多级冷却涡轮通流伴随优化理论与设计方法研究

The S2 surface direct problem analysisfor multi-stage cooling turbine plays fundamental and crucial role in the aerodynamic design of turbine and the turbine performance can bedirectly improved via S2 surface through-flow optimization design. The conventional optimization algorithms have limited application in the aerodynamic design of turbine due to the high computing cost. Whereas, the adjoint method can eliminate the dependence of the gradient of the objective function on the flow variables, which makes the optimization period of the adjoint mehtod independent of the number of the design variables. Therefore, for fast optimization of multi-stage turbine with a large number of design variables, the adjoint method has a wider application than the traditional algorithms. So far, the adjoint mehod has been applied in the S1 surface and 3D simulation. The applicants have done some preliminary work on the adjoint optimizationbased on the S2 surface direct problem. This project aims to further previous work, and has three objectives: the first one is to build the adjoint equation of S2 surface direct problem and improve the optimization ability. The second objective is to conduct the research on the adjoint optimization with the cooling air injection and the goal is to solve the cooling air injection issue of the adjoint flow field, achieving the optimization of turbine in aerodynamics and mixing of cooling air and main gas flow. Finally we aim to study the effect of cooling air injection on the main gas flow under the non-design point and enhance the performance of the multi-stage turbine with the influence of the mixing of cooling air and main gas flow under the multi-working condition.

多级冷却涡轮S2通流正问题分析是涡轮气动设计基础与核心，基于S2通流优化设计可直接提升涡轮气动性能。传统优化方法因其时间成本问题较难在涡轮优化设计中推广应用，而伴随方法因其消除了目标函数梯度对流场参量的依赖，使得计算量与设计变量数目无关，适用于多变量、多工况涡轮性能快速优化。迄今叶轮机内流伴随方法研究均落在S1流面与三维叶片上，而基于三维CFD的冷却涡轮伴随优化因冷却孔巨量网格问题难以实用。本课题拟基于S2流面正问题开展伴随优化理论与设计方法研究：一是建立S2流面正问题的通用伴随方程，提高多自由度下优化设计能力；二是研究冷气掺混下的伴随耦合优化方法，解决伴随场的冷气入射问题，实现多级涡轮气动和冷气掺混的耦合优化；三是研究非设计工况下冷气入射对主流燃气流动的影响，提高多级涡轮多工况、冷气掺混影响下的整体性能。

多级冷却涡轮S2通流正问题分析是涡轮气动设计基础与核心，基于S2通流优化设计可直接提升涡轮气动性能。课题研究针对带冷气的多级涡轮，开展了基于S2流面正问题的伴随优化理论与设计方法研究。主要研究内容：（1）推导并编程实现了曲线坐标系下涡轮S2流面正问题欧拉方程求解;(2) 开发了涡轮S2流面正问题伴随方法气动优化设计系统；（3）四级低压涡轮和涡轮过渡段算例验证了气动优化设计系统的有效性与实用性；（4）开发了考虑冷气掺混的涡轮气动/冷气掺混耦合优化设计系统；（5）以带冷气单级高压涡轮和四级低压涡轮算例验证了开发系统的有效性与实用性。研究解决了考虑叶片/流道几何多自由度下的多级涡轮优化模型问题，以及多级涡轮气动/冷气掺混伴随方程冷气源项模化问题。. 鉴于伴随优化方法的局域性缺憾以及叶轮机优化变量多、工程应用困难，课题研究开发了叶轮机全局优化方法与几何降维技术。主要研究内容：（6）提出改进人工蜂群算法(IABC）的全局优化方法，编程实现并验证了方法的有效性；（7）构建了基于Isight、IABC及并行超算的叶轮机全局优化系统；（8）提出了基于Bezier曲面特性的叶轮机曲面参数化降维方法，提出了三维叶片的半曲面、全曲面降维技术；（9）四级压气机、两级涡轮算例验证了IABC、曲面参数化方法的工程实用价值。研究解决了多级叶轮机气动优化的多极值黑箱特性问题，以及传统二维叶型加积叠线优化参数化方法控制变量多、高维特性显著问题。. 课题研究获批软件著作权：4项；发表论文：30篇，其中SCI Q1区1篇、Q3、Q4区2篇，EI收录11篇，国内会议12篇、国际会议9篇；培养博士生3人，毕业2人，培养毕业硕士生7人。. 研究获取的优化方法、降维技术及构建的优化平台，可直接应用于多级轴流压气机、涡轮的气动优化设计工程实际中，有效促进了“两机”专项叶轮机气动设计的优化能力提升与技术进步。