基于叶片载荷分布的离心风机多目标优化设计
Multi-objective Optimization Design of a Centrifugal Fan Based on the Blade Loading Distribution
崔文豪 1冯建军 1朱国俊 1刘博星 1戈振国 1罗兴锜1
作者信息
- 1. 西安理工大学水利水电学院,西安 710048
- 折叠
摘要
为实现离心风机气动性能的综合提升,将叶轮的叶片载荷分布作为设计参数,以风机全压效率、静压效率及全压作为优化目标建立优化设计体系.采用Morris方法分析设计变量与目标函数间的敏感性关系,综合计算流体力学和径向基神经网络建立设计变量与优化目标间的响应关系,结合遗传算法对离心叶轮进行优化.结果表明:叶片轮盖侧的前缘载荷是风机全压效率的主要影响参数;风机全压对于叶片载荷中间直线段斜率的敏感性较强;风机静压效率主要受到轮盘侧叶片载荷分布中间直线段斜率的影响.优化后风机全压效率、静压效率、全压分别提升了 2％、1.4％、34 Pa;优化后叶片载荷分布呈后加载形式,叶片吸力面靠近轮盖侧的流动分离现象明显改善,尾缘二次泄漏流得到抑制.
Abstract
To realize the comprehensive improvement of the aerodynamic performance of the cen-trifugal fan,the blade loading distribution of the centrifugal impeller is taken as the design parameter,and the total pressure efficiency,static pressure efficiency and total pressure of the fan are taken as the optimization objectives.Morris method was used to analyze the sensitivity relationship between the design variables and the objective function.The response relationship between design variables and optimization objectives is established by combining computational fluid dynamics and radial basis neural network.The centrifugal fan impeller is optimized and improved by genetic algorithm.The results show that the leading edge load on shroud is the main influencing parameter of the fan total pressure efficiency.The total pressure is more sensitive to the slope of the middle straight line of the blade loading.The static pressure efficiency of the fan is mainly affected by the slope of the middle straight line of the blade loading distribution on hub.After optimization,the total pressure efficiency,static pressure efficiency and total pressure are increased by 2％,1.4％and 34 Pa,respectively.For the optimized blade loading distribution,the main loading point on both hub and shroud locates at the rear of its curve.The flow separation on the blade suction side near shroud is significantly reduced,and the secondary leakage flow at the trailing edge is suppressed.
关键词
离心风机/叶片载荷分布/Morris敏感性分析/径向基神经网络/多目标优化Key words
centrifugal fan/blade loading distribution/Morris sensitivity analysis/radial basis neural network/multi-objective optimization引用本文复制引用
出版年
2024
NETL
NSTL
万方数据