dc.description.abstract |
The aim of this study is to understand the microscopic behavior of heat and momentum
transfer in nanofluids. With nanofluids reporting enhanced thermal conductivities (𝜅)
and viscosities (𝜂), a microscopic understanding is essential for engineering nanofluids
to be practical in heat transfer applications. Therefore, to study the microscopic
transport behavior, copper-argon nanofluids simulated by classical molecular
dynamics are employed. The Applicability of the Green-Kubo (GK) method in
nanofluid 𝜅 evaluation is questioned as the calculated thermal conductivities through
the GK method are considerably higher than the direct method in Non-EquilibriumMolecular-Dynamics
(NEMD). Green-Kubo calculations are found to be very
sensitive to the ill-defined partial enthalpy computation, resulting in an overestimation
of the 𝜅. However, the Green-Kubo and the direct method viscosity calculations
demonstrate a reasonable agreement.
Following the more reliable method, the NEMD direct approach, 𝜅 of the nanofluids
consisting of spherical nanoparticles with different diameters are investigated. The
computational results are compared with the classical effective medium theories and
no anomalous 𝜅 enhancements are observed in the nanofluids having fully dispersed
spherical particles. Various microscopic mechanisms such as liquid layering and
micro-convection are found to be ineffective for 𝜅 enhancements in nanofluids.
However, greatly enhanced 𝜅 are achieved, a maximum of 63% relative to pure argon,
in nanofluids consisting of chain-like particle arrangements. This demonstrates the
potential origin of anomalous 𝜅 enhancements in experimental measurements and the
capability of nanofluids with extended nanostructures to deliver better 𝜅
enhancements.
Further investigating the capability of extended nanostructures in nanofluid thermal
transport, 𝜅 enhancements of nanofluids consisting of nanowires with different lengths
and diameters are evaluated. It is shown that the heat conduction in the parallelly
arranged liquid and the nanowires exhibit a coupled thermal behavior owing to the
interface thermal resistance (R
b
). This contradicts with the predictions of the classical parallel heat conduction model and therefore, a novel model is proposed taking this
coupled behavior into account. New heat transfer characteristics at the nanoscale are
identified including the R
b
-driven coupled heat conduction, the reduced 𝜅 of suspended
nanowires, and the solid-like liquid layering. Using the new model, the importance of
these microscopic thermal characteristics in accurately predicting the effective 𝜅 is
shown. The sole contribution from the solid-like liquid layer to the 𝜅 enhancement is
found to be in between 20-30% for the nanofluids considered.
Extending the investigation of heat transfer phenomena in nanofluids based on
spherical nanoparticles, 𝜂 of nanofluids with different nanoparticle sizes,
concentrations, and arrangements are evaluated. Both the Green-Kubo and the direct
methods are employed and unlike the 𝜅, both methods show a reasonable agreement
with one another. Viscosity is observed to decrease as the particle diameter increases
in fully dispersed nanofluids. The ratio 𝐶
β
shows a decreasing trend indicating
better heat transfer performance in nanofluids with large particles. Nanofluid 𝜂 is
𝜂
𝐶
𝜅
observed to increase rapidly as the concentration increase. This makes 𝐶
β
to
increase as well indicating the diminished heat transfer performance in nanofluids with
high particle concentrations. As the particles in the nanofluid arrange into chain-like
structures, 𝜂 remains unaffected. This makes 𝐶
β
to decrease rapidly indicating the
greater heat transfer performance in nanofluids with chain-like nanoparticle
arrangements or in general, extended nanostructures. |
en_US |