Skapa referens, olika format (klipp och klistra)
Harvard
Dorendorf, G. (2017) Variability of nominally identical components and their influence in an assembly - Applications to a Volvo XC90. Göteborg : Chalmers University of Technology (Diploma work - Department of Applied Mechanics, Chalmers University of Technology, Göteborg, Sweden, nr: 2017:20).
BibTeX
@mastersthesis{
Dorendorf2017,
author={Dorendorf, Gesa},
title={Variability of nominally identical components and their influence in an assembly - Applications to a Volvo XC90},
abstract={Mechanical engineering is important in many engineering industries, for instance the automotive industry. It
allows the developement of improved and more complex products. The automotive industry possesses one of the
biggest markets worldwide and is affected by a high level of competition. Thousands of nominally identical cars
are manufactured everyday. Due to production tolerances, variable properties as well as other uncertainties, a
spread of properties is observed for nominally identical products. To performe time and cost efficient analyses of
engineering products, computer aided engineering (CAE) is increasingly used. The finite element (FE) method
allows fast analyses and modifications of large, complex structures, prior to manufacturing the product. A
central issue with FE simulations is the verification of the model, i.e. that the FE model predicts the real
model behavior with proper accuracy. Thus, for the verification process, a data set apart from the FE model is
required, which is usually obtained by an experimental modal analysis (EMA).
The EMA of nominally identical components can help to identify uncertainties and variations in these
components. These uncertainties and variations can be used to improve the accuracy of corresponding FE
models due to implementation of variations. Otherwise, they indicate also an issue for FE models because
EMAs are not performed for all components and thus the upper and lower boundaries of the uncertainties and
variations are not completely known. This thesis comprises studies in this field of interest, i.e. the quantification
of the variability influence of nominally identical components in an assembly. Two rear subframes of a Volvo
XC90 (2015) car are assembled to the same body in white (BIW) and analyzed experimentally as well as with
CAE.
The FE models are validated for a defined frequency range. A sensor placement for the BIW is evaluated
based on the method of effective independence. Furthermore, an experimental test setup is developed. An
FE model updating with respect to a data set of vibration measurements is performed for one rear subframe.
This includes the updating of stiffness and thickness parameters in the FE model by minimzing a frequency
based deviation metric using equalized damping. During the calibration procedure, the bootstrapping method is
used to estimate statistical informations, i.e. uncertainties of the calibration parameters. With the use of an additional
calibrated rear subframe, the spread in properties for the nominally identical rear subframes is analyzed.
Furthermore, an assembly of the BIW and the two subframes is analyzed numerically as well as experimentally.
The FE simulation uses a substructuring approach using the component mode synthesis method,
i.e. the Craig-Bampton method. In the EMA, the subframes are mounted as suggested by the Volvo Car
Corporation, and the dynamic responses of the two assemblies are analyzed with the main focus on the response
measured on the BIW. Results of the FE analysis are compared with results of the EMA and thus differences
between the two assemblies of a BIW and two nominally identical rear subframes analyzed.
The studies denote the rubber bushing stiffnesses in the interface of the subframes to the BIW as the
parameters which cause the largest differences between the two assemblies. Other properties such as plate
thicknesses and Young’s moduli of steel parts are found to cause less differences. Further, differences are related
to the fixed-interface modes of the individual subframes.
Keywords: Experimental modal analysis, model updating, structural dynamics, component mode synthesis,
Craig-Bampton, uncertainty quantification, effects in assemblies, system identification, automotive industry
},
publisher={Institutionen för tillämpad mekanik, Dynamik, Chalmers tekniska högskola},
place={Göteborg},
year={2017},
series={Diploma work - Department of Applied Mechanics, Chalmers University of Technology, Göteborg, Sweden, no: 2017:20},
}
RefWorks
RT Generic
SR Electronic
ID 250365
A1 Dorendorf, Gesa
T1 Variability of nominally identical components and their influence in an assembly - Applications to a Volvo XC90
YR 2017
AB Mechanical engineering is important in many engineering industries, for instance the automotive industry. It
allows the developement of improved and more complex products. The automotive industry possesses one of the
biggest markets worldwide and is affected by a high level of competition. Thousands of nominally identical cars
are manufactured everyday. Due to production tolerances, variable properties as well as other uncertainties, a
spread of properties is observed for nominally identical products. To performe time and cost efficient analyses of
engineering products, computer aided engineering (CAE) is increasingly used. The finite element (FE) method
allows fast analyses and modifications of large, complex structures, prior to manufacturing the product. A
central issue with FE simulations is the verification of the model, i.e. that the FE model predicts the real
model behavior with proper accuracy. Thus, for the verification process, a data set apart from the FE model is
required, which is usually obtained by an experimental modal analysis (EMA).
The EMA of nominally identical components can help to identify uncertainties and variations in these
components. These uncertainties and variations can be used to improve the accuracy of corresponding FE
models due to implementation of variations. Otherwise, they indicate also an issue for FE models because
EMAs are not performed for all components and thus the upper and lower boundaries of the uncertainties and
variations are not completely known. This thesis comprises studies in this field of interest, i.e. the quantification
of the variability influence of nominally identical components in an assembly. Two rear subframes of a Volvo
XC90 (2015) car are assembled to the same body in white (BIW) and analyzed experimentally as well as with
CAE.
The FE models are validated for a defined frequency range. A sensor placement for the BIW is evaluated
based on the method of effective independence. Furthermore, an experimental test setup is developed. An
FE model updating with respect to a data set of vibration measurements is performed for one rear subframe.
This includes the updating of stiffness and thickness parameters in the FE model by minimzing a frequency
based deviation metric using equalized damping. During the calibration procedure, the bootstrapping method is
used to estimate statistical informations, i.e. uncertainties of the calibration parameters. With the use of an additional
calibrated rear subframe, the spread in properties for the nominally identical rear subframes is analyzed.
Furthermore, an assembly of the BIW and the two subframes is analyzed numerically as well as experimentally.
The FE simulation uses a substructuring approach using the component mode synthesis method,
i.e. the Craig-Bampton method. In the EMA, the subframes are mounted as suggested by the Volvo Car
Corporation, and the dynamic responses of the two assemblies are analyzed with the main focus on the response
measured on the BIW. Results of the FE analysis are compared with results of the EMA and thus differences
between the two assemblies of a BIW and two nominally identical rear subframes analyzed.
The studies denote the rubber bushing stiffnesses in the interface of the subframes to the BIW as the
parameters which cause the largest differences between the two assemblies. Other properties such as plate
thicknesses and Young’s moduli of steel parts are found to cause less differences. Further, differences are related
to the fixed-interface modes of the individual subframes.
Keywords: Experimental modal analysis, model updating, structural dynamics, component mode synthesis,
Craig-Bampton, uncertainty quantification, effects in assemblies, system identification, automotive industry
PB Institutionen för tillämpad mekanik, Dynamik, Chalmers tekniska högskola,
T3 Diploma work - Department of Applied Mechanics, Chalmers University of Technology, Göteborg, Sweden, no: 2017:20
LA eng
LK http://publications.lib.chalmers.se/records/fulltext/250365/250365.pdf
OL 30
