【摘要】 应用流行的神经网络技术与先进的几何描述方法NURBS技术,对某两级风扇进行优化改型设计,特别针对于转子2的根部截面进行优化,通过对优化前后的转子2以及两级风扇的流场气动结构与特性的对比分析表明,该优化设计方法确实可行,具有一定的先进性,能够改善叶轮机械的整体气动特性,可以为叶轮机械的优化设计提供一种新的途径。更多还原

【Abstract】 A two stage fan especially the second rotor’s blade was reshaped to optimize its aero-dynamical performance by applying a variable learning rate momentum BP neural network arithmetic adjoined with NURBS method.The contrastive analysis of aero-dynamical performance between the original blade and the optimized one shown that this optimization method not only was feasible and advanced but also it could improved the flow field structure of turbomachinery.It was certain that this method would provide a new optimization way for turbomachinery blade design.更多还原