• CN: 11-2187/TH
  • ISSN: 0577-6686

Journal of Mechanical Engineering ›› 2022, Vol. 58 ›› Issue (12): 283-292.doi: 10.3901/JME.2022.12.283

Previous Articles     Next Articles

Blades Optimization of Pumps as Turbines Based on RBF-HDMR Model and PSO Algorithm

JIANG Bingxiao1, YANG Junhu1,2, WANG Xiaohui1,2, MIAO Senchun1,2   

  1. 1. School of Energy and Power Engineering, Lanzhou University of Technology, Lanzhou 730050;
    2. Key Laboratory of Fluid Machinery and Systems of Gansu Province, Lanzhou 730050
  • Received:2021-09-24 Revised:2022-03-25 Online:2022-06-20 Published:2022-09-14

Abstract: In view of the low efficiency of pumps as turbines(PAT), a hybrid model combining radial basis function(RBF) and high-dimensional model representation(HDMR) is proposed to optimize the performance of PAT. The impeller blade of PAT is taken as the research object. The blade profile is parameterized with Bezier curves, the independent variables of the surrogate model are separated and the training interval of the model is determined. The surrogate model based on blade profile control variables for PAT efficiency prediction is constructed with Matlab, Creo and Fluent software. The particle swarm optimization(PSO) algorithm is utilized to optimize the surrogate model in the training interval by global optimization, the optimal efficiency point of the PAT and the corresponding blade geometry parameters are obtained. The prediction data of the surrogate model is verified through numerical simulation and experimental research. The results show that the performance curves of the numerical simulation of PAT optimized by the surrogate model are in good agreement with the test results. At the optimal operating point, the numerical simulation efficiency value of the optimized PAT is 4.78% higher than that of the prototype PAT, the shaft power is 6.22% higher, the test efficiency value is 3.7% higher than that of the prototype PAT, and the shaft power is 4.8% higher.

Key words: pumps as turbines, blade optimization, radial basis function, high-dimensional model representation, particle swarm optimization

CLC Number: