Research article

RBF simulation of natural convection in a nanofluid-filled cavity

  • Received: 31 July 2016 Accepted: 26 August 2016 Published: 30 August 2016
  • In this study, natural convection in a cavity filled with a nanofluid is solved numerically utilizing a radial basis function pseudo spectral (RBF-PS) approach in the space domain and a differential quadrature method (DQM) in the time domain. The governing dimensionless equations are solved in terms of stream function, temperature and vorticity. In the cavity, thermally insulated top and bottom walls are maintained while the left and right walls are at constant temperatures. Numerical solutions present the average Nusselt number variation as well as streamlines, isotherms and vorticity contours. The non-dimensional problem parameters, Rayleigh number Ra, solid volume fraction χ and aspect ratio AR are varied as 103Ra106,0χ0.2 and AR=0.25,0.5,1,2,4, respectively. It is found that the fluid velocity and the heat transfer are enhanced in presence of nanoparticles, and the convective heat transfer is reduced in a rectangular cavity.

    Citation: Bengisen Pekmen Geridonmez. RBF simulation of natural convection in a nanofluid-filled cavity[J]. AIMS Mathematics, 2016, 1(3): 195-207. doi: 10.3934/Math.2016.3.195

    Related Papers:

    [1] Bengisen Pekmen Geridönmez . Numerical investigation of ferrofluid convection with Kelvin forces and non-Darcy effects. AIMS Mathematics, 2018, 3(1): 195-210. doi: 10.3934/Math.2018.1.195
    [2] Bin He . Developing a leap-frog meshless methods with radial basis functions for modeling of electromagnetic concentrator. AIMS Mathematics, 2022, 7(9): 17133-17149. doi: 10.3934/math.2022943
    [3] Fahad Alsharari, Mohamed M. Mousa . New application of MOL-PACT for simulating buoyancy convection of a copper-water nanofluid in a square enclosure containing an insulated obstacle. AIMS Mathematics, 2022, 7(11): 20292-20312. doi: 10.3934/math.20221111
    [4] Suliman Khan, M. Riaz Khan, Aisha M. Alqahtani, Hasrat Hussain Shah, Alibek Issakhov, Qayyum Shah, M. A. EI-Shorbagy . A well-conditioned and efficient implementation of dual reciprocity method for Poisson equation. AIMS Mathematics, 2021, 6(11): 12560-12582. doi: 10.3934/math.2021724
    [5] Saleh Mousa Alzahrani . Enhancing thermal performance: A numerical study of MHD double diffusive natural convection in a hybrid nanofluid-filled quadrantal enclosure. AIMS Mathematics, 2024, 9(4): 9267-9286. doi: 10.3934/math.2024451
    [6] C. S. K. Raju, S.V. Siva Rama Raju, S. Mamatha Upadhya, N. Ameer Ahammad, Nehad Ali Shah, Thongchai Botmart . A numerical study of swirling axisymmetric flow characteristics in a cylinder with suspended PEG based magnetite and oxides nanoparticles. AIMS Mathematics, 2023, 8(2): 4575-4595. doi: 10.3934/math.2023226
    [7] Kuiyu Cheng, Abdelraheem M. Aly, Nghia Nguyen Ho, Sang-Wook Lee, Andaç Batur Çolak, Weaam Alhejaili . Exothermic thermosolutal convection in a nanofluid-filled square cavity with a rotating Z-Fin: ISPH and AI integration. AIMS Mathematics, 2025, 10(3): 5830-5858. doi: 10.3934/math.2025268
    [8] Kawther Al Arfaj, Jeremy Levesly . Lagrange radial basis function collocation method for boundary value problems in 1D. AIMS Mathematics, 2023, 8(11): 27542-27572. doi: 10.3934/math.20231409
    [9] Kottakkaran Sooppy Nisar, Muhammad Shoaib, Muhammad Asif Zahoor Raja, Yasmin Tariq, Ayesha Rafiq, Ahmed Morsy . Design of neural networks for second-order velocity slip of nanofluid flow in the presence of activation energy. AIMS Mathematics, 2023, 8(3): 6255-6277. doi: 10.3934/math.2023316
    [10] Imran Siddique, Yasir Khan, Muhammad Nadeem, Jan Awrejcewicz, Muhammad Bilal . Significance of heat transfer for second-grade fuzzy hybrid nanofluid flow over a stretching/shrinking Riga wedge. AIMS Mathematics, 2023, 8(1): 295-316. doi: 10.3934/math.2023014
  • In this study, natural convection in a cavity filled with a nanofluid is solved numerically utilizing a radial basis function pseudo spectral (RBF-PS) approach in the space domain and a differential quadrature method (DQM) in the time domain. The governing dimensionless equations are solved in terms of stream function, temperature and vorticity. In the cavity, thermally insulated top and bottom walls are maintained while the left and right walls are at constant temperatures. Numerical solutions present the average Nusselt number variation as well as streamlines, isotherms and vorticity contours. The non-dimensional problem parameters, Rayleigh number Ra, solid volume fraction χ and aspect ratio AR are varied as 103Ra106,0χ0.2 and AR=0.25,0.5,1,2,4, respectively. It is found that the fluid velocity and the heat transfer are enhanced in presence of nanoparticles, and the convective heat transfer is reduced in a rectangular cavity.


    1. Introduction

    Natural convective heat transfer has seen a great deal of interest in the last decades due to many engineering applications such as insulation of buildings, solar energy collectors, cooling systems for electronic devices, etc. In order to improve the heat transfer characteristics of traditional liquids, nanosized metallic particles are inserted into the liquid. This increases the thermal conductivity of the fluid.

    There are a considerable amount of contributions modelling the heat transfer of the nanofluid by using appropriate numerical approaches. The most common numerical approach is the finite volume method [6,8,10,12,15,17,19]. Jou et al. [9] implemented a finite difference solution with the line it-erative method. Fattahi et al. [5] and Ashorynejad et al. [1] employed the Lattice Boltzmann method to simulate the heat transfer comparing different nanofluids, and considering the presence of magnetic field, respectively. Gumgum and Tezer-Sezgin showed the effeciency of dual reciprocity boundary element method on natural convection flow of nanofluids. A novel approach is seen in Serna et al. [16] with the usage of the network simulation method to show the influence of a pulsating flow in a heated lid-driven cavity filled with a nanofluid.

    To the best of the author’s knowledge, this is the first application of RBF-PS and DQM to simulate natural convection in an enclosure filled with a nanofluid. The main effect of this process is to be able to use a small number of grid points in the space domain, and to find the solution both in the space and the time domain at once. The average Nusselt number variation in different nanofluids with different solid volume fractions, and in different aspect ratios is presented. Streamlines, isotherms and vorticity contours are also illustrated.


    2. Problem Setup

    The problem geometry is described in Figure 1. No-slip boundary conditions on the walls (u=v=ψ=0) are imposed while the left wall is the hot wall and right wall is the cold wall. The top and bottom walls are adiabatic. Aspect ratio of the cavity is W:H.

    Figure 1. Problem Configuration.

    The nanofluid in the enclosure is laminar, incompressible and Newtonian. The nanoparticles are assumed to be in spherical shape, and the nanoparticles and the fluid are assumed to be in thermal equilibrium. Thermophysical properties of the nanofluid are also assumed constant except the density variation in the buoyancy force term treated by the Boussinesq approximation. The radiation effect and viscous dissipation are negligible.

    For some nanoparticles copper (Cu), silver (Ag), aluminium oxide (Al2O3), silicon oxide (SiO2), titanium oxide (TiO2), copper oxide (CuO), thermophysical properties are given in the Table 1.

