【作者】 张宇；

【导师】 陈长征；

【作者基本信息】 沈阳工业大学 ， 机械设计及理论， 2013， 博士

【摘要】 研究叶片的气动性能对风力机叶片的设计意义重大。本文以风力机叶片为研究对象，首先用三维建模软件建立叶片的模型，在风机额定工况下对叶片进行了动态特性分析和模态分析，求解了叶片的前七阶振型和固有频率，这为实际设计和制造风机叶片提供了参考。为比较升力系数和阻力系数在翼型结冰和无冰两种状态下的变化趋势，文中采用CFD计算方法对叶片的气动性能进行了仿真，分析结果得知，升力系数受结冰影响较大，相比之下，阻力系数所受影响有限；雷诺数对升阻力系数影响也有限；叶片结冰对流场影响很大，叶片压力突变，尾流涡旋提前发生，这是结冰改变翼型外形，导致流场不对称所致。通过本文研究，为叶片设计者提供了理论参考，同时，也提供了风机叶片的分析的新思路。其次，介绍了基于模糊神经网络的风力机叶片故障识别方法，该方法根据风力机叶片振动频谱图的各个频段范围，对叶片是否发生故障进行诊断。实验结果表明，该方法可以对风力机叶片结冰、裂纹、点蚀、磨损等故障进行有效地识别，识别准确率均达到90%以上。最后，提出了风力机叶片的优化设计模型，该模型以动量—叶素理论为基础，根据作用在风力机叶片上的气动力，以风力机叶片每段的年能量输出最大为设计目标，使用遗传算法进行寻优搜索。利用开发的优化设计程序，针对特定风场设计截面的最佳弦长和扭角。仿真结果表明，采用改进的遗传算法在对风力机叶片在扭角值基本吻合的前提下，各种气动性能均较为理想。更多还原

【Abstract】 The study of aerodynamic performance of wind turbine blades has the majorsignificance for the design of wind turbine blades. For the model of1.5MW wind turbineit is established by applying the first three-dimensional modeling software. Under the fanrated conditions, the blades of dynamic characteristics and modal are analyzed theseventh-order vibration and natural frequency are also solved, which provides a referencefor the actual design and manufacture of wind turbine blades.In order to obtain the comparing results of the lift coefficient and drag coefficientunder the condition of airfoil and icing airfoil, with the way of CFD calculations, thesimulation of aerodynamic performance is given. From the results, it is obtained that thelift coefficient is greatly impacted by icing airfoil; however, the drag coefficient has thelimited impact. When the values of Reynolds number are different, there is also has limitedimpact on the lift coefficients and drag coefficients. Convection field is greatly affected byicing airfoil, which leads to pressure mutations and wake vortex occurring in advance. Thekey reason for the phenomenon is the asymmetric of the flow field. Through this study, atheoretical reference and the analysis of new ideas are provided for blade designers.Secondly, this paper presents fault recognition method of wind turbine blade based onfuzzy neural network. This method carries out fault diagnosis according to vibrationspectrum diagram in each frequency band range of the wind turbine blade. Experimentalresults show that the method can recognize various faults of wind turbine blade such as icy,crack, wear and tear effectively, which has an accuracy rate above90%.Finally, a model for optimization design of wind turbine blades is presented. Thismodel bases on momentum-leaf pigment theory. Considering the aerodynamic forces ofwind turbine blades, optimization search is carried out by using of genetic algorithm with agoal of the annual energy output in every section of the wind turbine blades. The bestchord length and torsion angle of specific section wind field is designed by using thedeveloped program of the optimum design, Simulation results show that, with theoptimization of improved genetic algorithm, aerodynamic performance of wind turbine blade can be obviously i mproved.更多还原