Showing 1–7 of 7 results for author: Aubry, N

Data-Driven Modeling of Geometry-Adaptive Steady Heat Transfer based on Convolutional Neural Networks: Heat Convection

Authors: Jiang-Zhou Peng, Xianglei Liu, Nadine Aubry, Zhihua Chen, Wei-Tao Wu

Abstract: In this paper, based on neural networks, we develop a data-driven model for extremely fast prediction of steady-state heat convection of a hot object with arbitrary complex geometry in a two-dimensional space. According to the governing equations, the steady-state heat convection is dominated by convection and thermal diffusion terms, thus the distribution of the physical fields would exhibit stro… ▽ More In this paper, based on neural networks, we develop a data-driven model for extremely fast prediction of steady-state heat convection of a hot object with arbitrary complex geometry in a two-dimensional space. According to the governing equations, the steady-state heat convection is dominated by convection and thermal diffusion terms, thus the distribution of the physical fields would exhibit stronger correlations between adjacent points. Therefore, the proposed neural network model uses Convolutional Neural Network (CNN) layers as the encoder and Deconvolutional Neural Network (DCNN) layers as the decoder. Compared with fully connected (FC) network model, the CNN based model is good for capturing and reconstructing the spatial relationships of low-rank feature spaces, such as edge intersections, parallelism and symmetry. Furthermore, we applied the signed distance function (SDF) as the network input for representing the problem geometry, which contains more information compared to a binary image. For displaying the strong learning and generalization ability of the proposed network model, the training dataset only contains hot objects with simple geometries: triangles, quadrilaterals, pentagons, hexagons and dodecagons; while the validating cases use arbitrary and complex geometries. According to the study, the trained network model can accurately predict the velocity and temperature field of the problems with complex geometries which has never been seen by the network model during the model training; and the prediction speed is four orders faster than the CFD. The ability of the accurate and extremely faster prediction of the network model suggests the potentials of applying such kind of models to the applications of real-time control, optimization, and design in future. △ Less

Submitted 10 January, 2021; originally announced January 2021.

Comments: arXiv admin note: substantial text overlap with arXiv:2010.03854

Data-Driven Modeling of Geometry-Adaptive Steady Heat Transfer based on Convolutional Neural Networks: Heat Conduction

Authors: Jiang-Zhou Peng, Xianglei Liu, Nadine Aubry, Zhihua Chen, Wei-Tao Wu

Abstract: Numerical simulation of steady-state heat conduction is common for thermal engineering. The simulation process usually involves mathematical formulation, numerical discretization and iteration of discretized ordinary or partial differential equations depending on complexity of problems. In current work, we develop a data-driven model for extremely fast prediction of steady-state heat conduction of… ▽ More Numerical simulation of steady-state heat conduction is common for thermal engineering. The simulation process usually involves mathematical formulation, numerical discretization and iteration of discretized ordinary or partial differential equations depending on complexity of problems. In current work, we develop a data-driven model for extremely fast prediction of steady-state heat conduction of a hot object with arbitrary geometry in a two-dimensional space. Mathematically, the steady-state heat conduction can be described by the Laplace's equation, where a heat (spatial) diffusion term dominates the governing equation. As the intensity of the heat diffusion only depends on the gradient of the temperature field, the temperature distribution of the steady-state heat conduction displays strong features of locality. Therefore, in current approach the data-driven model is developed using convolution neural networks (CNNs), which is good at capturing local features (sub-invariant) thus can be treated as numerical discretization in some sense. Furthermore, in our model, a signed distance function (SDF) is proposed to represent the geometry of the problem, which contains more information compared to a binary image. For the training datasets, the hot objects are consisting of five simple geometries: triangles, quadrilaterals, pentagons, hexagons and dodecagons. All the geometries are different in size, shape, orientation and location. After training, the data-driven network model is able to accurately predict steady-state heat conduction of hot objects with complex geometries which has never been seen by the network model; and the prediction speed is three to four orders faster than numerical simulation. According to the outstanding performance of the network model, it is hoped that this approach can serve as a valuable tool for applications of engineering optimization and design in future. △ Less

Submitted 8 October, 2020; originally announced October 2020.

A non-linear fluid suspension model for blood flow

Authors: Wei-Tao Wu, Nadine Aubry, James F. Antaki, Mehrdad Massoudi

Abstract: Motivated by the complex rheological behaviors observed in small/micro scale blood vessels, such as the Fahraeus effect, plasma-skimming, shear-thinning, etc., we develop a non-linear suspension model for blood. The viscosity is assumed to depend on the volume fraction (hematocrit) and the shear rate. The migration of the red blood cells (RBCs) is studied using a concentration flux equation. A par… ▽ More Motivated by the complex rheological behaviors observed in small/micro scale blood vessels, such as the Fahraeus effect, plasma-skimming, shear-thinning, etc., we develop a non-linear suspension model for blood. The viscosity is assumed to depend on the volume fraction (hematocrit) and the shear rate. The migration of the red blood cells (RBCs) is studied using a concentration flux equation. A parametric study with two representative problems, namely simple shear flow and a pressure driven flow demonstrate the ability of this reduced-order model to reproduce several key features of the two-fluid model (mixture theory approach), with much lower computational cost. △ Less

Submitted 10 September, 2018; v1 submitted 24 August, 2018; originally announced August 2018.

Transport of platelets induced by red blood cells based on mixture theory

Authors: Wei-Tao Wu, Nadine Aubry, James F. Antaki, Mehrdad Massoudi

Abstract: Thrombosis is a common complication following the surgical implantation of blood contacting devices, and is strongly influenced by the phenomenon of near-wall enrichment of platelets. This paper describes a multi-constituent continuum approach to study this phenomenon. A mixture-theory model is used to describe the motion of the plasma and the red blood cells (RBCs) and the interactions between th… ▽ More Thrombosis is a common complication following the surgical implantation of blood contacting devices, and is strongly influenced by the phenomenon of near-wall enrichment of platelets. This paper describes a multi-constituent continuum approach to study this phenomenon. A mixture-theory model is used to describe the motion of the plasma and the red blood cells (RBCs) and the interactions between the two components. A transport model is developed to study the influence of the RBC field on the platelets. The model is used to study blood flow in a rectangular micro-channel, a sudden expansion micro-channel, and a channel containing micro crevices (representing a practical problem encountered in most blood-wetted devices). The simulations show that in the rectangular channel the concentration of platelets near the walls is about five times higher than the concentration near the centerline of the channel. It is also noticed that in the channel with crevices, extremely a large number of platelets accumulate in the deep part of the crevices and this may serve as the nidus for thrombosis occurring in medical devices. Keywords: Platelets; Blood flow; RBCs; Micro-channel; Mixture theory; Two-fluids; Thrombosis △ Less

Submitted 7 May, 2017; v1 submitted 1 March, 2017; originally announced March 2017.

Normal stress effects in the gravity driven flow of granular materials

Authors: Wei-Tao Wu, Nadine Aubry, James F. Antaki, Mehrdad Massoudi

Abstract: In this paper, we study the fully developed gravity-driven flow of granular materials between two inclined planes. We assume that the granular materials can be represented by a modified form of the second-grade fluid where the viscosity depends on the shear rate and volume fraction and the normal stress coefficients depend on the volume fraction. We also propose a new isotropic (spherical) part of… ▽ More In this paper, we study the fully developed gravity-driven flow of granular materials between two inclined planes. We assume that the granular materials can be represented by a modified form of the second-grade fluid where the viscosity depends on the shear rate and volume fraction and the normal stress coefficients depend on the volume fraction. We also propose a new isotropic (spherical) part of the stress tensor which can be related to the compactness of the (rigid) particles. This new term ensures that the rigid solid particles cannot be compacted beyond a point, namely when the volume fraction has reached the critical/maximum packing value. The numerical results indicate that the newly proposed stress tensor has an obvious and physically meaningful effects on both the velocity and the volume fraction fields. △ Less

Submitted 10 February, 2017; v1 submitted 6 February, 2017; originally announced February 2017.

Comments: This manuscript has been submitted to the International Journal of Non-Linear Mechanics

Multi-Constituent Simulation of Thrombus Deposition

Authors: Wei-Tao Wu, Megan A. Jamiolkowski, William R. Wagner, Nadine Aubry, Mehrdad Massoudi, James F. Antaki

Abstract: In this paper, we present a spatio-temporal mathematical model for simulating the formation and growth of a thrombus. Blood is treated as a multi-constituent mixture comprised of a linear fluid phase and a thrombus (solid) phase. The transport and reactions of 10 chemical and biological species are incorporated using a system of coupled convection-reaction-diffusion (CRD) equations to represent th… ▽ More In this paper, we present a spatio-temporal mathematical model for simulating the formation and growth of a thrombus. Blood is treated as a multi-constituent mixture comprised of a linear fluid phase and a thrombus (solid) phase. The transport and reactions of 10 chemical and biological species are incorporated using a system of coupled convection-reaction-diffusion (CRD) equations to represent three processes in thrombus formation: initiation, propagation and stabilization. Computational fluid dynamic (CFD) simulations using the libraries of OpenFOAM were performed for two illustrative benchmark problems: in vivo thrombus growth in an injured blood vessel and in vitro thrombus deposition in micro-channels (1.5mm x 1.6mm x 0.1mm) with small crevices (125μm x 75μm and 125μm x 137μm). For both problems, the simulated thrombus deposition agreed very well with experimental observations, both spatially and temporally. Based on the success with these two benchmark problems, which have very different flow conditions and biological environments, we believe that the current model will provide useful insight into the genesis of thrombosis in blood-wetted devices, and provide a tool for the design of less thrombogenic devices. △ Less

Submitted 1 December, 2016; v1 submitted 25 January, 2016; originally announced January 2016.

Comments: Submission planned to Biophysical Journal

Controlling Mixing Inside a Droplet by Time Dependent Rigid-body Rotation

Authors: R. Chabreyrie, D. Vainchtein, C. Chandre, P. Singh, N. Aubry

Abstract: The use of microscopic discrete fluid volumes (i.e., droplets) as microreactors for digital microfluidic applications often requires mixing enhancement and control within droplets. In this work, we consider a translating spherical liquid droplet to which we impose a time periodic rigid-body rotation which we model using the superposition of a Hill vortex and an unsteady rigid body rotation. This… ▽ More The use of microscopic discrete fluid volumes (i.e., droplets) as microreactors for digital microfluidic applications often requires mixing enhancement and control within droplets. In this work, we consider a translating spherical liquid droplet to which we impose a time periodic rigid-body rotation which we model using the superposition of a Hill vortex and an unsteady rigid body rotation. This perturbation in the form of a rotation not only creates a three-dimensional chaotic mixing region, which operates through the stretching and folding of material lines, but also offers the possibility of controlling both the size and the location of the mixing. Such a control is achieved by judiciously adjusting the three parameters that characterize the rotation, i.e., the rotation amplitude, frequency and orientation of the rotation. As the size of the mixing region is increased, complete mixing within the drop is obtained. △ Less

Submitted 22 September, 2008; originally announced September 2008.

Comments: 6 pages, 6 figures

Journal ref: IMECE2008-68026