Flow Field Simulation and Model Order Reduction Techniques for Hydro Turbines

To explore the operational patterns of pumped storage power stations in order to meet the demand for real-time acquisition of three-dimensional velocity field data of turbines, the paper establishes a three-dimensional reduced-order model of a turbine based on finite volume method simulation. Using Pera Sim Fluid software to perform numerical simulations of a Francis-99 turbine under various operating conditions and construct a dataset. Modal and modal coefficient extraction is performed on the dataset through Proper Orthogonal Decomposition (POD), and a mathematical relationship between operating variables and modal coefficients is established using Radial Basis Function (RBF) method, thereby achieving rapid prediction of the three-dimensional velocity field of turbines under arbitrary operating variables. The results show that this model can achieve interpolation prediction within the operating range with a small number of samples, demonstrating high generalization and accuracy. The average relative error of the computational results is 0.02%, with a numerical error of less than 0.5 m/s, and the average time for obtaining turbine velocity field data is reduced to 0.5 seconds.