Multi-objective optimization of 200 kW air centripetal turbine based on artificial neural networks

The air-Brayton cycle has the characteristics of safety and high efficiency, which can be used as an energy conversion system for mobile small reactors. The air turbine is one of the key components in the cycle system, and improving its performance is of great significance. In this paper, an artificial neural network model combined with a genetic algorithm was used to optimize the rotor of an air centrifugal turbine with axial thrust and efficiency as the objective. The results show that the artificial neural network model can fit the CFD numerical simulation results well, with a coefficient of determination larger than 0.97. Then, after optimizing the artificial neural network model with a genetic algorithm, the total -total efficiency of the air centrifugal turbine was improved by 1.479 %, while the axial thrust was reduced by 1.07 %.