高级检索+

基于BP神经网络与NSGA-Ⅱ算法的低比转速离心泵优化

Optimization of Low Specific Speed Centrifugal Pump Based on BP Network and NSGA-Ⅱ Algorithm

  • 摘要: 低比转速离心泵具有流量小、扬程高、效率低的特点。在满足扬程的使用要求条件下,为进一步提高低比转速离心泵的效率,通过修改叶轮子午面优化低比转速离心泵,再选择叶片出口宽度、进口shroud安放角、进口hub安放角、出口安放角和包角5个参数进行优化,以扬程和效率作为共同优化目标,采用最优拉丁超立方设计方法设计40组实验,利用BP神经网络作为优化参数与优化目标之间的代理模型,并运用NSGA-Ⅱ多目标遗传算法进行扬程和效率的寻优。寻优之后通过斜切叶轮出口抑制驼峰现象,通过多参数优化结果表明,离心泵在标准工况下扬程提升了13 m、效率提升了8.07%。

     

    Abstract: The low specific speed centrifugal pump has the characteristics of small flow, high head and low efficiency. In order to improve the efficiency of the low specific speed centrifugal pump under the condition of meeting the use requirements of the head, this paper optimizes the low specific speed centrifugal pump by modifying the meridian surface of the impeller, and then the blade outlet width, the inlet shroud placement angle, the inlet hub placement angle, the outlet placement angle and wrap angle are used as optimization parameters. Head and efficiency are selected as optimization goals. 40 groups of experiments are designed by using the optimal Latin Hypercube Design Method. BP neural network is used as the surrogate model between optimization parameters and optimization goals, and NSGA-Ⅱ multi-objective genetic algorithm is used to optimize the head and efficiency. After optimization, the hump phenomenon is suppressed by chamfering the impeller outlet. The final optimized centrifugal pump has an increase of 13 m in head and an efficiency increase of 8.07% under standard conditions.

     

/

返回文章
返回