Experimental study on heat-transfer performance of supercritical-pressure R134a and predictions based on artificial neural network

Abstract

The supercritical organic Rankine cycle is a critical technology for efficiently using medium- or low-temperature heat sources, where heat transfer by supercritical organic fluids is crucial. This study experimentally investigates the heat-transfer performance of supercritical-pressure R134a in a heated tube and predicts its heat transfer coefficient. The wall-temperature distributions and heat transfer coefficients are obtained. Moreover, the buoyancy effect on the heat transfer is discussed. The results show that the buoyancy effect is evident in low-enthalpy regimes, and it is more significant at lower mass and higher heat fluxes but is slightly affected by pressure. Furthermore, empirical correlations and an artificial neural network (ANN) model are adopted to predict the heat transfer coefficient, and their prediction performances are quantitatively evaluated. The results show that the ANN prediction performance is significantly improved over traditional correlations, demonstrating that ANN is an effective tool for predicting the heat-transfer performance of supercritical-pressure fluids.