{Schmidt2017,
title = {Thermal imaging as a solution for reliable monitoring of AFP processes},
author = {Carsten Schmidt and Berend Denkena and Tristan Hocke and Klaas Völtzer},
year = {2017},
date = {2017-04-20},
booktitle = {3rd ACM Automated Composites Manufacturing},
address = {Montreal, Canada},
keywords = {Automated Fiber Placement, Industry 4.0, Manufacturing Quality, Process Monitoring, Thermal Imaging},
pubstate = {published},
tppubtype = {conference}
}
2016
Schmidt, Carsten; Weber, Patricc; Völtzer, Klaas; Deniz, Onur
Self-Configurable Production of CFRP Aerospace Components Based on Multi-Criteria Structural Optimization Konferenzbeitrag
In: CFK-Valley Convention, 2016.
Abstract | BibTeX | Schlagwörter: Composite Structures, Industry 4.0, Self Configurable Production
@inproceedings{Schmidt2016,
title = {Self-Configurable Production of CFRP Aerospace Components Based on Multi-Criteria Structural Optimization },
author = {Carsten Schmidt and Patricc Weber and Klaas Völtzer and Onur Deniz},
year = {2016},
date = {2016-06-17},
booktitle = {CFK-Valley Convention},
abstract = {Increased utilization of composite materials in a wide range of applications and industries due to their specific properties such as strength-to weight ratio, damage tolerance, reduced maintenance costs and flexibility let to advanced requirements on production methodologies and increased demands on lightweight construction. To be well prepared for tomorrow´s market, lightweight industries must be able to response quickly to customer needs as well as controlling costs and manufacturing quality. Accompanying challenges of manufacturing individually designed lightweight components with predominantly small lot sizes have to be met with flexible production systems in a quick reconfigurable production environment. Process reliability particularly in single part production is a big challenge. To minimize the risk associated therewith, criteria like component producibility already have to be considered in the design phase and monitoring of the manufacturing process becomes more relevant to ensure desired material properties.
Within the interdisciplinary research project “High Performance Production of CFRP-Structures” (HP CFK), a generic approach for simultaneously developing lightweight aerospace components, automated fiber placement system and processes is developed. This presentation introduces the new multicriteria optimization framework that efficiently involves corresponding manufacturing analysis of developed aerospace composite structures with respect to production costs, producibility and material characteristics. In addition, a newly developed reconfigurable automated fiber placement system with adaptive force controlled compaction unit, whose technical characteristics are represented within the optimization framework, and an online thermal imaging system for monitoring fiber placement manufacturing processes are presented. It is shown, how component topologies easily improve and adapt themselves to the corresponding production technology due to returned process knowledge and manufacturing restrictions and hereby avoiding time consuming redesign iterations. The coupling of optimization framework and manufacturing system lead to a self-configuration of automated fiber placement processes and an individually parametrization of the monitoring system for part specific online quality control.
},
keywords = {Composite Structures, Industry 4.0, Self Configurable Production},
pubstate = {published},
tppubtype = {inproceedings}
}
Increased utilization of composite materials in a wide range of applications and industries due to their specific properties such as strength-to weight ratio, damage tolerance, reduced maintenance costs and flexibility let to advanced requirements on production methodologies and increased demands on lightweight construction. To be well prepared for tomorrow´s market, lightweight industries must be able to response quickly to customer needs as well as controlling costs and manufacturing quality. Accompanying challenges of manufacturing individually designed lightweight components with predominantly small lot sizes have to be met with flexible production systems in a quick reconfigurable production environment. Process reliability particularly in single part production is a big challenge. To minimize the risk associated therewith, criteria like component producibility already have to be considered in the design phase and monitoring of the manufacturing process becomes more relevant to ensure desired material properties.
Within the interdisciplinary research project “High Performance Production of CFRP-Structures” (HP CFK), a generic approach for simultaneously developing lightweight aerospace components, automated fiber placement system and processes is developed. This presentation introduces the new multicriteria optimization framework that efficiently involves corresponding manufacturing analysis of developed aerospace composite structures with respect to production costs, producibility and material characteristics. In addition, a newly developed reconfigurable automated fiber placement system with adaptive force controlled compaction unit, whose technical characteristics are represented within the optimization framework, and an online thermal imaging system for monitoring fiber placement manufacturing processes are presented. It is shown, how component topologies easily improve and adapt themselves to the corresponding production technology due to returned process knowledge and manufacturing restrictions and hereby avoiding time consuming redesign iterations. The coupling of optimization framework and manufacturing system lead to a self-configuration of automated fiber placement processes and an individually parametrization of the monitoring system for part specific online quality control.
Within the interdisciplinary research project “High Performance Production of CFRP-Structures” (HP CFK), a generic approach for simultaneously developing lightweight aerospace components, automated fiber placement system and processes is developed. This presentation introduces the new multicriteria optimization framework that efficiently involves corresponding manufacturing analysis of developed aerospace composite structures with respect to production costs, producibility and material characteristics. In addition, a newly developed reconfigurable automated fiber placement system with adaptive force controlled compaction unit, whose technical characteristics are represented within the optimization framework, and an online thermal imaging system for monitoring fiber placement manufacturing processes are presented. It is shown, how component topologies easily improve and adapt themselves to the corresponding production technology due to returned process knowledge and manufacturing restrictions and hereby avoiding time consuming redesign iterations. The coupling of optimization framework and manufacturing system lead to a self-configuration of automated fiber placement processes and an individually parametrization of the monitoring system for part specific online quality control.