Heat pumps are today a valuable asset in climate change mitigation and energy decarbonization. To meet the Net Zero Emissions target by 2050, the International Energy Agency forecasts that the global heat pump stocks should increase by over 300% by 2030.

It is widely recognized that the choice of refrigerant used in heat pumps significantly affects their performance. Therefore, accurately ranking available refrigerants is of great importance. However, literature has shown that the compressor behavior strongly depends on the refrigerant selected. As a result, it seems appropriate to perform a new classification that takes into account the effects of refrigerants on the performance of the heat pump and the compressor.

The authors propose an integrated design of the refrigerant and heat pump process with a refrigerant-dependent compressor design model in the form of a nonconvex Mixed Integer Non-Linear Program (MINLP). The MINLP is solved using the Non-Linear Branch and Bound algorithm of Artelys Knitro. To determine the ranking, the optimization is repeatedly executed. Whenever a feasible refrigerant molecular design solution is found, an integer-cut constraint is added to the model to exclude this solution from the design space.

The study conducted with the help of Artelys Knitro shows that when considering the compressor’s refrigerant dependency, the ranking of the top refrigerants changes substantially. This indicates that refrigerant-dependent compressor models are crucial for identifying the optimal combination of refrigerants and heat pump processes. Finally, the authors declare that the complexification brought by integrating a refrigerant-dependent compressor model can only be exploited by using computationally efficient implementations.