Study on performances of supercritical CO2 recompression Brayton cycles with multi-objective optimization

Supercritical carbon dioxide recompression Brayton cycles (SCO2RBC) have the potential for higher practicability and better economy, because its high-efficiency, compact size, using standard construction materials, low-cost, and so on. Based on the mathematical model of thermodynamics, the thermodynamic parameters optimizing method for SCO2RBC by the non-dominated sorting genetic algorithm II (NSGA-II) with exergy efficiency and net power output as its objective functions is presented in this paper. The parametric study shows that the cycle performance could be greatly affected by some key parameters. And the optimization gets the Pareto optimal relation curve, which indicates that they cannot reach the maximum at the same time and there is a conflicting relation between them. In addition, the relationship of the key cycle parameters is established and then a fast query system is built for SCO2RBC design based on artificial neural network (ANN). It shows that the predicted results by ANN are in substantial agreement with optimization results by NSGA-II with a good accuracy and the designers could firstly determine the numerical value of the cycle net power output or the exergy efficiency as needed, and then the key cycle parameters could be retrieved by the query system fast and accurately.