Combined learning processes for injection moulding based on simulation and experimental data. Hopmann, C., Jeschke, S., Meisen, T., Thiele, T. D., Tercan, H., Liebenberg, M., Heinisch, J., & Theunissen, M. In 33rd International Conference of the Polymer Processing Society : PPS-33 : Cancun, Mexico, December 10-14, 2017, Lawrence, KS, Dec, 2017. 33rd International Conference of the Polymer Processing Society, Cancún (Mexico), 10 Dec 2017 - 14 Dec 2017, Polymer Processing Society. Paper abstract bibtex Injection moulding enables the production of complex formed high-quality plastics parts in a single production step. To achieve a high and constant product quality, an appropriate process set-up with regards to product quality and process robustness is essential. A conventional process set-up requires expensive and time consuming experiments and know-how from the machine operator. One way to overcome this challenge is to make use of machine learning methods for process set-up. These methods can model the relationship between setting parameters and quality values and thus enable the identification of optimal working points. However, the training required for accurate modelling needs experimental data from extensive experiments for each process, too. Numerical simulation can predict quality values based on setting parameters without practical experiments. Whereas trends and the general dependencies between the parameters can be predicted with a satisfying accuracy, a certain discrepancy between the prediction of the simulation and the real process cannot be excluded. A combined approach using data from injection moulding simulations as well as experimental data appears promising to overcome the detriments of the solitary use of simulation for the training of machine learning algorithms. General dependencies could be gained from simulations without practical experiments and fine-tuning could be achieved by experimental trials with a minimal scope. In this paper, data obtained from a 2.5 D injection moulding simulation is compared with experimental data from a plate specimen and a complex formed injection moulded part. Therefore, central composed designs of experiments are used to identify differences in the effects and interdependencies of six setting parameters on quality values like part weight and dimensions. Furthermore, differences in the absolute values and the functionality of the effects are considered. On this basis, a combined machine learning concept using simulation and experimental data is presented.
@inproceedings {LiebenbergPPS17,
author = {Hopmann, Christian and Jeschke, Sabina and Meisen, Tobias
and Thiele, Thomas David and Tercan, Hasan and Liebenberg,
Martin and Heinisch, Julian and Theunissen, Matthias},
title = {{C}ombined learning processes for injection moulding based
on simulation and experimental data},
address = {Lawrence, KS},
publisher = {Polymer Processing Society},
reportid = {RWTH-2018-222255},
year = {2017},
comment = {33rd International Conference of the Polymer Processing
Society : PPS-33 : Cancun, Mexico, December 10-14, 2017},
booktitle = {33rd International Conference of the
Polymer Processing Society : PPS-33 :
Cancun, Mexico, December 10-14, 2017},
month = {Dec},
date = {2017-12-10},
organization = {33rd International Conference of the
Polymer Processing Society, Cancún
(Mexico), 10 Dec 2017 - 14 Dec 2017},
cin = {417810 / 416910 / 052100},
cid = {$I:(DE-82)417810_20140620$ / $I:(DE-82)416910_20140620$ /
$I:(DE-82)052100_20140620$},
typ = {PUB:(DE-HGF)8},
url = {https://publications.rwth-aachen.de/record/719619},
abstract = {Injection moulding enables the production of complex
formed high-quality plastics parts in a single production
step. To achieve a high and constant product quality, an
appropriate process set-up with regards to product quality and
process robustness is essential. A conventional process
set-up requires expensive and time consuming experiments
and know-how from the machine operator. One way to overcome
this challenge is to make use of machine learning methods
for process set-up. These methods can model the relationship
between setting parameters and quality values and thus enable
the identification of optimal working points. However,
the training required for accurate modelling needs experimental
data from extensive experiments for each process, too.
Numerical simulation can predict quality values based on
setting parameters without practical experiments. Whereas
trends and the general dependencies between the parameters
can be predicted with a satisfying accuracy, a certain
discrepancy between the prediction of the simulation and
the real process cannot be excluded. A combined approach
using data from injection moulding simulations as well as
experimental data appears promising to overcome the detriments
of the solitary use of simulation for the training of machine
learning algorithms. General dependencies could be gained
from simulations without practical experiments and fine-tuning
could be achieved by experimental trials with a minimal scope.
