Experimental study of a coiled flow inverter as heat exchanger for its use with TiO2–Water nanofluids and its potential implementation as an enhanced heat exchanger for desorption in a modified commercial Direct–Fired Ammonia–Water absorption chiller

Abstract

The chaotic mixing occurring in the chaotic flow and developed inside a coiled flow inverter (CFI) produces the removal or weakening of thermal and hydrodynamic boundary layers, causing heat and mass transfer augmentation. The present work examines the synergy of incorporating a thermal improved working fluid (i.e. nanofluid) as a means of enhancing the in-tube forced convection heat transfer in a CFI. Also, this work considers a CFI as a potential enhanced desorber in an absorption refrigeration system. Aqueous solutions of dispersed TiO2 nanometer-sized particles (i.e., TiO2–Water nanofluids) were prepared and characterized by measuring their physical and transport properties. The effects on forced convective heat transfer were experimentally investigated in a test rig including a CFI prototype. The convective heat transfer coefficients were obtained at Reynolds numbers (N_Re) and TiO2 nanoparticle volume concentrations ranging from 1400–9500 and 0–1.5 v/v%, respectively. The Nusselt number (N_Nu) in the CFI containing 1.0% v/v nanofluid was 41%–52% higher than in the CFI containing pure base fluid (i.e., water), while the 1.5 v/v% nanofluid increased the N_Nu by 4%–8% compared to water. Two new correlations to predict the N_Nu of TiO2–water nanofluids in the CFI at Reynolds numbers of 1400 ≤ N_Re ≤ 9500 and nanoparticle volume concentrations ranges of 0.2–1.0 v/v% and 0.2–1.5 v/v% are proposed. A fully instrumented test bench was built from a gas–fired activated commercial ammonia–water absorption chiller of 3–ton (i.e. 10 kW cooling capacity). The test bench was charged with ammonia–water solution and startup trials were executed. They showed instabilities in the solution flow. A coiled flow inverter desorber prototype with s curvature ratio () value of 10 was constructed based on the models established on the doctoral thesis of Vargas Bautista [1] which precedes this thesis.