An efficient strategy for reliability-based multidisciplinary design optimization of twin-web disk with non-probabilistic model |
| |
| Institution: | 1. College of Physical Science and Technology, Research Institute for Biomimetics and Soft Matter, Xiamen University, Xiamen 361005, China;2. School of Mechanics, Civil Engineering and Architecture, Northwestern Polytechnical University, Xi''an 710129, China;3. State Key Laboratory of Physical Chemistry of Solid Surfaces, College of Chemistry and Chemical Engineering, Collaborative Innovation Center of Chemistry for Energy Materials, Xiamen University, Xiamen 361005, China;4. Fujian Provincial Key Laboratory for Soft Functional Materials Research, Jiujiang Research Institute, Xiamen University, Xiamen 361005, China;1. National Key Laboratory of Science and Technology on Aero-Engine & Aero-Thermodynamics, Beihang University, Beijing 100191, China;2. Department of Aircraft Airworthiness Certification, Civil Aviation Management Institute of China, Beijing 100102, China,;1. State Key Laboratory of Digital Manufacturing Equipment and Technology, School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, P. R. China;2. Department of Industrial and Manufacturing Systems Engineering, University of Michigan-Dearborn, Dearborn, MI 48128, USA |
| |
| Abstract: | The twin-web disk holds big promise for increasing efficiency of the aircraft engine. Its reliability-based multidisciplinary design optimization involves several disciplines including fluid mechanics, heat transfer, structural strength, and vibration. The solution to this optimization problem requires three-loop calculations including loops for optimization, reliability, and interdisciplinary consistence often making its computational cost unacceptably high. The lack of sufficient amount of probabilistic data, especially for this brand-new turbine disk, makes matters worse. In this paper, the non-probabilistic uncertain variables are described by an evidence theory-based fuzzy set method, which we extend to general structure of uncertain data. We also propose two modifications of the active learning kriging model: one of them for the purpose of optimization with respect to the distance from the optimum point and another one for the purpose of assessing reliability by introducing the importance concept. Applications of these two modifications are demonstrated in this paper. Finally, a multi-adaptive learning kriging strategy for non-probabilistic reliability-based multidisciplinary design optimization of twin-web disk is proposed to improve its power efficiency and reliability in a computationally effective way. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