    Table 1.Physical properties of fluid and solid phases.
    PropertyWaterCuAgAl2O3SiO2TiO2CuO
    Cp (J/kgK)4179385235765765686.2535.6
    ρ (kg/m3)997.18933105003970397042506320
    k (W/mK)0.61340142925368.9576.5
    β×105(K1)211.671.890.850.630.91.8
     | Show Table
    DownLoad: CSV

    3. Governing Equations

    The continuity, momentum and energy equations are [2,13,19]

    ux+vy=0 (3.1a)
    ut+uux+vuy=1ρnfPx+μnfρnf(2ux2+2uy2) (3.1b)
    vt+uvx+vvy=1ρnfPy+μnfρnf(2vx2+2vy2)+(ρβ)nfρnfg(TTc) (3.1c)
    Tt+uTx+vTy=αnf(2Tx2+2Ty2), (3.1d)

    where

    the effective density of the nanofluid is ρnf=(1χ)ρf+χρs,

    the effective dynamic viscosity is μnf=μf(1χ)2.5,

    the thermal expansion coefficient of the nanofluid is βnf,

    the heat capacity of the nanofluid is (Cp)nf,

    the thermal diffusivity of the nanofluid is αnf=knf(ρCp)nf,

    the effective thermal conductivity of the nanofluid is knf=kfks+2kf2χ(kfks)ks+2kf+χ(kfks),

    and χ is the solid volume fraction, subindices f and s refer to fluid and nanosized solid particle, respectively. Also, the nanofluid has the following properties

    (ρβ)nf=(1χ)(ρβ)f+χ(ρβ)s, (3.2)
    (ρCp)nf=(1χ)(ρCp)f+χ(ρCp)s. (3.3)

    In order to get the non-dimensional form of the governing equations, the non-dimensional parameters are defined as

    x=x, y=y, u=uαf, v=vαf, P=p2ρnfα2f, T=TTcThTc, (3.4)

    in which is the characteristic length.

    Substituting these parameters in Eq.(3.4) into the governing equations Eq.(3.1), and then dropping the prime notation, the following dimensionless equations are obtained

    ux+vy=0 (3.5a)
    ut+uux+vuy=ρfρnfPx+Prμnfρfμfρnf(2ux2+2uy2) (3.5b)
    vt+uvx+vvy=ρfρnfPy+Prμnfρfμfρnf(2vx2+2vy2)+(ρβ)nfρnfβfRaPrT (3.5c)
    Tt+uTx+vTy=αnfαf(2Tx2+2Ty2). (3.5d)

    The definitions of velocity components in terms of stream function ψ, u=ψ/y, v=ψ/x (which satisfy the continuity equation) and of vorticity result in ω=×u=vxuy=2ψ. Pressure terms in Eqs.(3.5b)-(3.5c) are eliminated by applying the definition of vorticity to these equations. By this way, vorticity equation is derived. Thus, the dimensionless governing equations in terms of stream function ψ, temperature T and vorticity ω are

    2ψ=ω (3.6a)
    2T=αfαnf(Tt+uTx+vTy) (3.6b)
    2ω=ρnfμfμnfρfPr(ωt+uωx+vωy)μf(ρβ)nfμnfρfβfRaTx, (3.6c)

    where Pr=νfαf is the Prandtl number and Ra=gβf(ThTc)3αfνf is the Rayleigh number.


    4. RBF-PS in space and DQM in time

    The diffusion-convection type partial differential equation

    φi=Ni+Nbj=1αjfij, (4.1)

    may be approximated by RBFs of the form

    φi=Ni+Nbj=1αjfij, (4.2)

    where φ is an unknown (ψ,T or ω), αj's are initially unknown coefficients, f's are approximating functions formed by RBFs depending on radial distance r=||xxj|| in which x=(x,y) is the field point and xj=(xj,yj) is the collocation point, Ni is the number of interior nodes, and Nb is the number of boundary nodes.

    In matrix-vector form, Eq.(4.2) may also be written as

    φ=Fα  α=F1φ. (4.3)

    The matrix F of size (Ni+Nb)×(Ni+Nb) is the matrix formed by fj's columnwise, and α={α1,α2,,αNi+Nb} is the coefficient vector.

