Abstract: Aiming at the problem of parametric modeling and profile reconstruction in the intelligent optimization design of steam turbine blades， the coordinate equation of blade profile was derived based on the cubic Bezier curve， and the parametric expression of turbine blade profile was realized according to the basic geometric parameters of blade. On the basis of the parameterized model， according to the known blade coordinate point datas， the actual turbine blade profile was reconstructed by using the seeker optimization algorithm （SOA）， and the comparison between SOA and PSO （Particle Swarm Optimization） was carried out through the example of blade profile reconstruction. The results show that the parameterized profile based on Bezier curve has good smoothness and high quality， the proposed method can arbitrarily modify the geometric parameters of the blade and regenerate the model， which is more practical and efficient. The blade profile reconstruction method based on SOA optimization algorithm can accurately reconstruct the blade profile of Dykas steam turbine， and the convergence speed and convergence result of the method based on SOA are faster and more stable. The model and the algorithm lay a solid foundation for the design optimization of turbine blades.

Key words: Steam turbine blade;Spontaneous condensation;Parameterized;Seeker optimization algorithm;Bezier curve;Reconstruction of blade profile

摘要： 针对汽轮机叶片智能优化设计中的叶片参数化建模与型线重构问题，首先基于三次Bezier曲线推导了叶片型线的坐标方程，根据叶片基本几何参数实现了汽轮机叶片型线的参数化表达。在构建的参数化模型基础上，根据已知的叶型坐标点数据，采用人群搜索算法（SOA）重构实际汽轮机叶片型线，并通过叶型重构算例对SOA和粒子群算法（PSO）进行了对比分析。结果表明：基于Bezier曲线构建的参数化型线光滑性好、质量高，所提方法可以任意修改叶型几何参数并再生模型，实用性较强且效率较高。基于SOA优化算法的叶片型线重构方法可以准确重构Dykas汽轮机叶片型线，且SOA重构叶型的方法收敛速度快、收敛结果更稳定。

关键词: 汽轮机叶片;自发凝结;参数化;人群搜索算法;Bezier曲线;叶型重构