Superstructure-free synthesis and multi-objective optimization of supercritical CO2 cycles

The optimization of the supercritical CO2 cycle structures is mostly experience-intensive and computation-demanding. Using the superstructure-free method can effectively realize computer-aided intelligent construc-tion and optimization of cycle configurations. In this paper, an improved superstructure-free algorithm is pro-posed which goes beyond the state-of-the-art methods by expanding the searching space and evaluation matrices. The improved algorithm can intelligently and simultaneously optimize the cycle structure and design parameters by considering both the net power output and specific investment cost as objective functions. By applying the algorithm, several previously unseen cycle structures were retrieved, and the maximum net power output in-creases by a range of 15.45% to 37.71%. The optimal cycle obtained by entropy weight and TOPSIS analysis of the Pareto solution set is a cascaded configuration of a double compression intercooling cycle and a basic cycle through sharing cooling process, which achieves a net power output of 37.75 kW and a specific investment cost of 5.33 $/W. The improved superstructure-free algorithm can also be applied to other thermodynamic cycles and can greatly contribute to the intelligent configuration design and optimization of similar thermodynamic cycles.