A critical review and comparative study on positive displacement compressor models for fast performance prediction in wide working condition ranges

Positive displacement compressors, e.g., reciprocating, rotary, and scroll compressors, are commonly employed in vapor compression refrigeration systems using various refrigerants, and they may operate under a wide range of working conditions. In the simulation-based design of refrigeration systems, a suitable compressor model for fast and stable prediction of compressor performance is extremely important. This study aims to investigate the applicability of existing models in wide working condition ranges, and to recommend suitable ones which could simultaneously satisfy the following three requirements: high accuracy, reliable extrapolation capability, and small amount of data required for model calibration. The most representative compressor models, including eight data-driven models and four semi-empirical ones, are selected from publications, and they are validated using data from experiments conducted by the present authors as well as those from publications. Model validation results show that the top-performing data-driven and semi-empirical models can achieve sufficient accuracy with an average deviation of less than 2%; compared to data-driven models, semi-empirical models can better maintain accuracy and reasonable trends when extrapolating, and require fewer calibration data to achieve the same level of accuracy. In conclusion, among the existing fast performance-prediction compressor models, semi-empirical ones should be preferentially recommended.