叶栅叶型正反设计的伴随优化方法

Adjoint Optimization for Direct and Inverse Design of Cascade Blade Profiles

针对基于流场Navier-Stokes方程求解的叶栅叶型正反问题设计优化中,计算量随设计变量数目急剧增加的问题,采用伴随方法建立了集叶片几何参数化、网格生成、流场求解、伴随场求解与优化求解于一体的优化求解方法。针对反问题设计中目标控制参数分布难以给定的问题,通过分析叶栅叶型正问题优化求解结果,给出了反问题优化求解所需的较优目标压力分布。利用自主程序完成了以减弱或消除流动分离、提高叶栅气动性能为目的的叶栅叶型正、反问题自动优化求解,优化后的叶型叶栅总压损失系数分别降低了5.69%、4.50%。研究表明,叶栅叶型吸力面曲率的减小、叶片前加载和中后部逆压梯度的减小,可有效抑制叶片尾缘附近的流动分离。该研究工作对发展高效宽工况叶栅叶型设计技术具有一定的参考价值。

To overcome the problem that the calculation cost drastically increases with the design variable number in Navier-Stokes equations based design optimization for direct and inverse design of cascade blade profiles,an optimization system,including blade geometry parameterization,grid generation,flow solving,adjoint field solving and optimization,is established by means of the adjoint method.In view of the problem that the target control parameter distribution in the design of inverse problem is difficult to be defined,the target pressure distribution of the present inverse design is given by analyzing the optimization result of the direct design.An in-house code is used to solve both direct and inverse problems of the cascade blade profile design for the purposes of reducing or eliminating the flow separation and improving the aerodynamic performance.After optimization,the total pressure loss coefficients of the cascades for the direct and inverse problems are reduced by 5.69% and 4.50%,respectively.The results show that the curvature of the blade suction surface profile is reduced,the blade is front-loaded and the pressure gradient after the blade middle chord is reduced,which effectively inhibits the flow separation near the blade trailing edge.This work provides a reference for the design of cascade blade profiles with high efficiency.