In this paper, data obtained from a 2.5 D injection moulding
simulation is compared with experimental data from a plate
specimen and a complex formed injection moulded part.
Therefore, central composed designs of experiments are used
to identify differences in the effects and interdependencies
of six setting parameters on quality values like part weight
and dimensions. Furthermore, differences in the absolute
values and the functionality of the effects are considered.
On this basis, a combined machine learning concept using
simulation and experimental data is presented.},
}
Downloads: 0
{"_id":"8DtE9HyFBdc5RPoGM","bibbaseid":"hopmann-jeschke-meisen-thiele-tercan-liebenberg-heinisch-theunissen-combinedlearningprocessesforinjectionmouldingbasedonsimulationandexperimentaldata-2017","author_short":["Hopmann, C.","Jeschke, S.","Meisen, T.","Thiele, T. D.","Tercan, H.","Liebenberg, M.","Heinisch, J.","Theunissen, M."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"propositions":[],"lastnames":["Hopmann"],"firstnames":["Christian"],"suffixes":[]},{"propositions":[],"lastnames":["Jeschke"],"firstnames":["Sabina"],"suffixes":[]},{"propositions":[],"lastnames":["Meisen"],"firstnames":["Tobias"],"suffixes":[]},{"propositions":[],"lastnames":["Thiele"],"firstnames":["Thomas","David"],"suffixes":[]},{"propositions":[],"lastnames":["Tercan"],"firstnames":["Hasan"],"suffixes":[]},{"propositions":[],"lastnames":["Liebenberg"],"firstnames":["Martin"],"suffixes":[]},{"propositions":[],"lastnames":["Heinisch"],"firstnames":["Julian"],"suffixes":[]},{"propositions":[],"lastnames":["Theunissen"],"firstnames":["Matthias"],"suffixes":[]}],"title":"Combined learning processes for injection moulding based on simulation and experimental data","address":"Lawrence, KS","publisher":"Polymer Processing Society","reportid":"RWTH-2018-222255","year":"2017","comment":"33rd International Conference of the Polymer Processing Society : PPS-33 : Cancun, Mexico, December 10-14, 2017","booktitle":"33rd International Conference of the Polymer Processing Society : PPS-33 : Cancun, Mexico, December 10-14, 2017","month":"Dec","date":"2017-12-10","organization":"33rd International Conference of the Polymer Processing Society, Cancún (Mexico), 10 Dec 2017 - 14 Dec 2017","cin":"417810 / 416910 / 052100","cid":"$I:(DE-82)417810_20140620$ / $I:(DE-82)416910_20140620$ / $I:(DE-82)052100_20140620$","typ":"PUB:(DE-HGF)8","url":"https://publications.rwth-aachen.de/record/719619","abstract":"Injection moulding enables the production of complex formed high-quality plastics parts in a single production step. To achieve a high and constant product quality, an appropriate process set-up with regards to product quality and process robustness is essential. A conventional process set-up requires expensive and time consuming experiments and know-how from the machine operator. One way to overcome this challenge is to make use of machine learning methods for process set-up. These methods can model the relationship between setting parameters and quality values and thus enable the identification of optimal working points. However, the training required for accurate modelling needs experimental data from extensive experiments for each process, too. Numerical simulation can predict quality values based on setting parameters without practical experiments. Whereas trends and the general dependencies between the parameters can be predicted with a satisfying accuracy, a certain discrepancy between the prediction of the simulation and the real process cannot be excluded. A combined approach using data from injection moulding simulations as well as experimental data appears promising to overcome the detriments of the solitary use of simulation for the training of machine learning algorithms. General dependencies could be gained from simulations without practical experiments and fine-tuning could be achieved by experimental trials with a minimal scope. In this paper, data obtained from a 2.5 D injection moulding simulation is compared with experimental data from a plate specimen and a complex formed injection moulded part. Therefore, central composed designs of experiments are used to identify differences in the effects and interdependencies of six setting parameters on quality values like part weight and dimensions. Furthermore, differences