A Scaling Methodology for Dynamic Systems: Quantification of Approximate Similitude and Use in Multiobjective Design.

Scaling;Multiobjective Design;Dynamic Systems;Similitude;Mechanical Engineering;Engineering;Mechanical Engineering

Kittirungsi, BuritStefanopoulou, Anna ;

University of Michigan

关键词: Scaling;    Multiobjective Design;    Dynamic Systems;    Similitude;    Mechanical Engineering;    Engineering;    Mechanical Engineering;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/58383/burit_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

A design technique that adapts or scales a system design to meet new requirements is developed. This scaling technique is potentially useful because it focuses on retaining existing desirable characteristics (e.g., efficiency, stability) of the original design through minimal modifications. Previous work in the literature explored this notion by developing scaling techniques based on the dynamic similitude principle.However, such similitude-based scaling is often found too restrictive because it may not be feasible to satisfy all of the scaling laws designated by the similitude principle exactly. Moreover, the literature only defines such similitude discretely in terms of whether the scaled design satisfies these scaling laws. This definition then makes it impossible to assess the degree to which two designs are close to satisfy similitude. The work in this dissertation mitigates these difficulties as follows: First, it uses a novel combination of activity-based model reduction and dimensional analysis to assess the relative importance of each scaling law and permit neglecting the least important ones, thereby providing more freedom than strict similitude-based scaling.Next, a metric is developed to cope with the situation in which the most important scaling law(s) cannot be followed due to other conflicting requirements and constraints. This metric allows one to quantify approximate similitude, that is, the degree to which the scaled design is close to satisfying the discrete definition of exact similitude. Then, this quantification is utilized in a multiobjective scaling framework that trades off approximate similitude versus the conflicting requirements and constraints. The applicability of the methodology is demonstrated through three case studies. The first study applies the methodology to a linear quarter-car system to scale the chassis vibrations. The second study scales a fuel cell’s nonlinear air supply system subject to different power requirements. The last case study represents a scaling design study of a complex multi-body dynamic vehicle design to maintain rollover safety properties when subject to extra roof-top loads.These examples demonstrate that the proposed method does provide a systematic, computationally efficient approach to redesign as compared with casting the redesign as an optimization problem.