Aerodynamic optimization of a SCO2 radial-inflow turbine based on an improved simulated annealing algorithm

Supercritical carbon dioxide plays a vital role in the development of power generation applications. It owns the characteristics of high density and low viscosity, which can ensure a compact structure for turbomachinery. With the blossom of optimization algorithm, an interdisciplinary research which applies optimization method to a traditional design process of turbomachinery can accelerate the course and promote the validity by leaps and bounds. We improve the traditional simulated annealing algorithm and establish an optimization process containing the optimization of rotor meridian plane and nozzle profile. This process can effectively reduce the computation time by establishing a surrogate model of coarse mesh simulation. The effects of traditional simulated annealing algorithm (SAA), genetic algorithm (GA) and improve simulated annealing algorithm (ISAA) are compared. As a result, we realize a maximum of 4.94% promotion for isentropic efficiency in ISAA computation. Also, ISAA method saves the computation time by 59.6% compared to GA and by 41.5% compared to SAA. Applying ISAA optimization method to the turbine in a kW-scale solar-driven Brayton cycle power system, we realize a 1.17% increase for the system efficiency.