% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@PHDTHESIS{Mann:985472,
author = {Mann, Samuel Micha},
othercontributors = {Reisgen, Uwe and Rethmeier, Michael},
title = {{D}atenbasierte {E}rfassung und {R}egelung transienter
{Q}ualitätsmerkmale beim {M}etall-{S}chutzgasschweißen},
volume = {2024,2},
school = {RWTH Aachen University},
type = {Dissertation},
address = {Düren},
publisher = {Shaker},
reportid = {RWTH-2024-04642},
isbn = {978-3-8440-9478-7},
series = {Aachener Berichte Fügetechnik},
pages = {1 Online-Ressource : Illustrationen},
year = {2024},
note = {Druckausgabe: 2024. - Auch veröffentlicht auf dem
Publikationsserver der RWTH Aachen University; Dissertation,
RWTH Aachen University, 2023},
abstract = {Gas metal arc welding technology is confronted with a high
demand for quality and competence, one that presently can
only be fulfilled by highly trained but poorly available
specialists. The overarching goal of this work is therefore
to transfer parts of the process competence into the welding
system. The focus is further on the compliance with quality
features of the non-volatile product quality (weld seam
geometry) and the volatile process quality (welding fume
emission). With the acquisition of the transient process and
product quality followed by the closing of the quality
control loop, two research objectives are then specified and
investigated using the introduced conceptof data-based
quality control. The first part of this work considers
suitable sensor technology as well as data pro-cessing to
capture meaningful process features that can be used for
statistical modeling of quality features. Hybrid process
imaging is used to investigate an approach for
simultaneously capturing process features from the joint,
process zone, and weld seam in one sensor system. Here, the
position of the joint can be detected at just 1-2°mm from
the weld pool, thus minimizing the lead time error compared
to conventional sensor systems. Electrical and optical time
series are characterized by high availability yet require a
distinct degree of modeling. With an introduced feature
extraction methodology, time series are made usable, which
is demonstrated using neural networks to identify process
deviations. Building on the previously studied sensor
technology and feature extraction, the second part
demonstrates data-based quality control using two case
studies for the control of fillet weld flanks (product
quality) and welding fume emission (process quality). In the
welding fume emission study, the FER is modeled using
current and voltage time series and reduced by $12-40\%$
over wide power ranges of the standard GMAW process. As part
of the data-based quality control of the fillet weld
geometry, a symmetrical flank ratio is controlled in the
PA,PB and PC welding positions using process images. Based
upon these case studies, the concept of data-based quality
control is further developed and ultimately provides a
methodological basis for the acquisition and control of
further quality features.},
cin = {417610 / 080067},
ddc = {620},
cid = {$I:(DE-82)417610_20140620$ / $I:(DE-82)080067_20181221$},
pnm = {DFG project 390621612 - EXC 2023: Internet of Production
(IoP) (390621612) / WS-B2.I - Connected Job Shop
(X080067-WS-B2.I)},
pid = {G:(GEPRIS)390621612 / G:(DE-82)X080067-WS-B2.I},
typ = {PUB:(DE-HGF)11 / PUB:(DE-HGF)3},
doi = {10.18154/RWTH-2024-04642},
url = {https://publications.rwth-aachen.de/record/985472},
}
guest ::