基于多岛遗传算法的大流量离心泵水力性能优化研究  

Optimization Design of Centrifugal Pump based on MIGA

作　　者：张广 冯雪萍[3] 曾庚运 吴喜东 ZHANG Guang;FENG Xueping;ZENG Gengyun;WU Xidong(State Key Laboratory of Hydropower Equipment,Harbin 150040,China;Harbin Electric Machinery Company Limited,Harbin 150040,China;State Grid Gansu Liujiaxia Hydropower Plant,Yongjing 731600,China;Guangdong Hydropower Planning&Design Institute Co.,Ltd.,Guangzhou 510635,China)

机构地区：[1]水力发电设备全国重点实验室,哈尔滨150040 [2]哈尔滨电机厂有限责任公司,哈尔滨150040 [3]国网甘肃刘家峡水电厂,甘肃永靖731600 [4]广东省水利电力勘测设计研究院有限公司,广州510635

出　　处：《大电机技术》2024年第4期111-117,共7页Large Electric Machine and Hydraulic Turbine

基　　金：珠江三角洲水资源配置工程课题研究(ZSJ-XMKT-2020-0026)。

摘　　要：离心泵的效率和汽蚀性能是关系到离心泵能效和稳定性的关键指标,本文以Isight多学科优化平台为基础,将参数化建模、网格划分、数值计算、优化分析等水力设计流程有机结合,构建了离心泵自动优化设计平台。以效率指标和汽蚀余量为目标函数,采用多岛遗传算法对比转速为202m, m^(3)/s的大流量离心泵进行了多目标性能优化。结果表明:多岛遗传算法能够有效提升大流量离心泵水力性能,其主要表现为离心泵设计效率提高0.4%,小流量工况效率整体提高2%~3%,设计工况汽蚀余量降低3m。Pump efficiency and cavitation performance are key indicators related to pump energy characteristic and stability.Based on the Isight multidisciplinary optimization platform,an automatic optimization design platform for centrifugal pumps was constructed by organically combing hydraulic design processes such as parametric modeling,meshing,numerical calculation,optimization analysis and so on.The multi-objective performance of a large flow centrifugal pump with the specific speed of 202m,m^(3)/s is optimized by the multi-island genetic algorithm,which takes the efficiency index and cavitation allowance as objective functions.The results show that the multi-island genetic algorithm can effectively improve the hydraulic performance of the large-flow centrifugal pump.The main performance is that the best efficiency optimization of the centrifugal pump increases by 0.4%,the whole efficiency under the condition of low flowrate is improved within the range of 2%~3%,and the cavitation margin get about 3m reducement under the design conditions.

关 键 词：多岛遗传算法 Isight平台 离心泵 优化设计 数值仿真 

分 类 号：TH311[机械工程—机械制造及自动化]