Systematic Empirical Modelling and evaluation of supercritical Nusselt number correlations

Supercritical fluids are increasingly utilized in various industrial applications due to their unique properties. Accurate prediction of heat transfer in supercritical fluids is crucial for the design and optimization of these systems. However, the existing empirical correlations for supercritical Nusselt number often exhibit large uncertainties and discrepancies. In this study, a comprehensive Systematic Empirical Modelling (SEM) and evaluation of supercritical Nusselt number correlations for carbon dioxide flow in bare tubes was conducted. A total of 42 dimensionless variables were investigated, generating over 6 million potential forms of empirical correlations. The most accurate correlations, involving three to six dimensionless variables, were identified based on the root mean square percentage error between experimental and recalculated or predicted values of the Nusselt number and wall temperature. The identified correlations exhibited good agreement with the experimental data, with root mean square percentage errors ranging from 4.949% to 5.759% for the wall temperature. This study underscores the substantial benefits of employing SEM and systematic evaluation in identifying the most accurate correlations, and facilitates the development of robust and dependable correlations for system-level thermal-hydraulic codes. [Display omitted] •The SEM method is employed to fit approximately 6.2 million variations of correlations.•42 dimensionless variables are systematically tested to model the Nusselt number.•Identification of the best Nusselt number correlation with MPE of wall temperature at 0.129%.•Multi-range bisection method is adopted for comprehensive wall temperature approximation.•This work provides valuable insights into the conventional empirical modeling techniques.