Experimental performance characterization of R-454B and R-410A in data center server rack mount cooling unit: Energy efficiency and environmental impact assessment.

Data center cooling accounts for a substantial fraction of facility energy consumption, with environmental pressures driving the transition from high-GWP refrigerants to sustainable alternatives. However, rigorous experimental comparisons of next-generation refrigerants under realistic operational conditions remain limited, and existing benchmarking protocols often fail to account for ambient variability across test conditions. This study addresses these gaps by conducting a comprehensive experimental evaluation of an in-rack air-conditioning unit, comparing the conventional refrigerant R410A with the low-GWP alternative R454B (GWP = 466) across 24 tests spanning varied server loads and thermal conditions. The methodology integrates a transparent steady-state detection algorithm for time-series measurement processing, normalized Energy Efficiency Ratio (EER) calculations at fixed reference conditions to enable equitable cross-test comparison, and sensitivity analyses quantifying the influence of ambient and evaporator-side temperatures on system performance. Cooling capacity, raw and normalized EER, facility-level Power Usage Effectiveness (PUE), and Total Equivalent Warming Impact (TEWI) were derived for each refrigerant. Results demonstrate that R454B exhibits loaddependent performance advantages: under high-load conditions (9 kW), R454B achieved 38 % higher EER (~5.8 vs. ~4.2) and superior cooling capacity (13 kW vs. 9.5 kW median) compared to R410A, with reduced operational variability. However, under supply-air matched baseline conditions (16–20 ◦C), both refrigerants exhibited equivalent performance, confirming that R454B’s efficiency gains emerge primarily under elevated thermal stress and higher refrigerant flow rates. PUE analysis showed equivalent facility-level efficiency, enabling R454B as a direct drop-in replacement. TEWI analysis revealed that indirect emissions dominate climate impact (>90 %), establishing operational efficiency optimization as the primary environmental lever, with refrigerant selection providing secondary benefits through GWP reduction. These findings support the adoption of R454B in high-load data center environments, while the transparent methodological framework provides a reproducible benchmark for condition-aware refrigerant evaluation in mission-critical cooling systems.