Compressor models under pressure: modeling impact on refrigerant performance in vapor-injection heat pumps
In cold climate regions with ambient temperatures down to -20 °C, air-to-water heat pumps (AWHP) using single-stage standard cycles face technical limitations. Low source temperatures reduce evaporation pressure and suction gas density, decreasing refrigerant mass flow and heating capacity. Simultaneously, high pressure ratios lead to low isentropic efficiencies and elevated discharge temperatures, restricting compressor operation. Both effects reduce system efficiency, expressed by the coefficient of performance (COP), and increase reliance on auxiliary heating, raising operational costs. To address these limitations, quasi-two-stage vapor injection is identified in literature as a promising enhancement. However, the effectiveness of such cycle modifications depends on accurate compressor modeling. This study investigates how the choice of compressor model affects refrigerant performance predictions in vapor-injection cycles. Three compressor models are compared in a quasi-steady-state simulation framework with thermodynamic properties from REFPROP. One assumes constant isentropic and volumetric efficiencies, while furthermore an empirical and semi-physical model is used. The vapor injection cycle is benchmarked against a simple cycle across source temperatures from -20 °C to 0 °C and a sink temperature of 75 °C. Key performance indicators are COP, heating capacity, and compressor discharge temperature. Refrigerants considered are R290, R1270, and R1243zf. Results show that refrigerant ranking changes with the compressor model. For example, R290 outperforms R1243zf in the semi-empirical model but not in the constant-efficiency case. Different compressor models lead to deviations within the vapor-injection cycle up to 15.9 % for COP, 5.4 % for heating capacity and 11.4 K in discharge temperatures. The findings show that refrigerant selection depends on compressor modeling assumptions. Since performance rankings of refrigerants shift depending on the used model, a more detailed refrigerant-dependent model is required. Experimental validation on a dedicated vapor injection test rig will follow to quantify these interactions under real conditions.