中开多级离心泵效率优化计算方法

对离心泵水力效率及高效区相对宽度的优化计算方法进行研究，在水力损失模型的基础上提出基于近似模型的多目标优化计算方法. 以中开多级离心泵的优化设计为例，基于水力损失模型进行设计变量灵敏度分析，选出关键设计变量. 分别利用水力损失、完全二次响应面（RSF）、径向基高斯响应面（RBF）和克里金响应面（KRG）4种近模型优化离心泵的关键设计变量，分析4种效率优化计算方法的精确性和有效性. 结果表明：基于理论公式计算的第1种优化方法耗时少，但结果误差较大；后3种优化方法基于计算流体动力学（CFD）数值仿真分析，结果准确，其中RSF模型的结果最精确且计算时间较短. 比较3种不同近似模型的计算精度，RSF的计算结果最精确，RBF结果次之，KRG结果最差. 在设计流量下，基于RSF的Pareto最优解的扬程为83.77 m，效率为77.26%，基于RBF的Pareto最优解的扬程为83.09 m，效率为76.63%.

Optimization calculation method for efficiency of multistage split case centrifugal pump

The optimization calculation method of the hydraulic efficiency and the relative width of high efficient area for centrifugal pumps was researched. The multi-objective optimization calculation method of the approximate models was proposed based on the hydraulic loss model. Optimization design of multi-stage dual split centrifugal pump was conducted as an example. The key design variables were selected out through sensitivity analysis based on the hydraulic loss model. The hydraulic loss model, the complete quadratic response surface function (RSF) model, the radial basis Gaussian response surface function (RBF) model and the Kriging response surface function (KRG) model were used respectively to optimize the key design variables of centrifugal pumps. The accuracy and efficiency of the four methods were analyzed as well. Results showed that the calculation time of the first optimization method based on the theoretical formula was the shortest, but the error was big. The latter three optimization methods were based on the computational fluid dynamics (CFD) numerical simulation analysis and the results were accurate. The results of RSF model were the most accurate and the calculation time was short. The calculation results of RSF was the most accurate, followed by that of RBF, and the worst was that of KRG by the comparison of the calculation accuracies of the three approximate models. The Pareto optimal solution based on RSF had the head of 83.77 m and the efficiency of 77.26% with the design flow. The Pareto optimal solution based on RBF had the head of 83.09 m and the efficiency of 76.63% with the design flow.