    The first and second order space derivatives of φ are derived by using F and Eq.(4.3) as

    φx=Fxα=FxF1φ,φy=Fyα=FyF1φ, (4.4)
    2φx2=x(φx)=2Fx2F1φ,2φy2=2Fy2F1φ. (4.5)

    Using Eqs.(4.4)-(4.5), Eq.(4.1) is expressed in matrix-vector form as

    D2φ=([u]dDx+[v]dDy)φ (4.6)
    or equivalently, (D2M)φ=0, (4.7)

    where the matrices are D2=(2F2x+2Fy2)F1, Dx=FxF1, Dy=FyF1 and M=([u]dDx+[v]dDy), and the subscript d refers to diagonal.

    The Dirichlet type boundary conditions ({dbc}) are inserted to the system matrix of Eq.(4.7) as

    [[D2M]Ni[D2M]Nb0I][φiφb]=[0dbc], (4.8)

    with the identity matrix I of size Nb×Nb. In case of Neumann type boundary conditions (nbc), the given boundary conditions are added to the system matrix of Eq.(4.7) with the help of Eq.(4.4) as

    [[D2M]NiDn][φiφb]=[0nbc], (4.9)

    in which Dn is either Dx or Dy with + or sign with respect to normal vector on the boundary.

    In order to handle time derivatives, differential quadrature method (DQM) is used. DQM approximates the derivatives of a function at a grid point by a linear summation of all functional values in the whole problem domain.

    The time interval is firstly considered as [0,tmax] with a maximum value of time, tmax. System matrices of the governing equations are formed by space and time domains as a box, and these matrices are not too large for small values of tmax. However, for large values of tmax, any system matrix become too large to compute. To overcome this difficulty for a large system, time is divided into equal time subintervals, and each subinterval is also divided into L number of nonuniform grid points.

    DQM manages the first order time derivative as

    φt=Lk=1alkφijk, (4.10)

    where alk is the first order weighting coefficient given explicitly in [18] as

    alk=M(1)(tl)(tltk)M(1)(tk),lk,all=Lk=1,klalk, (4.11)

    where M(1)(tl)=Lk=1,kl(tltk), and l=1,2,,L.

    The system in each time subinterval is regarded as a block consisting of space and time domain. Aniterative system at each block on the dimensionless governing equations are built as follows

    D2ψn+1ijl=ωnijl, (4.12a)
    un+1ijl=Dyψn+1ijl,vn+1ijl=Dxψn+1ijl, (4.12b)
    (D2αfαnf(At+M))Tn+1ijl=0, (4.12c)
    (D2ρnfμfPrρfμnf(At+M))wn+1ijl=μf(ρβ)nfμnfρfβfRaDxTn+1ijl, (4.12d)

    where M=[u]m+1dDx+[v]m+1dDy, At is the matrix of size L×L formed by weighting coefficient in Eq.(4.11), and n shows the iteration level.

    In the first block, ψ,T and ω are attained as zero at n=0 except the known boundary at n=0. Once this iteration is completed at a block, the next block starts with the initial values taken as the results of the previous block. This enables one to reach aimed value of tmax.

    The unknown vorticity boundary conditions are handled by using the definition of vorticity as

    ω=vxuy=Dxvn+1Dyun+1. (4.13)

    The resulting systems of equations in the form Ax=b are solved by Gaussian elimination with partial pivoting, and QR factorization which is for overdetermined system of temperature equation because of adiabatic boundary conditions.

    The average Nusselt number through the heated left wall is defined by ¯Nu=10Txdy, and computed by Clenshaw Curtis quadrature due to the usage of Chebyshev non-uniform grid distribution.

    Multiquadric (MQ) f=r2+c2 and inverse multiquadric (IMQ) RBFs f=1/r2+c2 are employed in this study. The shape parameter c controls the shape of the basis functions. As c gets larger, the shape becomes flat and the matrix becomes more ill-conditioned. MQ collocation matrices are conditionally positive definite [14], and the exponential convergence of the error of MQ approximation have been demonstrated by Madych et al. [11]. IMQs are strictly positive definite.

    In order to determine a suitable shape parameter depending on the problem parameters, an initial interval for the shape parameter where the iterative system converges (or catches the expected behaviour) is determined. This interval is divided into equal c values.

