Hydrodynamic Characteristic and Prediction Study of 1, 1, 1, 2-Tetrafluoroethane under Supercritical Pressure

The supercritical organic Rankine cycle (S-ORC) is highly effective for utilizing medium- and low-temperature heat sources. This study investigated the hydrodynamic behavior of supercritical-pressure 1,1,1,2-tetrafluoroethane (R134a) within a horizontal 2 mm circular tube through integrated experimental and machine learning techniques. Experimental investigations spanned pressures between 4.2 and 5.4 MPa, inlet temperatures between 20 and 50 °C, and heat fluxes between 60 and 300 kW/m². Systematic analysis of hydrodynamic characteristics was accompanied by predictive modeling using an extreme learning machine (ELM) framework to forecast pressure drop trends. The hydrodynamic characteristic (HDC) curve of supercritical R134a exhibits significant differences from subcritical flow behavior—it lacks a negative-slope region but features a distinct “pressure drop stabilization region,” where pressure drop remains consistent across a broad range of mass flow rates. The pressure-drop stabilization region diminishes with elevated system pressure or inlet temperature but enhanced with heat flux. Mechanistic analysis revealed that the initiation of this region is predominantly influenced by frictional pressure drop, whereas its termination correlates with acceleration pressure drop. Crucially, no flow instabilities were detected within the pressure-drop stabilization region. However, operation in the low-mass-flow-rate regime of the curve induced dynamic oscillatory behavior, characterized by periodic fluctuations in the mass flow rate, wall and fluid outlet temperatures, system pressure, and pressure drop. These instabilities are attributed to axial fluid density gradients arising from localized thermal nonequilibrium. The ELM model demonstrated robust predictive performance, maintaining errors within ±10% across all operating conditions, highlighting its effectiveness in analyzing supercritical hydrodynamic phenomena.