Development of a One-Dimensional Framework for the Performance Prediction and Optimization of Centrifugal Compressors

Energy storage is essential to enable large-scale renewable integration and ensure grid stability. While electrochemical batteries are poorly suited to long-duration storage and pumped hydro is geographically constrained, advanced Compressed Air Energy Storage (CAES) concepts are gaining attention. In particular, isobaric adiabatic CAES systems based on constant-pressure air storage in hard-rock caverns offer the potential for stable operation and high round-trip efficiency, provided that sufficiently high compressor outlet temperatures can be achieved with limited inter-cooling. This work presents an in-house, one-dimensional modular framework for the performance prediction of single-stage centrifugal compressors, conceived as a building block for the preliminary design of multistage configurations in such CAES systems. The model integrates state-of-the-art empirical loss correlations and modular descriptions of compressor components. Validation against benchmark experimental datasets shows good agreement in terms of total-to-total pressure ratio and isentropic efficiency over a relevant mass-flow range. To further assess the robustness of the 1D model, it is coupled with an NSGA-II multi-objective optimization routine and applied to existing impellers geometry. The optimization focuses on improving efficiency and pressure ratio over a limited interval around the design mass flow rate, rather than maximizing the full operating range, consistently with the expected operating conditions of multistage compressors in CAES applications. The selected optimized geometry is reconstructed in 3D and analyzed by RANS simulations, confirming that the 1D tool captures the main performance trends. Finally, a road map is outlined for embedding the present 1D+NSGA-II framework into a flexible design environment for multistage, CAES-oriented compressor architectures.