Development of a generalized ANN-based evaporation heat transfer correlation for refrigerants in turbulent flow

Abstract Global warming, primarily caused by the accumulation of greenhouse gases, poses a major threat due to its impact on climate systems. Transitioning to low Global Warming Potential (GWP) refrigerants, such as R1234yf, in cooling systems is crucial to mitigate this. Although numerous evaporation correlations exist in the literature for turbulent flow, each correlation is typically valid only for a specific range of operating conditions (e.g., Reynolds number, Prandtl number, heat flux, saturation temperature). No generalized correlation currently exists that can accurately predict the evaporation heat transfer coefficient of refrigerants across the combined range of conditions covered by these individual correlations. As a novel contribution, present study an ANN-based generalized correlation developed using an aggregated dataset spanning these diverse ranges. The proposed artificial neural network-based correlation achieved high accuracy, with an R $$^2$$ 2 value exceeding 98% and a MAPE of 8.3%, indicating an acceptable prediction error. By covering a wide range of conditions with a single equation, suggested correlation reduces the experimental burden and accelerates the development of low-GWP refrigerant systems.