<p>The present work contributes to advancing the theoretical understanding of ferroconvection phenomena in porous media, which is critical for engineering applications involving magnetic nanofluids, such as biomedical cooling, energy systems, and microfluidics. In this work, we investigate convective heat transfer and subcritical dynamics in rotating ferrofluids with couple stresses, incorporating local thermal non-equilibrium effects (LTNE). The study employs a two-temperature model to describe heat exchange between solid and liquid phases and the Darcy-Brinkman model for ferrofluid flow in porous media. Stability and convection onset are analyzed under free-free boundary conditions using linear and nonlinear methods, with the Galerkin technique solving the resulting eigenvalue problems. To enhance the predictive capabilities of the study, a hybrid Artificial Neural Network (ANN) model was developed, trained using key parameters (magnetization, couple stresses, permeability, rotation, porosity-modified conductivity ratio, and interface heat transfer coefficient) as inputs and corresponding Rayleigh numbers as outputs. Rayleigh number differences reveal a significant subcritical region. The results indicated that increasing magnetization, permeability, and porosity-modified conductivity ratio reduced the critical Rayleigh number, thus destabilizing the system. Conversely, higher couple stresses, and interphase heat transfer coefficient delayed the onset of convection and stabilized the system. Furthermore, the subcritical region was notably expanded under strong couple stress. The ANN demonstrated high accuracy (<InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="11242_2025_2204_Article_IEq1.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="103" /> </InlineMediaObject> <EquationSource Format="TEX">$R^2 = 0.999287$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msup> <mi>R</mi> <mn>2</mn> </msup> <mo>=</mo> <mn>0.999287</mn> </mrow> </math></EquationSource> </InlineEquation>) in predicting Rayleigh numbers, closely matching analytical results. This hybrid approach offers novel insights into ferrofluid convection dynamics under LTNE conditions and highlights the interplay between thermal and flow control mechanisms.</p>

Modeling Heat Transfer in Rotating Ferrofluids with Couple Stresses Under Local Thermal Non-equilibrium Using ANN

期刊论文
发表日期：

2025年8月5日

查看全文

Akanksha Thakur,
Sunil Kumar,
Reeta Devi

摘要

The present work contributes to advancing the theoretical understanding of ferroconvection phenomena in porous media, which is critical for engineering applications involving magnetic nanofluids, such as biomedical cooling, energy systems, and microfluidics. In this work, we investigate convective heat transfer and subcritical dynamics in rotating ferrofluids with couple stresses, incorporating local thermal non-equilibrium effects (LTNE). The study employs a two-temperature model to describe heat exchange between solid and liquid phases and the Darcy-Brinkman model for ferrofluid flow in porous media. Stability and convection onset are analyzed under free-free boundary conditions using linear and nonlinear methods, with the Galerkin technique solving the resulting eigenvalue problems. To enhance the predictive capabilities of the study, a hybrid Artificial Neural Network (ANN) model was developed, trained using key parameters (magnetization, couple stresses, permeability, rotation, porosity-modified conductivity ratio, and interface heat transfer coefficient) as inputs and corresponding Rayleigh numbers as outputs. Rayleigh number differences reveal a significant subcritical region. The results indicated that increasing magnetization, permeability, and porosity-modified conductivity ratio reduced the critical Rayleigh number, thus destabilizing the system. Conversely, higher couple stresses, and interphase heat transfer coefficient delayed the onset of convection and stabilized the system. Furthermore, the subcritical region was notably expanded under strong couple stress. The ANN demonstrated high accuracy ( $R^2 = 0.999287$ $R^{2} = 0.999287$ ) in predicting Rayleigh numbers, closely matching analytical results. This hybrid approach offers novel insights into ferrofluid convection dynamics under LTNE conditions and highlights the interplay between thermal and flow control mechanisms.