• 论文 •    

基于Kriging模型的涡轮盘优化设计方法

黄章俊,王成恩   

  1. 1.东北大学 辽宁省复杂装备多学科设计优化技术重点实验室,辽宁沈阳110004;2.东北大学 流程工业综合自动化教育部重点实验室,辽宁沈阳110004
  • 出版日期:2010-05-15 发布日期:2010-05-25

Turbine discs optimization design based on Kriging model

HUANG Zhang-jun, WANG Cheng-en   

  1. 1.Liaoning Provincial Key Laboratory of Multidisciplinary Optimal Design for Complex Equipment of Northeastern University, Shenyang 110004, China;2.Ministry of Education Key Laboratory of Process Industry Automation,Northeastern University,Shenyang 110004,China
  • Online:2010-05-15 Published:2010-05-25

摘要: 研究了航空发动机涡轮盘设计问题,将Kriging方法与全局约束优化算法相结合,提出了基于Kriging近似响应模型的优化求解方法,有效减少了有限元分析的次数和计算量。该优化设计方法采用正交试验设计和有限元分析方法获得涡轮盘设计空间的初始样本数据;利用已有的设计样本数据建立涡轮盘重量和应力响应近似计算的Kriging模型,在此基础上利用全局优化算法在设计空间进行设计方案优化搜索;利用有限元分析方法计算近似最优设计方案的输出响应,并以此更新已有的设计样本数据,不断提高Kriging模型的近似精度。数值计算结果表明,基于Kriging模型的涡轮盘优化设计方法不仅可以获得良好的设计结果,还能大大降低计算代价,提高设计效率。

关键词: 涡轮盘, 优化设计, Kriging模型, 有限元方法

Abstract: Aero-engine turbine disc design problems were studied and an optimization design approach based on Kriging approximate models by combining Kriging method and a global constrained optimization algorithm was put forward, which significantly reduced the number of finite element analyses and computations. Initial sampling data were obtained with the orthogonal experiment design and finite element method. The Kriging model for calculating the approximate values of turbine disc weight and stresses was constructed by using existing design data, which was invoked iteratively by the global optimization algorithm during search for an approximate optimum through the design space. Subsequently, the approximate optimum as a result of search based on the Kriging model was analyzed using finite element method, and the computation outputs were utilized to update the design dataset so as to improve the accuracy of the Kriging model. Numerical computation demonstrated that the presented optimization design approach based on Kriging models could obtain satisfactory solutions while considerably decreased computation cost and increased design efficiency as well.

Key words: turbine disc, optimization design, Kriging model, finite element method

中图分类号: