Experimental and theoretical investigation of saturated flow boiling of R410A refrigerant and nanoparticle laden lubricant mixtures in a smooth tube.

Abstract

In air conditioning systems, a small amount of lubricating oil inevitably leaves the compressor and circulates through the other components. This lubricant acts as a contaminant affecting heat transfer and pressure losses in the heat exchangers. Mixtures of refrigerant and nanolubricants, that is, nanoparticles dispersed in the lubricant oils, have shown potential to augment heat transfer in the refrigerant evaporators. However, the mechanisms of heat transfer enhancement due to the nanolubricants are still not well explained. Two-phase flow boiling heat transfer coefficient superposition models, available in the literature, used Dittus-Boelter or Gnielinski correlations to predict convective heat transfer, and Forster-Zuber or Cooper correlations to estimate the nucleate boiling heat transfer. These correlations do not account for the presence of nanoparticles in the flow and are not able to predict the heat transfer enhancements, or sometimes the degradation, observed during flow boiling experiments of refrigerant and nanoparticle laden lubricant mixtures. A new comprehensive model was developed by modifying and integrating existing convective heat transfer models originally developed for nanofluids and pool boiling models for nanolubricants. The new model accounted for the mass conservation of nanoparticles during their migration from the laminar sublayer near the wall to the adjacent turbulent layer near the gas core, and for the effect of nanoparticle concentration on two-phase convective heat transfer multiplier. A new test apparatus was constructed and the newly developed model was experimentally verified with heat transfer data of the single-phase convective heat transfer processes and saturated two-phase flow boiling heat transfer processes of refrigerant R410A with two nanolubricants in a 9.5 mm I.D. smooth copper tube. The selected nanolubricants had non-spherical ZnO nanoparticles and spherical γ-Al2O3 nanoparticles dispersed in Polyolester (POE) lubricant. In the smooth tube, R410A-nanolubricant mixtures, which had higher thermal conductivity and kinematic viscosity than R410A-POE lubricant mixture, showed some degradation in heat transfer coefficient compared to R410A-POE case; but they also had lower pressure drops. The model explained this by accounting for the effect of increased laminar sublayer thickness and reduced thermal conductivity in laminar sublayer due to diffusion of nanoparticles towards the turbulent layer. These phenomena were responsible for the observed two-phase flow boiling heat transfer coefficient degradation when using the present R410A-nanolubricant mixtures Al2O3 nanolubricant shared similar thermal conductivity in the wet state as that of ZnO nanolubricant. However, both data and simulations showed that Al2O3 nanolubricant had about 15% higher heat transfer coefficient that ZnO nanolubricant. The ZnO nanoparticles, with their large aspect ratio, were predicted to diffuse slowly under Brownian motion than Al2O3 nanoparticles. Experiments also showed that long-term flow boiling testing of R410A-nanolubricant mixtures resulted in a continuous and gradual increase of the heat transfer coefficient. A possible explanation was that the nanoparticle deposition on the tube inner wall and its near wall interaction led to small but incremental enhancements in the nucleate boiling phenomena. Finally, if nanoparticles were to be constrained in the laminar sublayer and near the wall of the tube, the predicted heat transfer coefficients from the simulations were higher than that of R410A- POE mixture. The analysis of this case revealed that the nucleate boiling contribution was significantly augmented for such scenario.