Technoeconomic analysis of CO2-Mixture cycles under variable operation for concentrating solar power

This study evaluates the thermodynamic optimization and system-level annual performance of CO₂-mixture power cycles for Gen2 concentrating solar power plants under variable off-design operation. Starting from multi-objective optimization that identified Pareto-optimal solutions across five cycle configurations and twelve dopant options, thirteen cycles are selected to explore two strategic questions: whether compromise solutions between thermal efficiency and primary heat exchanger temperature difference yield better LCOE, and how to optimally utilize additional temperature gains from CO₂-mixtures. For cycles offering temperature differences above 177°C, parallel “adjusted” and “normalized” configurations test whether additional temperature gains should reduce thermal storage costs or primary heat exchanger sizing. A key methodological contribution involves translating existing off-design operational principles into System Advisor Model’s user-defined power cycle format while employing a hybrid operational strategy combining fixed and sliding pressure modes. Variable operation consistently outperforms constant operation by 3–4 $/MWh. Annual simulations for a 100 MW plant in Seville demonstrate that 53%CO₂-47%COS mixture in recompression with intercooling achieves the minimum LCOE of 103.6 $/MWh, outperforming higher-efficiency CO₂-mixture and pure CO₂ cycles. A precompression cycle with 55%CO₂-45%COS mixture achieves the lowest total capital cost and second-lowest LCOE of 104.2 $/MWh despite moderate thermal efficiency through reduced thermal storage costs, suggesting extreme temperature gains can compensate for lower efficiency. While the trade-off between thermal storage and primary heat exchanger cost reduction remains inconclusive for Gen2 systems, with “adjusted” configurations experiencing higher annual molten salt pumping power consumption than “normalized” configurations of the same cycles, it presents promising potential for Gen3 applications.