Optimization of Guide Vane Airfoil Shape of Pump Turbine Based on SVM-MDMR Model

Abstract

Pumped storage is an important green, low-carbon, and clean flexible regulating power source in the power system, which can provide regulation services for the power system, promote the construction of a new type of power system, and facilitate the green transformation of energy. To improve the efficiency and stability of the centrifugal pump turbine under multiple operating conditions, a surrogate model combining radial basis functions with a high-dimensional model is used for performance optimization. Taking the active guide vane of the centrifugal pump turbine as the research object, the airfoil profile is parameterized, and the surrogate model's independent variables and training range are determined. Combining programming and numerical simulation software, an efficiency prediction model for the centrifugal pump and water turbine based on guide vane airfoil control variables is constructed. The particle swarm algorithm is used to globally optimize the constructed model to obtain the optimal efficiency point and corresponding airfoil-related parameters. Finally, numerical simulation and experimental research methods are used to validate the predicted data. The results show that under the premise of ensuring grid performance and operational stability, the numerical simulation efficiency of the pump turbine under the optimization scheme is increased by 1.6 and 0.32%, respectively, compared to the numerical efficiency of the prototype guide vane. In the experimental case, the efficiency of the water turbine and pump is increased by 0.76 and 0.14%, respectively, compared to the prototype guide vane.