Benchmarking and Validation of Reciprocating Piston Compressor Models for Industrial Heat Pump Design
The economic viability of industrial heat pumps largely depends on compressor efficiency. Therefore, designing a heat pump requires selecting an appropriate compressor model. Several reciprocating compressor models have been introduced, each differing in approach and underlying physics. However, their relative strengths and weaknesses remain unclear. In this study, we examine three recently published compressor models: an empirical correlation, a loss-based model, and a differential model. We establish a comprehensive compressor performance dataset, by combining published experimental data and synthetic data generated with standardized compressor performance correlations (DIN EN 12900 + manufacturer coefficients). Over 20,000 data points are made publicly available with this work. We then use the dataset to evaluate the models’ predictions across various refrigerants, operating points, and cylinder geometries relevant to industrial heat pumps. Our results show that both the correlation and the loss-based model deliver good prediction accuracy when properly fitted. These models predict compressor isentropic efficiency with average deviations of 2% to 6% across most of the investigated parameter space. Good accuracy is also demonstrated for out-of-sample refrigerants, including CO2 and refrigerant mixtures. The loss-based model is more robust than the correlation for variations in compressor geometry. Both models are fast, with average runtimes on the order of milliseconds. The differential model, fitted to residential-regime compressor data but with greater physical detail, achieves lower accuracy and requires longer runtimes. Our findings suggest that targeted parameter re-fitting is more effective than adding physical detail for improving compressor efficiency predictions in industrial heat pump modeling.