in the absolute values and the functionality of the effects are considered. On this basis, a combined machine learning concept using simulation and experimental data is presented.","bibtex":"@inproceedings {LiebenbergPPS17,\n author = {Hopmann, Christian and Jeschke, Sabina and Meisen, Tobias\n and Thiele, Thomas David and Tercan, Hasan and Liebenberg,\n Martin and Heinisch, Julian and Theunissen, Matthias},\n title = {{C}ombined learning processes for injection moulding based\n on simulation and experimental data},\n address = {Lawrence, KS},\n publisher = {Polymer Processing Society},\n reportid = {RWTH-2018-222255},\n year = {2017},\n comment = {33rd International Conference of the Polymer Processing\n Society : PPS-33 : Cancun, Mexico, December 10-14, 2017},\n booktitle = {33rd International Conference of the\n Polymer Processing Society : PPS-33 :\n Cancun, Mexico, December 10-14, 2017},\n month = {Dec},\n date = {2017-12-10},\n organization = {33rd International Conference of the\n Polymer Processing Society, Cancún\n (Mexico), 10 Dec 2017 - 14 Dec 2017},\n cin = {417810 / 416910 / 052100},\n cid = {$I:(DE-82)417810_20140620$ / $I:(DE-82)416910_20140620$ /\n $I:(DE-82)052100_20140620$},\n typ = {PUB:(DE-HGF)8},\n url = {https://publications.rwth-aachen.de/record/719619},\n abstract = {Injection moulding enables the production of complex\n formed high-quality plastics parts in a single production\n step. To achieve a high and constant product quality, an\n appropriate process set-up with regards to product quality and\n process robustness is essential. A conventional process\n set-up requires expensive and time consuming experiments\n and know-how from the machine operator. One way to overcome\n this challenge is to make use of machine learning methods\n for process set-up. These methods can model the relationship\n between setting parameters and quality values and thus enable\n the identification of optimal working points. However,\n the training required for accurate modelling needs experimental\n data from extensive experiments for each process, too.\n Numerical simulation can predict quality values based on\n setting parameters without practical experiments. Whereas\n trends and the general dependencies between the parameters\n can be predicted with a satisfying accuracy, a certain\n discrepancy between the prediction of the simulation and\n the real process cannot be excluded. A combined approach\n using data from injection moulding simulations as well as\n experimental data appears promising to overcome the detriments\n of the solitary use of simulation for the training of machine\n learning algorithms. General dependencies could be gained\n from simulations without practical experiments and fine-tuning\n could be achieved by experimental trials with a minimal scope.\n In this paper, data obtained from a 2.5 D injection moulding\n simulation is compared with experimental data from a plate\n specimen and a complex formed injection moulded part.\n Therefore, central composed designs of experiments are used\n to identify differences in the effects and interdependencies\n of six setting parameters on quality values like part weight\n and dimensions. Furthermore, differences in the absolute\n values and the functionality of the effects are considered.\n On this basis, a combined machine learning concept using\n simulation and experimental data is presented.},\n}\n\n","author_short":["Hopmann, C.","Jeschke, S.","Meisen, T.","Thiele, T. D.","Tercan, H.","Liebenberg, M.","Heinisch, J.","Theunissen, M."],"key":"LiebenbergPPS17","id":"LiebenbergPPS17","bibbaseid":"hopmann-jeschke-meisen-thiele-tercan-liebenberg-heinisch-theunissen-combinedlearningprocessesforinjectionmouldingbasedonsimulationandexperimentaldata-2017","role":"author","urls":{"Paper":"https://publications.rwth-aachen.de/record/719619"},"metadata":{"authorlinks":{}}},"bibtype":"inproceedings","biburl":"https://kbsg.rwth-aachen.de/files/kbsgweb.bib","dataSources":["dqRQPSg6Hy3ZXQg7z"],"keywords":[],"search_terms":["combined","learning","processes","injection","moulding","based","simulation","experimental","data","hopmann","jeschke","meisen","thiele","tercan","liebenberg","heinisch","theunissen"],"title":"Combined learning processes for injection moulding based on simulation and experimental data","year":2017}