基于神经网络模型和CFD的轴流泵自动优化

发布时间：2018-01-20 05:29

本文关键词： 轴流泵程序集成最优拉丁超立方设计优化设计径向基神经网络　出处：《排灌机械工程学报》2017年06期 　论文类型：期刊论文

【摘要】：参数化设计和计算流体力学被广泛应用于流体机械的优化设计.采用旋转机械设计软件CFturbo对轴流泵进行水力设计.为缩短优化周期,基于Isight多学科优化平台,通过编写批处理命令将CFturbo与PumpLinx集成,实现了轴流泵的CFD自动优化.以提高轴流泵的水力效率为优化目标,采用最优拉丁超立方设计对叶轮和导叶的7个设计变量进行空间采样,设计了72组方案.基于PumpLinx的数值模拟结果,建立了目标函数与设计变量之间的径向基神经网络模型,并采用多岛遗传算法对其进行优化,结果表明:数值模拟结果与试验结果吻合较好,且径向基神经网络模型能准确预测轴流泵效率与设计变量的关系.优化后,设计点效率提高了4.46%,而扬程几乎不变.通过Pareto图分析,获得了设计变量对目标影响的显著水平,可为轴流泵的优化设计提供一定的参考.
[Abstract]:Parametric design and computational fluid dynamics (CFD) are widely used in the optimization design of fluid machinery. The hydraulic design of axial flow pump is carried out by the rotating machine design software CFturbo. Based on the Isight multidisciplinary optimization platform, CFturbo and PumpLinx are integrated by writing batch commands. The CFD automatic optimization of axial flow pump is realized. In order to improve the hydraulic efficiency of axial flow pump, the optimal Latin hypercube design is used to sample the seven design variables of impeller and guide vane. 72 groups of schemes are designed. Based on the numerical simulation results of PumpLinx, the radial basis function neural network model between the objective function and the design variables is established, and the multi-island genetic algorithm is used to optimize the model. The results show that the numerical simulation results are in good agreement with the experimental results, and the radial basis function neural network model can accurately predict the relationship between the axial flow pump efficiency and the design variables. After optimization, the design point efficiency is increased by 4.46%. By Pareto diagram analysis, the significant level of the influence of the design variables on the target is obtained, which can provide a certain reference for the optimization design of the axial flow pump.
【作者单位】：江苏大学国家水泵及系统工程技术研究中心;江苏省水利工程科技咨询有限公司;
【基金】：国家科技支撑计划项目(2015BAD20B01) 江苏省水利科技项目(2015042) 江苏高校自然科学研究项目(09KJB570001)
【分类号】：TH312
【正文快照】： 陆荣,袁建平,李彦军,等.基于神经网络模型和CFD的轴流泵自动优化[J].排灌机械工程学报,2017,35(6):481-487.LU Rong,YUAN Jianping,LI Yanjun,et al.Automatic optimization of axial flow pump based on radial basis functions neural net-work and CFD[J].Journal of drain

【相似文献】