混流泵叶轮的三维反问题设计多目标优化  

Multi-objective optimization of mixed-flow pump impeller based on 3-D inverse design

作　　者：王梦成 袁建平[1] 李彦军[1] 郑云浩 WANG Mengcheng;YUAN Jianping;LI Yanjun;ZHENG Yunhao(National Research Center of Pumps, Jiangsu University, Zhenjiang 212013, China)

机构地区：[1]江苏大学国家水泵及系统工程技术研究中心,江苏镇江212013

出　　处：《哈尔滨工程大学学报》2020年第12期1854-1860,共7页Journal of Harbin Engineering University

基　　金：国家重点研发计划(2018YFB0606103).

摘　　要：为进一步提高混流泵运行效率及高效区范围,本文对比转速为530的混流泵进行了综合优化设计。采用由反问题设计法、数值模拟、拉丁超立方抽样法、克里金模型和非支配遗传算法构成的综合优化系统,以叶片载荷为设计参数,以0.7Qdes和1.1Qdes处的泵段效率为优化目标。在累计探索10000种不同叶轮配置后,得到了最优配置集构建的帕累托前沿,并根据工程需求选取了适当的参数配置。结果表明:与原始混流泵模型相比,优化后模型在较大的流量范围内(0.7Qdes^1.3Qdes)泵段效率有明显提升。在设计点处,扬程基本不变,效率提高约4.2%。该方法具有良好的工程应用价值。To further improve the mixed-flow pump operating efficiency and the range of the high efficiency, the mixed-flow pump with a specific speed of 530 was optimized. A comprehensive optimization system consisting of the inverse problem design method, numerical simulation, Latin hypercube sampling method, Kriging model, and non-dominated genetic algorithm was adopted, with the blade load as the design parameter and the pump section efficiency at 0.7Qdes and 1.1Qdes as the optimization objective. After 10 000 different impeller configurations were explored, the Pareto front of the optimal configuration set was obtained, and the appropriate parameter configurations were selected according to engineering requirements. The pump efficiency of the optimized model was significantly improved in a larger flow range(0.7Qdes^1.3Qdes). At the design point, the head was basically unchanged, and the efficiency was increased by about 4.2%. The research results show that this method has good engineering application value.

关 键 词：导叶式混流泵 叶轮 三维反问题设计 数值模拟 多目标优化 拉丁超立方取样 克里金模型 非支配遗传算法 

分 类 号：TH313[机械工程—机械制造及自动化]