    The shape parameter value c providing ψmax zero or closest to zero is chosen. If more than one c values give ψmax=0, then the first c value giving 0 is taken.


    5. Numerical Results and Discussion

    The implementations and computations are done in Matlab. In implementation of space derivatives Eqs.(4.4)-(4.5), right back slash operator, which makes use of Gaussian elimination with partial pivoting, is used instead of taking the inverse of F directly.

    The base fluid is taken as pure water with Prandtl number 6.2.

    A relaxation parameter 0<γ<1 is used as φn+1γφn+1+(1γ)φn once eqs.(4.12a), (4.12c) or (4.12d) are performed. In particular, γ=0.1 is employed for Ra=103,Ra=104 and Ra=105 once eqs.(4.12a), (4.12c) and (4.12d) are solved, and γ=0.01 is carried out for Ra=106 once eq.(4.12d) is solved.

    tmax is fixed at 20 for AR=1, at 21 for AR1, and each block has an up time level L=5 for AR=1, L=3 for AR1.

    Table 2 describes the well agreement of the proposed scheme with the benchmark problem in [3] in which 41×41 number of grid points are used. Both MQ RBF and IMQ RBF are carried out, and c values are larger in IMQ than in MQ. In both cases, c values decrease as Ra increases.

    Table 2.Comparison of ¯Nu values.
    RaNb,Ni¯Nu (MQ)c (MQ)¯Nu (IMQ)c (IMQ)[3]
    10364, 2251.11810.1251.11570.191.12
    10464, 2252.24920.1452.24440.1552.243
    10580, 3614.53570.084.51260.14.52
    10696, 6258.87160.068.84510.088.8
     | Show Table
    DownLoad: CSV

    When there is no nanoparticle inside of the water, the average Nusselt number is ¯Nu=4.7062 in Ra=105 utilizing MQ RBF. As is seen from Table 3, the presence of any type of nanoparticle enhances heat transfer in the cavity. Also, the increase in solid volume fraction is resulted with the increase in ¯Nu. The improved heat transfer is pronounced better by the insertion of Cu nanoparticles than the other type of nanoparticles.

    Table 3.¯Nu with solid volume fraction variation in different nanofluids when Ra=105 with c=0.09.
    χCuAgAl2O3SiO2TiO2CuO
    0.044.91624.91574.86354.85904.81024.8992
    0.085.10975.10764.99994.99064.89435.0748
    0.165.45075.44365.21275.19255.00495.3771
    0.25.60135.59155.29045.26405.03205.5056
     | Show Table
    DownLoad: CSV

    In Figure 2, the variation in Ra is shown in Cu-water nanofluid with χ=0.2. Streamlines in the center of the cavity expands with the increase in Ra. Also, the strong temperature gradient in isotherms through the left and right walls is noticed due to the increase in buoyancy force. The centered vorticity contour is divided into two cells, and these two cells are shrunk through the left and right walls. At a large Rayleigh number (Ra=106), the circulation in the main centered cell of streamlines in case of χ=0 (Figure 3) is not much different than the case of χ=0.2 (Figure 2). This indicates that laminar flow regime is preserved.

    Figure 2. Streamlines, isotherms and vorticity contours in different Ra numbers as χ=0.2 is fixed with copper filled water.
    Figure 3. Contours in Ra=106 with χ=0.

    In Table 4, fluid velocity and ¯Nu values are presented in different concentration of Cu-water nanofluid. The values Nb,Ni and c are also the same for χ=0.02 and χ=0.2 with χ=0. In each concentration, the increase in fluid velocity as Ra increases is noted. Fluid velocity does not change from χ=0 to χ=0.02. However, the remarkable change occurs in χ=0.2. This also points out that fluid flows faster in highly concentrated Cu-water nanofluid than pure water. For small Ra values (Ra=103,104), the change in χ has no much effect on fluid velocity. For Ra=105 and Ra=106, the increase in |ψ|max and umax are seen in presence of nanoparticle in solid volume fraction χ=0.2. In each Ra, the existence of nanoparticle inside water causes ¯Nu to increase, and so the convective heat transfer is pronounced.

