再压缩 S-CO₂布雷顿循环性能分析及多目标优化

The solar thermal power generation technology combined with heat storage has stable output and strong peak shaving capabilities. The introduction of supercritical carbon dioxide (S-CO₂) Brayton cycle can further improve thermoelectric conversion efficiency. Most of the existing studies evaluated the performance of S-CO₂ cycle based on a single index, leading to inconsistent evaluation results. Hence, it is necessary to carry out multi-index comprehensive evaluation to objectively reflect the cycle performance. In the present paper, mathematical models were established to investigate the thermodynamic performance and economy of a 35 MW recompression S-CO₂ cycle, and the effects of critical parameters on cycle performance was analyzed. A BP-GA optimization method of back propagation neural network combined with elitist nondominated sorting genetic algorithm was constructed for multi-objective optimization of cycle performance. The results indicate that the cycle efficiency increases with an increasing total thermal conductivity of the recuperators, but there is a ceiling on growth. There are significant non-monotonic relations between turbine inlet temperature, minimum cycle pressure, maximum cycle pressure, split ratio and cycle performance, and the corresponding optimal values are 639.14℃, 8.10 MPa, 29.74 MPa and 0.70, respectively. Compared with the cycle performance based on the design conditions, the optimized cycle shows a reduction of 11.1% in LCOE, and an increase of 5.1% and 27.6% in efficiency and specific work, respectively.