改进离散型遗传算法在离心泵性能优化中的应用

Application of modified discrete genetic algorithm in performance optimization of centrifugal pump

针对离心泵性能优化中设计参数多且优化周期长的问题,提出了适用于离散变量优化的改进遗传算法.以泵效率为优化目标,选取叶轮6个设计参数,采用拉丁超立方设计方法创建90组设计样本,用响应面模型对泵效率和设计参数进行拟合.通过改进的离散型遗传算法对其进行寻优,得到叶轮最优设计参数组合.结果表明:响应面模型能够准确反映效率和叶轮设计参数之间的关系,设计工况下预测值与数值模拟的偏差为0.82%;优化后泵效率在0.6Qd,1.0Qd和1.4Qd下的增幅分别为7.05%,5.98%和1.90%,且输入功率显著降低;优化后叶轮和蜗壳内部的流动损失面积明显减小,整体流动损失显著降低.参数敏感性分析表明:叶片出口安放角、叶轮出口直径、叶轮出口宽度和叶片包角对泵性能的影响较大.

Aiming at the problem of many design parameters and long optimization period in centrifugal pump performance optimization,a modified genetic algorithm applicable to discrete variable optimization was proposed.With pump efficiency as the optimization objective,six design parameters of the impeller were selected,and 90 sets of design samples were created using the Latin hypercube design method.A response surface model was used to fit the pump efficiency and the design parameters,which is optimized by a modified discrete genetic algorithm to obtain the optimal combination of impeller design parameters.The results showed that the response surface model can accurately reflect the relationship between efficiency and impeller design parameters,and the deviation between the predicted value and numerical simulation under design conditions is 0.82%,the optimized pump efficiency increases by 7.05%,5.98%and 1.90%at 0.6Qd,1.0Qd and 1.4Qd.The input power is significantly reduced,the flow loss area inside the impeller and worm gear is significantly reduced after optimization,and the overall flow loss is significantly reduced.Parameter sensitivity analysis showed that blade outlet placement angle,impeller outlet diameter,impeller outlet width,and blade wrap angle have a large impact on pump performance.