摘要
根据离散伴随方法理论和自动微分技术,由流场求解器源代码构造了相应的离散伴随场求解器.通过耦合参数化程序、网格生成程序、流场求解器和伴随场求解器,建立了适用于叶轮机械叶栅的气动优化设计系统.利用该系统,对某二维跨音速透平叶栅在给定叶型表面目标压力分布的情况下,通过构造合适的目标函数将叶栅反设计问题转化为叶栅气动优化设计问题,成功进行了气动反设计.反设计结果表明,文中建立的叶栅气动优化设计系统能够根据给定的叶型表面压力分布有效地进行压力反设计,验证了该方法的正确性与有效性.
A discrete adjoint solver is constructed according to the discrete adjoint theory and the automatic differentiation technology based on an in-house flow solver code. An aerodynamic optimization design system for turbomachinery cascades is established by coupling a parameterization program, a structured multi-block grid generator, a flow solver and the discrete adjoint solver established in this work. Based on the optimization system, the pressure inverse design problem could be translated to an optimization problem by constructing the objective function appropriately. An inverse design for a typical 2D transonic turbomachinery cascade is preformed under the inviscid flow environment according to the target pressure distribution specified beforehand. The inverse design result indicates that the optimization system established in this article could implement the pressure inverse design efficiently. The validity and efficiency of the optimization design system are proved.
出处
《江西理工大学学报》
CAS
2012年第3期47-50,共4页
Journal of Jiangxi University of Science and Technology
基金
国家自然科学基金资助项目(51076121)
江西理工大学科研基金项目计划(JXXJ11038)
关键词
离散伴随方法
自动微分
透平叶栅
反设计
灵敏度
discrete adjoint method
automatic differentiation
turbomachinery cascade
inverse design
sensitivity