摘要
为了提高优化设计效率,节约成本,将试验设计方法、响应面近似模型与遗传算法优化方法结合在一起应用于跨音速转子叶片气动优化设计中.首先采用拉丁超立方试验设计方法生成样本点构建目标函数的响应面模型,然后利用遗传算法对响应面模型进行目标寻优,并对优化解进行流场分析验证.为了验证方法的可行性,以NASA rotor37转子叶片前缘积迭线弯与掠优化设计为例,总压损失系数为目标函数进行单目标优化设计,优化后总压损失系数降低了0.528%,总耗时20h,算例结果表明基于响应面近似模型的优化设计方法可用于叶片气动优化中,提高了优化设计效率.
In order to increase the efficiency of the optimization design and reduce the computation, a method was applied to the aerodynamic optimization design of the rotor blade of the transonic compressor,which integrated the experiment of design method and response surface model with the genetic algorithm. First,the samples were generated by the Latin hypercuhe sampling （LHS）. Then the response surface model was constructed by the samples, which was applied to replace the CAE software. Finally, the rotor blade was optimized by the genetic algorithm and the solution of the optimal design was verified by the software of the numerical computation. In order to prove the feasibility of the method, the method was applied to the stacking lihe of NASA Rotor37, the objective function was the loss coefficient of the total pressure. The result showed that the loss coefficient of the total pressure decreased 0. 528%, the time was 28 hours.
出处
《西安工业大学学报》
CAS
2013年第2期114-117,150,共5页
Journal of Xi’an Technological University
基金
863国家自然科学基金资助项目(2007AA04Z184)
关键词
跨音速转子叶片
试验设计
响应面模型
气动优化设计
rotor blade of transonic compressor
experiment of design method
response surface model
aerodynamic optimization design
