Refine
Year of publication
- 2010 (342) (remove)
Institute
- Fachbereich Medizintechnik und Technomathematik (72)
- Fachbereich Wirtschaftswissenschaften (41)
- IfB - Institut für Bioengineering (40)
- Fachbereich Elektrotechnik und Informationstechnik (37)
- Fachbereich Energietechnik (35)
- Fachbereich Chemie und Biotechnologie (34)
- Fachbereich Maschinenbau und Mechatronik (31)
- Fachbereich Bauingenieurwesen (24)
- Fachbereich Luft- und Raumfahrttechnik (24)
- INB - Institut für Nano- und Biotechnologien (24)
Language
- English (185)
- German (155)
- Italian (1)
- Multiple languages (1)
Document Type
- Article (178)
- Conference Proceeding (73)
- Book (38)
- Part of a Book (20)
- Conference: Meeting Abstract (12)
- Patent (7)
- Report (5)
- Other (2)
- Part of a Periodical (2)
- Contribution to a Periodical (1)
Keywords
Recently, we introduced and mathematically analysed a new method for grid deformation (Grajewski et al., 2009) [15] we call basic deformation method (BDM) here. It generalises the method proposed by Liao et al. (Bochev et al., 1996; Cai et al., 2004; Liao and Anderson, 1992) [4], [6], [20]. In this article, we employ the BDM as core of a new multilevel deformation method (MDM) which leads to vast improvements regarding robustness, accuracy and speed. We achieve this by splitting up the deformation process in a sequence of easier subproblems and by exploiting grid hierarchy. Being of optimal asymptotic complexity, we experience speed-ups up to a factor of 15 in our test cases compared to the BDM. This gives our MDM the potential for tackling large grids and time-dependent problems, where possibly the grid must be dynamically deformed once per time step according to the user's needs. Moreover, we elaborate on implementational aspects, in particular efficient grid searching, which is a key ingredient of the BDM.
During the development process of a complex technical product, one widely used and important technique is accelerated testing where the applied stress on a component is chosen to exceed the reference stress, i.e. the stress encountered in field operation, in order to reduce the time to failure. For that, the reference stress has to be known. Since a complex technical product may fail regarding numerous failure modes, stress in general is highly dimensional rather than scalar. In addition, customers use their products individually, i.e. field operation should be described by a distribution rather than by one scalar stress value. In this paper, a way to span the customer usage space is shown. It allows the identification of worst case reference stress profiles in significantly reduced dimensions with minimal loss of information. The application example shows that even for a complex product like a combustion engine, stress information can be compressed significantly. With low measurement effort it turned out that only three reference stress cycles were sufficient to cover a broad range of customer stress variety.