Optimal control strategy for compressed air energy storage under off-design conditions

Compressed air energy storage (CAES) serves as a critical enabling technology for grid-scale integration of renewable energy. Most existing studies on off-design CAES operation focus on improving efficiency through individual control strategies. However, systematic comparisons of different operating strategies, especially their combined effects on volumetric storage density and the underlying thermodynamic trade-offs, are still lacking. This study develops a thermodynamic model for a four-stage CAES system integrated with a dual-tank water-based thermal storage module, and systematically evaluates three operating strategies, namely unregulated operation, inlet guide vane (IGV) regulation, and speed regulation, under variable backpressure conditions. The results demonstrate that speed regulation yields the highest round-trip efficiency (RTE) of 65.68%, which is 1.00 percentage point higher than the 64.68% achieved under unregulated operation. This improvement stems from optimised off-design flow-matching within compressors, which reduces the specific charging energy consumption to 549.33 kJ/kg and lowers the compressor exergy destruction ratio from 36.68% to 33.51%. By comparison, IGV regulation achieves a maximum volumetric energy storage density of 2.78 kWh/m 3 , representing a 0.72% increment over unregulated operation. This density advantage is attributed to an increase in the cumulative stored air mass and an elevated thermal storage temperature. A fundamental thermodynamic trade-off between RTE and storage density is revealed under different control strategies. The quantitative findings provide a rigorous decision-making framework for CAES operational optimization: speed regulation is preferable for efficiency-focused peak shaving, whereas IGV regulation is well-suited for capacity-constrained operational scenarios. The study offers actionable insights for enhancing the thermodynamic performance and grid integration of large-scale CAES systems.