    Table 4.With and without Cu inside water.
    χ=0
    Ra|ψ|maxumaxvmax¯NuNb,Nic
    1031.17833.61383.68441.113964,2250.14
    1045.238716.348119.98082.263364,2250.14
    10511.214939.518173.66974.706280,3610.09
    10620.320999.2457228.86669.169096,6250.05
    χ=0.02
    1031.12863.46173.51711.1598
    1045.284716.418920.11982.3066
    10511.492339.430074.20294.8133
    10620.947298.5315229.169511.2539
    χ=0.2
    1030.67772.08022.08471.7668
    1044.933215.069515.91732.6060
    10513.890244.167470.83335.6013
    10626.5623104.7299237.696511.2539
     | Show Table
    DownLoad: CSV

    The increase in fluid velocity with the increase in χ is also illustrated in Figure 4 with u and v velocity profiles at mid sections of the unit square cavity. The fluid moves at higher velocities close to the boundaries than the center of the cavity as can be seen in Figure 4(a)-(b). In each cases, absolute value of velocities is the largest at the highest concentration of nanofluids χ=0.2.

    Figure 4. Velocity profiles for Cu-water nanofluid when Ra=105 in different χ.

    In Figures 5 and 6, different aspect ratios of the Cu-water-filled cavity are taken into account performing IMQ RBF. Similar behaviour as in Ra=105 in Figure 2 occurs. Table 5 indicates the average Nusselt number values through the heated wall. On one hand, convective heat transfer decreases inside the cavity when the cavity sizes increase. On the other hand, ¯Nu values in case of 0<AR<1 are greater than the case of AR>1 which may be a result of the effect of longer heated wall on the fluid flow and heat transfer inside the cavity.

    Figure 5. IMQ in different AR>1 with Ra=105,χ=0.2.
    Figure 6. IMQ in different 0<AR<1 with Ra=105,χ=0.2.
    Table 5.¯Nu values in different AR when Ra=105,χ=0.2.
    W:H¯Nuc
    AR>12:14.790.17
    4:13.510.15
    0<AR<11:25.440.12
    1:44.770.11
     | Show Table
    DownLoad: CSV

    6. Conclusion

    In this study, a numerical investigation on natural convection in a cavity filled with a nanofluid is presented. The space derivatives in the dimensionless equations are discretised by RBF-PS and the time derivatives are handled by DQM. The nature of this approach provides one to obtain the results with small number of grid points. Instead of solving the system of time and space as an entire large system, time is divided into subintervals and the system is observed block by block. Also, the solution at each block is obtained at once. Convective heat transfer increase with the insertion of nanoparticles into the water. Further, the decrease in the convective heat transfer is noticed with the increase in Cu-water-filled cavity size.


    Conflict of Interest

    I have no financial or other relationship with other people or organizations in my study.


