基于人工神经网络与遗传算法的风力机翼型优化设计方法  

Optimal Design Method for Wind Turbine Airfoil Based on Artificial Neural Network Model and Genetic Algorithm

作　　者：琚亚平[1] 张楚华[1] 

机构地区：[1]西安交通大学能源与动力工程学院,陕西省西安市710049

出　　处：《中国电机工程学报》2009年第20期106-111,共6页Proceedings of the CSEE

基　　金：国家高技术研究发展计划项目(863计划)(2007AA05Z203);国家自然科学基金项目(10572113);新世纪优秀人才支持计划资助项目(NCET-06-0847)~~

摘　　要：开发风力机专用翼型是风电技术中的一个关键问题,高升阻比翼型的优化设计方法是目前风力机空气动力学的重要研究课题。该文建立了多运行工况下升阻比最高的风力机翼型优化设计方法,运用Bezier函数建立了翼型的数字化参数表征方法,根据完全析因试验设计方法选取了翼型族的设计空间,利用计算流体力学方法获得了每个翼型样本的气动性能参数,采用人工神经网络和遗传算法相结合的现代优化方法数值求解了优化命题。基于上述模型对FX63–167翼型进行优化改进,重点研究风力机翼型在3种运行工况点下的气动优化设计命题及其求解方法。计算表明优化后翼型具有更佳的气动特性,各工况点下的升阻比均有所提高,验证了该优化方法的合理性、可行性。The development of special airfoils for wind turbines is one of the key problems in the wind power technology. The optimal design method for the high lift-drag ratio airfoil is currently the important issue in the area of wind turbine aerodynamics. An optimal design problem and the corresponding numerical method were established for wind turbine airfoils with the maximum lift-drag ratios at multiple working conditions as the optimal objective. Firstly, the airfoil profile was parameterized by Bezier curves and the design space for airfoil family was specified by the full factorial design of experiment method. Then, the aerodynamic parameters for each sample airfoil were calculated by the computational fluid dynamics method. The optimal problem was numerically solved by combining the genetic algorithm with the artificial neural network model. Finally, the above optimal method was applied to the optimal redesign of the FX 63-167 wind turbine airfoil. The results indicate that the optimized airfoil has better aerodynamic performance, whose lift-drag ratio is increased comparing with the original one at all the working conditions under consideration. It eventually confirms the feasibility and validity of the proposed optimal problem and the numerical solution method. This work is of academic and engineering values for guiding the aerodynamic design and analysis of wind turbine blades.

关 键 词：风力机翼型 参数化建模 优化设计 神经网络 遗传算法 

分 类 号：TM614[电气工程—电力系统及自动化]