Comparative Analysis of Direct Drop-In Fluid Replacement for a Centrifugal Compression System

The phasing out of high-GWP refrigerants and the growing diversity of working fluids used across heat pumps, refrigeration systems, and closed-cycle power applications have made drop-in fluid replacement a question of significant practical interest. Centrifugal compressors are designed around the thermophysical properties of a specific fluid, and the performance penalty is incurred when working fluid is replaced without redesigning the impeller. This study presents a CFD comparison of direct drop-in fluid replacement in a fixed geometry centrifugal compression system. Eight working fluids that span the property range relevant to current drop-in substitutions are evaluated: air, nitrogen, argon, carbon dioxide, R22, R134a, R1234yf, and R1234ze(E). A reference centrifugal impeller was reconstructed in ANSYS BladeGen, meshed in ANSYS TurboGrid using the Automatic Topology and Meshing method, and simulated in ANSYS CFX (2024 R2) as a single periodic passage with Frozen Rotor interfaces and Spalart–Allmaras turbulence closure. Performance maps were generated for each fluid across a range of rotational speeds and mass flow rates, with a common inlet reference condition applied across all cases to isolate the influence of fluid properties from an inlet state. The resulting dataset enables a like-for-like comparison of pressure ratio, efficiency, and shaft power requirement, providing a basis for assessing the aerodynamic implications of drop-in fluid substitution in centrifugal compression systems. Air, nitrogen, argon and carbon dioxide achieved similar peak efficiencies (~88%) and comparable pressure ratios (PR), indicating they can be used as drop-in substitutes without performance loss. Refrigerants R1234yf and R1234ze(E) matched R134a in efficiency (peak ~90%) while offering higher pressure ratios and significantly lower power requirements at peak efficiency. At 20,000 RPM and a mass flow rate of 2 kg/s, compared to a PR of 1.45 for air, nitrogen, carbon dioxide and argon achieved PRs of 1.4, 1.9 and 2.4, respectively. At the same settings, R134a and R1234 refrigerants reached PRs of 5 and 6, respectively. The power requirement was ~8 × 104 W for air and similar fluids, and ~11 × 104 W for refrigerants.