论文部分内容阅读
基于大转折角基元级速度三角形设计概念,使用CFD软件对3种进气马赫数、6种进气角、3种扩压因子、3种叶栅稠度、2种叶型最大厚度比的共计324个平面叶栅流场进行数值计算。为了提高新型叶栅的性能,利用人工神经网络构建基于数据库样本空间的近似函数,采用遗传算法来寻找新的设计,并对新的设计进行气动性能预测。对中弧线和厚度分布进行优化,优化后总压损失系数降低了14.48%。
Based on the concept of triangular design with large turning angle, a total of three kinds of intake Mach number, six kinds of intake angles, three kinds of diffusion factors, three kinds of cascade thicknesses and two kinds of maximum thickness ratio of blades were calculated using CFD software 324 plane cascade flow field numerical calculation. In order to improve the performance of the new type of cascade, the artificial neural network is used to construct the approximate function based on the sample space of the database, the genetic algorithm is used to find the new design and the aerodynamic performance of the new design is predicted. The arc line and thickness distribution are optimized. The optimized total pressure loss coefficient is reduced by 14.48%.