    [1] H. R Ashorynejad, A. A. Mohamad, and M. Sheikholeslami,Magnetic field e ects on naturalconvection flow of a nanofluid in a horizontal cylindrical annulus using Lattice Boltzmann method ,Int. J. Therm. Sci., 64 (2013), 240-250.
    [2] H. C. Brinkman, The viscosity of concentrated suspensions and solutions, J. Chem. Phys., 3 (1952),571-581.
    [3] G. de vahl Davis, Natural convection of air in a square cavity : A bench mark numerical solution,Int. J. Numer. Meth. Fl., 3 (1983), 249-264.
    [4] G. Fasshauer, Meshfree Approximation Methods with Matlab, World Scientific Publications, Sin-gapore, 2007.
    [5] E. Fattahi, M. Farhadi, K. Sedighi, and H. Nemati, Lattice Boltzmann simulation of natural con-vection heat transfer in nanofluids, Int. J. Therm. Sci., 52 (2012), 137-144.
    [6] B. Ghasemi, S. M. Aminossadati, and A. Raisi, Magnetic field e ect on natural convection in ananofluid-filled square enclosure, Int. J. Therm. Sci., 50 (2011), 1748-1756.
    [7] S. Gumgum and M. Tezer-Sezgin, DRBEM solution of natural convection flow of nanofluids with a heat source, Eng. Ana. Bound., 34 (2010), 727-737.
    [8] M. Jahanshahi, S. E. Hosseinizadeh, M. Alipanah, A. Dehghani, and G. R. Vakilinejad, Numerical simulation of free convection based on experimental measured conductivity in a square cavity using Water/SiO2 nanofluid, Int. Comm. Heat Mass, 37 (2010), 687-694.
    [9] R. Y. Jou and S. C. Tzeng, Numerical research of nature convective heat transfer enchancement filled with nanofluids in rectangular enclosures, Int. J. Heat Mass Tran., 33 (2006), 727-736.
    [10] K. Khanafer, K. Vafai, and M. Lightstone, Buoyancy-driven heat transfer enhancement in a two-dimensional enclosure utilizing nanofluids, Int. J. Heat Mass Tran., 46 (2003), 3639-3653.
    [11] W. R. Madych and S. A. Nelson, Bounds on multivariate polynomials and exponential errorestimates for multiquadric interpolation, J. Approx. Theory, 70 (1992), 94-114.
    [12] A. H. Mahmoudi, I. Pop, M. Shahi, and F. Talebi, MHD natural convection and entropy generation in a trapezoidal enclosure using Cu-water nanofluid, Comput. Fluids, 72 (2013), 46-62.
    [13] J. C. Maxwell-Garnett, Colors in metal glasses and in metallic films, Phil. Trans R. Soc. A, 203(1904), 385-420.
    [14] C. A. Michelli, Interpolation of scattered data: Distance matrices and conditionally positivedefinite functions, Constr. Approx., 2 (1986), 11-22.
    [15] M. Muthtamilselvan, P. Kandaswamy, and J. Lee, Heat transfer enhancement of copper-waternanofluids in a lid-driven enclosure, Commun. Nonlinear Sci. Numer. Simulat.,15 (2010), 1501-1510.
    [16] J. Serna, F. J. S. Velasco, and A. S. Meca, Application of network simulation method to viscousflow: The nanofluid heated lid cavity under pulsating flow, Comput. Fluids, 91 (2014), 10-20.
    [17] M. Shahi, A. H. Mahmoudi, and A. H. Raouf, Entropy generation due to natural convection coolingof nanofluid, Int. Comm. Heat Mass, 38 (2011), 972-983.
    [18] C. Shu, Di erential quadrature and its application in engineering, Springer-Verlag, 2000.
    [19] R. K. Tiwari and M. K. Das, Heat transfer augmentation in a two-sided lid-driven differentially heated square cavity utilising nanofluids, Int. J. Heat Mass Tran., 50 (2007), 2002-2018
  • This article has been cited by:

    1. Bengisen Pekmen Geridönmez, NUMERICAL SIMULATION OF NATURAL CONVECTION IN A POROUS CAVITY FILLED WITH FERROFLUID IN PRESENCE OF MAGNETIC SOURCE, 2017, 4, 2148-7847, 1756, 10.18186/journal-of-thermal-engineering.369169
    2. Bengisen Pekmen Geridonmez, Free Convection in a Wavy Walled Cavity With a Magnetic Source Using Radial Basis Functions, 2019, 141, 0022-1481, 10.1115/1.4042782
    3. Fahad Alsharari, Mohamed M. Mousa, New application of MOL-PACT for simulating buoyancy convection of a copper-water nanofluid in a square enclosure containing an insulated obstacle, 2022, 7, 2473-6988, 20292, 10.3934/math.20221111
    4. Kottakkaran Sooppy Nisar, Muhammad Shoaib, Muhammad Asif Zahoor Raja, Yasmin Tariq, Ayesha Rafiq, Ahmed Morsy, Design of neural networks for second-order velocity slip of nanofluid flow in the presence of activation energy, 2023, 8, 2473-6988, 6255, 10.3934/math.2023316
  • Reader Comments
  • © 2016 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(5894) PDF downloads(1281) Cited by(4)

Article outline

Figures and Tables

Figures(6)  /  Tables(5)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog