COMPUTATIONAL MODELING OF THERMAL FIELD-FLOW FRACTIONATION: LIMITATIONS OF THE TALBOT MODEL FOR LIQUID-PHASE NANOFLUIDS

Document Type

Conference Proceeding

Publication Date

3-2026

Department

Department of Mechanical and Aerospace Engineering

Abstract

The performance of nanofluids in thermal management systems is strongly influenced by thermodiffusion (Soret effect), which drives nanoparticle migration under temperature gradients and can lead to aggregation and property changes. This study employs a two-dimensional computational framework modeled after a Thermal Field-Flow Fractionation (ThFFF) channel to investigate thermophoretic transport in water-based nanofluids containing alumina, copper, and hybrid nanoparticles. Simulations were performed in ANSYS Fluent under a 50 K cross-stream gradient using two formulations: (i) the default Talbot model, based on ideal-gas assumptions, and (ii) a Buongiorno-based, property-dependent model implemented via a User-Defined Function (UDF). Results show that the Talbot model predicts cold-wall particle accumulation but fails to differentiate between particle materials or sizes, producing nearly identical concentration and residence-time profiles for all cases. In contrast, the Buongiorno model introduces temperature-dependent coupling among viscosity, density, and thermal conductivity, yielding more physically realistic migration patterns, including partial retention near the hot wall and measurable differences in particle trajectories and velocities. These findings confirm that accurate prediction of thermophoretic behavior in liquid-phase nanofluids requires property-sensitive models rather than constant-property assumptions. Future work will extend simulations to steady-state conditions, explore hybrid particle interactions, and extract thermal diffusion coefficients for validation against experimental ThFFF data.

Publication Title

Proceedings of the Thermal and Fluids Engineering Summer Conference

ISBN

9781567004885

Link to Full Text

Share

COinS