Session Details
MS12-1: Multidisciplinary and Structural Design Optimization (Ganzes Minisymposium anzeigen)
Thursday, 12. October 2017; 16:00 - 18:00 Uhr in Raum 7.12
Sitzungsleitung: Erich Wehrle
16:00
Industrial Design Optimization: From Research to Industry Application (Keynote)
Markus Schatz (Airbus DS GmbH)
Kurzfassung:
Time effective product development is crucial to success. To obtain viable compromises in aerospace applications, several arrangements with different disciplines have to be made. Those determine the frequency of numerical studies, wherefore time is essential. Starting from a research perspective, general design optimization will be presented. Thereafter, multiple industrial examples will be given, where invested effort returns in performance, robustness and fewer arrangements.
Industrial Design Optimization: From Research to Industry Application (Keynote)
Markus Schatz (Airbus DS GmbH)
Kurzfassung:
Time effective product development is crucial to success. To obtain viable compromises in aerospace applications, several arrangements with different disciplines have to be made. Those determine the frequency of numerical studies, wherefore time is essential. Starting from a research perspective, general design optimization will be presented. Thereafter, multiple industrial examples will be given, where invested effort returns in performance, robustness and fewer arrangements.
16:40
Regenerating CAD Models with OpenCASCADE and pythonOCC from Numerical Models with Application to Shape Optimization
Altug Emiroglu (Technical University of Munich), Andreas Apostolatos (Technical University of Munich), Roland Wüchner (Technical University of Munich), Kai-Uwe Bletzinger (Technical University of Munich)
Kurzfassung:
Sensitivity filtering methods are unavoidable when numerical shape optimization is considered. A mortar based sensitivity filtering method which incorporates underlying CAD parametrization of the numerical models is proposed. The method is combined with a software environment which utilizes the capabilities of the opensource library OpenCASCADE and pythonOCC module respectively for the regeneration of the CAD models directly from the optimized numerical models as a result of the procedure.
Regenerating CAD Models with OpenCASCADE and pythonOCC from Numerical Models with Application to Shape Optimization
Altug Emiroglu (Technical University of Munich), Andreas Apostolatos (Technical University of Munich), Roland Wüchner (Technical University of Munich), Kai-Uwe Bletzinger (Technical University of Munich)
Kurzfassung:
Sensitivity filtering methods are unavoidable when numerical shape optimization is considered. A mortar based sensitivity filtering method which incorporates underlying CAD parametrization of the numerical models is proposed. The method is combined with a software environment which utilizes the capabilities of the opensource library OpenCASCADE and pythonOCC module respectively for the regeneration of the CAD models directly from the optimized numerical models as a result of the procedure.
17:00
Variational sensitivity analysis in the scope of multiscale problems
Wojciech Kijanski (TU Dortmund University), Franz-Joseph Barthold (TU Dortmund University)
Kurzfassung:
The combination of methods for shape optimisation with different established approaches for analysis and simulation of complex heterogeneous materials on multiple scales based on numerical homogenisation techniques opens a remarkable range of applications introducing design variables, objective functions and constraints on different scales. To design micro-structures, the essential steps for sensitivity analysis on multiple scales will be outlined and accentuated by illustrative examples.
Variational sensitivity analysis in the scope of multiscale problems
Wojciech Kijanski (TU Dortmund University), Franz-Joseph Barthold (TU Dortmund University)
Kurzfassung:
The combination of methods for shape optimisation with different established approaches for analysis and simulation of complex heterogeneous materials on multiple scales based on numerical homogenisation techniques opens a remarkable range of applications introducing design variables, objective functions and constraints on different scales. To design micro-structures, the essential steps for sensitivity analysis on multiple scales will be outlined and accentuated by illustrative examples.
17:20
Kriging-guided Level Set Method for Crash Topology Optimization
Elena Raponi (University of Camerino), Mariusz Bujny (Technical University of Munich), Markus Olhofer (Honda Research Institute Europe GmbH), Nikola Aulig (Honda Research Institute Europe GmbH), Simonetta Boria (University of Camerino), Fabian Duddeck (Technical University of Munich)
Kurzfassung:
Crashworthiness optimization problems are characterized by strong nonlinearities and discontinuities. Hence, gradient-based methods cannot be used and alternative approaches have to be considered. Here, a novel, kriging-based method for level set topology optimization is proposed and validated on a crash test case. Compared to CMA-ES, this method demonstrates to be efficient in terms of convergence speed and promising in the context of crash topology optimization.
Kriging-guided Level Set Method for Crash Topology Optimization
Elena Raponi (University of Camerino), Mariusz Bujny (Technical University of Munich), Markus Olhofer (Honda Research Institute Europe GmbH), Nikola Aulig (Honda Research Institute Europe GmbH), Simonetta Boria (University of Camerino), Fabian Duddeck (Technical University of Munich)
Kurzfassung:
Crashworthiness optimization problems are characterized by strong nonlinearities and discontinuities. Hence, gradient-based methods cannot be used and alternative approaches have to be considered. Here, a novel, kriging-based method for level set topology optimization is proposed and validated on a crash test case. Compared to CMA-ES, this method demonstrates to be efficient in terms of convergence speed and promising in the context of crash topology optimization.