Abstract:  Aiming at the problems of a large number of variables required for parameterized expression of turbine blades and a large amount of calculation in the optimization process in the process of optimal design of hydro turbines, an optimal design method of hydro turbine runner blades was proposed. First, the model hydraulic turbine was numerically simulated by computational fluid dynamics (CFD) technology; then, the quadratic curve was used to parametrically fit the airfoil curve of the turbine blade, and the blade shape was controlled by changing the proportional coefficient of part of the blade position; through CFD numerical simulation technology, the calculation results of blades under different parameters under the same working conditions was calculated,the sample space was generated; through the back propagation neural network(BPNN) and genetic algorithm, the efficiency and hydraulic performance of the turbine were optimized. Finally, the obtained optimization results were calculated by CFD numerical simulation and compared with the prototype blade. The research results show that the efficiency of the optimized tubular turbine is increased by 2.5%, and the pressure distribution of the blade is effectively improved, which proves that the optimization of the blade airfoil curve by this method is effective and has practical engineering application value.

Key words: hydraulic machinery; runner blades; computational fluid dynamics (CFD); back propagation neural network (BPNN); blade airfoil