MAISRN Mathematical Analysis2090-46652090-4657International Scholarly Research Network93783010.5402/2011/937830937830Research ArticleAnalytical Approximate Solution of Nonlinear Differential Equation Governing Jeffery-Hamel Flow with High Magnetic Field by Adomian Decomposition MethodGanjiD. D.1SheikholeslamiM.1AshorynejadH. R.1GarceaG.MuchaP. B.1Faculty of Mechanical EngineeringBabol Noshirvani University of TechnologyMazandaran P.O. Box 484, Babol 47148-71167Irannit.ac.ir201131072011201121022011070420112011Copyright © 2011 D. D. Ganji et al.This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The magnetohydrodynamic Jeffery-Hamel flow is studied analytically. The traditional Navier-Stokes equation of fluid mechanics and Maxwell's electromagnetism governing equations reduce to nonlinear ordinary differential equations to model this problem. The analytical tool of Adomian decomposition method is used to solve this nonlinear problem. The velocity profile of the conductive fluid inside the divergent channel is studied for various values of Hartmann number. Results agree well with the numerical (Runge-Kutta method) results, tabulated in a table. The plots confirm that the method used is of high accuracy for different α, Ha, and Re numbers.

1. Introduction

The flow of fluid through a divergent channel is called Jeffery-Hamel flow since introducing this problem by Jeffery  and Hamel  in 1915 and 1916, respectively. On the other hand, the term of magnetohydrodynamic (MHD) was first introduced by Bansal  in 1994. The theory of magnetohydrodynamics is inducing current in a moving conductive fluid in presence of magnetic field; such induced current results in force on ions of the conductive fluid. The theoretical study of magnetohydrodynamic (MHD) channel has been a subject of great interest due to its extensive applications in designing cooling systems with liquid metals, MHD generators, accelerators, pumps, and flow meters .

In fluid mechanics most of the problems are nonlinear. It is very important to develop efficient methods to solve them. Up to now, it is very difficult to obtain analytical approximations of nonlinear partial differential equations even though there are high-performance computers and computation software. The small disturbance stability of Magnetohydrodynamic stability of plane Poiseuille flow has been investigated by Makinde and Motsa  and Makinde  for generalized plane Couette flow. Their results show that magnetic field has stabilizing effects on the flow. Considerable efforts have been done to study the MHD theory for technological application of fluid pumping system in which electrical energy forces the working conductive fluid. Damping and controlling of electrically conducting fluid can be achieved by means of an electromagnetic body force (Lorentz force) produced by the interaction of an applied magnetic field and an electric current that usually is externally supplied. Harada et al.  studied the fundamental characteristics of linear Faraday MHD theoretically and numerically. In 2005, Anwari et al.  continued the Harada et al.  work numerically and theoretically, for various loading configurations. Jang and Lee  emphasized on the idea that, in such problems, the moving ions drag the bulk fluid with themselves and such MHD system induces continues pumping of conductive fluid without any moving part. Homsy et al.  worked and developed the same idea mentioned above. The purpose of the current work is to study the mechanics of the fluid through a divergent channel in presence of electromagnetic field (Figure 1).

Geometry of the MHD Jeffery-Hamel flow.

The Adomian decomposition method (ADM) is used to solve a wide range of physical problems. One of the semiexact methods which does not need linearization or discretization is Adomian decomposition method, and several modifications have improved its ability .

An advantage of this method is that it can provide analytical approximation or an approximated solution to a rather wide class of nonlinear (and stochastic) equations without linearization, perturbation, closure approximation, or discretization methods. Unlike the common methods, that is, weak nonlinearity and small perturbation which change the physics of the problem due to simplification, ADM gives the approximated solution of the problem without any simplification. Thus, its results are more realistic . ADM abilities have attracted many authors to use this method for solving fluid dynamic problems.

Jafari and Daftardar-Gejji  presented a modified ADM to solve a system of nonlinear equations, which yielded a series of solutions with faster accelerated convergence than the series obtained by the standard ADM. Bulut et al.  studied viscous incompressible flow through orifice. Allen and Syam  and Wang  investigated nonhomogeneous and classical Blasius equation by ADM. Soh  applied ADM to solve thin film equation. Hashim  presented the Adomian decomposition method for solving BVPs for fourth-order integrodifferential equations and the Blasius equation . Kechil and Hashim  presented a nonperturbative solution of free-convective boundary-layer equation by ADM. Chang  presented a decomposition solution for fins with temperature-dependent surface heat flux. Arslanturk  inspected on the fins efficiency of convective straight fins with temperature-dependent thermal conductivity using the decomposition method. ADM also has been used by several researchers to solve a wide range of physical problems in various engineering fields such as fluid flow and porous media simulation  and other nonlinear systems . In recent years some researchers used new methods to solve these kinds of problems .

In this paper, we have applied ADM to find the approximate solutions of nonlinear differential equations governing the MHD Jeffery-Hamel flow, and a comparison between the results and the numerical solution has been provided. The numerical results of this problem are done using Maple 12.

2. Governing Equations

Consider a system of cylindrical polar coordinates (r,θ,z) which steady two-dimensional flow of an incompressible conducting viscous fluid from a source or sink at channel walls, lie in planes, and intersect in the axis of z, assuming purely radial motion which means that there is no change in the flow parameter along the z direction. The flow depends on  r, θ, and further assume that there is no magnetic field in the z-direction. In the reduced form of continuity, Navier-Stokes and Maxwell’s equations are ρrr(ru(r,θ))=0,u(r,θ)u(r,θ)r=-1ρPr+υ[2u(r,θ)r2+1ru(r,θ)r+1r22u(r,θ)θ2-u(r,θ)r2]-σB02ρr2u(r,θ),1ρrPθ-2υr2u(r,θ)θ=0, where B0 is the electromagnetic induction, σ the conductivity of the fluid, u(r) the velocity along radial direction, P the fluid pressure, υ the coefficient of kinematic viscosity, and ρ the fluid density. Considering uθ=0 for purely radial flow, one can define the velocity parameter as f(θ)=ru(r). Introducing the η=θ/α as the dimensionless degree, the dimensionless form of the velocity parameter can be obtained by dividing that to its maximum value asf(η)=f(θ)fmax. Substituting  (2.5) into (2.2) and (2.3) and eliminating P, one can obtain the ordinary differential equation for the normalized function profile as  f′′′(η)+2αRef(η)f(η)+(4-Ha)α2f(η)=0, with the following reduced form of boundary conditionf(0)=1,  f(0)=0,f(1)=0. We introduce the Reynolds number and the Hartmann number based on the electromagnetic parameter as follows, respectively:Re=fmaxαυ=Umaxrαυ(divergent-channel: α>0,fmax>0convergent-channel: α<0,fmax<0),Ha=σB02ρυ.

Consider equation Fu(t)=g(t), where F represents a general nonlinear ordinary or partial differential operator including both linear and nonlinear terms. The linear terms are decomposed into L+R, where L is easily invertible (usually the highest order derivative) and R is the remainder of the linear operator. Thus, the equation can be written as :Lu+Nu+Ru=g, where Nu indicates the nonlinear terms. By solving this equation for Lu, since L is invertible, we can writeL-1Lu=L-1g-L-1Ru-L-1Nu. If L is a second-order operator, L-1 is twofold in definite integral. By solving (3.2), we have u=A+Bt+L-1g-L-1Ru-L-1Nu, where A and B are constants of integration and can be found from the boundary or initial conditions. Adomian method assumes that the solution u can be expanded into infinite series asu=n=0un. Also, the nonlinear term Nu will be written asNu=n=0An, where An are the special Adomian polynomials. By specified An, the next component of u can be determined: un+1=L-1n=0nAn. Finally, after some iteration and getting sufficient accuracy, the solution of the equation can be expressed by (3.5).

In (3.6), the Adomian polynomials can be generated by several means. Here we used the following recursive formulation : An=1n![dndλn[N(i=0nλi  ui)]]λ=0,n=0,1,2,3,. Since the method does not resort to linearization or assumption of weak nonlinearity, the solution generated is in general more realistic than those achieved by simplifying the model of the physical problem.

4. Application

According to (3.1), (2.6) must be written as follows:Lf=-2Reαff-(4-Ha)α2f, where the differential operator L is given by L=d3/dη3. Assume that the inverse of the operator L exists and it can be integrated from 0 to η, that is, L-1=0η(·)dη  dη  dη.

Operating with L-1 on (4.1) and after exerting boundary condition on it, we havef(η)=f(0)+f(0)η+f′′(0)  η22+L-1(Nu), where Nu=-2Reαf(η)f(η)-(4-Ha)α2f(η). ADM introduced the following expression:f(η)=m=0fm(η),f(η)=m=0fm=f0+L-1(Nu). To determine the components of fm(η) the f0(η) is defined by applying the boundary condition of (2.7), and by assuming f′′(0)=β, f0(η)=1+β  η22,f1(η)=(-1120)αReβ2η6-14(13αReβ+16(4-Ha)α2β)η4-1120(4-Ha)α2βη5-16(4-Ha)α2η3,f2(η)=(110800)α2Re2β3η10+17560α3Re(4-Ha)β2η-18(121αRe(-112αReβ-124(4-Ha)α2β)β-1420α2Re2+142α2Reβ(-13αReβ-16(4-Ha)α2β)-1840α3Re(4-Ha)β2)η8-17(-11360α3Re(4-Ha)β-1720(4-Ha)2α4β)η7-16(110αRe(-13αReβ-16(4-Ha)α2β)+120(4-Ha)α2(-13αReβ-16(4-Ha)α2β))η6-15(-112α3Re(4-Ha)-1720(4-Ha)2α4)η5+140320(4-Ha)2α4βη8-1120(4-Ha)α2(-112αReβ-124(4-Ha)α2β)η7+1720(4-Ha)2α4η6

and f3(η),f4(η), can be determined in a similar way from (4.5).

Using f(η)=m=0fm(η)=f0(η)+f1(η)+f2(η)+f3(η)+, thusf(η)=1+βη22+(-1120)αReβ2η6-14(13αReβ+16(4-Ha)α2β)η4-124(4-Ha)α2βη4-12(4-Ha)α2η2+(110800)α2Re2β3η10-18(121αRe(-112αReβ-112(4-Ha)α2β)β-1420α2Re2β2+142αReβ(-13αReβ-13(4-Ha)α2β)-1840(4-Ha)α2β2Re)η8-16(-110α3Re(4-Ha)β+110αRe(-13αβRe-13(4-Ha)α2β))η6-14(-13α3Re(4-Ha)-16(4-Ha)2α4)η4+16720(4-Ha)α3Reβ3η8-130(4-Ha)α2(-112αβRe-16(4-Ha)2α2β)η6-124(4-Ha)2α4η4+. According to (4.8), the accuracy of ADM solution increases by increasing the number of solution terms (m). For the complete solution of (4.8), β should be determined, with boundary condition of f′′(0)=β.

5. Results and Discussion

In this study the objective was to apply Adomian decomposition method to obtain an explicit analytic solution of the MHD Jeffery-Hamel problem. The magnetic field acts as a control parameter such as the flow Reynolds number and the angle of the walls, in MHD Jeffery-Hamel problems. There is an additional nondimensional parameter that determines the solutions, namely, the Hartmann number. Table 1 shows the value of constant β for different α, Ha, and Re numbers at the divergent channel.

Value of f(0)=β at various Re, Ha, and α.

 Re α Ha = 100 Ha = 200 Ha = 300 f′′(0)=β f′′(0)=β f′′(0)=β 25 2.5 −2.42483 −2.54507 −2.66789 5 −3.09259 −3.57685 −4.09816 7.5 −3.98607 −5.14061 −6.44257 50 2.5 −2.77059 −2.87965 −2.99167 5 −3.85619 −4.25077 −4.69303 7.5 −5.16973 −6.04706 −7.13483

For comparison, a few limited cases of the ADM solutions are compared with the numerical results. The comparison between the numerical results and ADM solution for velocity when Re=25 and α=5° is shown in Table 2. The error bar shows an acceptable agreement between the results observed, which confirms the validity of the ADM. In these tables the error is introduced as follows:%Error=|f(η)NM-f(η)ADM|f(η)ADM. %Error of ADM for f(η) at different steps when α=5, Re=50, and Ha=100 can be seen in Figure 2; it shows that at first we have large error then it decreases and after 12 steps the error becomes minimized.

Comparison between the numerical results and ADM solution for velocity when Re=25, α=5.

 η Ha = 0 Ha = 100 Ha = 250 Ha = 500 Numerical ADM %Error Numerical ADM Error Numerical ADM Error Numerical ADM %Error 0 1 1 0 1 1 0 1 1 0 1 1 0 0.1 0.986671 0.9866365 0.000035 0.987488 0.9867671 0.00073 0.988606 0.9870331 0.001591 0.99022 0.992695 0.002475 0.2 0.947258 0.9471292 0.000136 0.950396 0.9477282 0.002807 0.9547 0.9488372 0.006141 0.960933 0.970544 0.009611 0.3 0.883419 0.8831778 0.000273 0.890009 0.884749 0.00591 0.899076 0.8873853 0.013003 0.912273 0.912273 0.006304 0.4 0.797697 0.7973476 0.000438 0.808288 0.8005834 0.009532 0.822925 0.8132968 0.0117 0.844383 0.832683 0.009146 0.5 0.693233 0.6928205 0.000595 0.70764 0.6984237 0.013024 0.727664 0.7175749 0.013865 0.757286 0.743421 0.010234 0.6 0.573424 0.5730174 0.000709 0.590631 0.5813941 0.015639 0.614709 0.6086584 0.009843 0.650719 0.643816 0.006902 0.7 0.441593 0.4412649 0.000743 0.459695 0.4520834 0.016558 0.485232 0.4810561 0.008606 0.523909 0.515303 0.018548 0.8 0.300674 0.300475 0.000662 0.316855 0.3121316 0.014907 0.33989 0.3351125 0.014056 0.37529 0.361234 0.013709 0.9 0.152979 0.1529137 0.000427 0.163467 0.1618745 0.009742 0.178555 0.1772346 0.007395 0.202125 0.19473 0.009368 1 0 0 0 0 0 0 0 0 0 0 0 0

%Error of ADM for f(η) at different steps when α=5, Re=50, and Ha=100.

Figures 3, 4, and 5 show the magnetic field effect on the velocity profiles for divergent channels. There are good agreements between the numerical solution obtained by the fourth-order Runge-Kutta method and the differential transformation method.

The ADM solution for velocity in divergent channel for α=2.5°,  Re=200.

The ADM solution for velocity in divergent channel for α=5°,  Re=200.

The ADM solution for velocity in divergent channel for α=7.5°,  Re=200.

Under magnetic field the Lorentz force effect is in opposite of the momentum’s direction that stabilizes the velocity profile.

The results show moderate increases in the velocity with increasing Hartmann numbers at small angle (α=2.5°) and differences between velocity profiles are more noticeable at greater angles. Backflow is excluded in converging channels  but it may occur for large Reynolds numbers in diverging channels. For specified opening angle, after a critical Reynolds number, we observe that separation and backflow are started.

Figures 68 show the magnetic field effects at constant α and different Reynolds numbers. At α=5°, Re=75 with increasing Hartmann number the velocity profile becomes flat and thickness of boundary layer decreases, but at this Reynolds number no backflow is observed as shown in Figure 6. It can be seen in Figure 9 that without magnetic field at α=5°, Re=150 the backflow starts and that with increasing Hartmann number this phenomenon eliminates.

The ADM solution for velocity in divergent channel for α=5°,  Re=75.

The ADM solution for velocity in divergent channel for α=5°,  Re=225.

The ADM solution for velocity in divergent channel for α=5°,  Re=300.

The ADM solution for velocity in divergent channel for α=5°,  Re=150.

By increasing Reynolds number the backflow expands and so greater magnetic field is needed in order to eliminate it. As shown in Figures 7 and 8 at α=5°, Re=225 the back flow eliminates at Ha=1000 while at α=5°, Re=300 this occurs at Ha=2000.

6. Conclusion

In this paper, magnetohydrodynamic Jeffery-Hamel flow has been solved via a sort of analytical method, Adomian decomposition method (ADM). Also this problem is solved by a numerical method (the Runge-Kutta method of order 4), and some conclusions are summarized as follows.

Adomian decomposition method is a powerful approach for solving MHD Jeffery-Hamel flow in high magnetic field, and it can be observed that there is a good agreement between the present and numerical results.

Increasing Hartmann number will lead to backflow reduction. In greater angles or Reynolds numbers high Hartmann number is needed for the reduction of backflow.

NomenclatureB0:

Magnetic field (wb·m2)

β:

Constant

Lu:

Linear term

Nu:

Nonlinear term

Ru:

The remainder of linear operator

F:

General nonlinear operator

A:

f(η):

Dimensionless velocity

Ha:

Hartmann number

ρ:

Density

P:

Pressure term

Re:

Reynolds number

r,  θ:

Cylindrical coordinates

Umax:

Maximum value of velocity

u,  v:

Velocity components along x-, y-axes, respectively.

Greek Symbolsυ:

Kinematic viscosity

τ:

Viscous stresses

α:

Angle of the channel

θ:

Any angle

η:

Dimensionless angle.

JefferyG. B.The two-dimensional steady motion of a viscous fluidPhilosophical Magazine19156455465HamelG.Spiralförmige Bewgungen Zäher FlüssigkeitenJahresbericht der Deutschen Mathematiker-Vereinigung1916253460BansalL.Magnetofluiddynamics of Viscous Fluids1994Jaipur, IndiaJaipur Publishing HouseChaJ. E.AhnY. C.KimM. H.Flow measurement with an electromagnetic flowmeter in two-phase bubbly and slug flow regimesFlow Measurement and Instrumentation2002125-63293392-s2.0-003574204510.1016/S0955-5986(02)00007-9TendlerM.Confinement and related transport in extrap geometryNuclear Instruments and Methods In Physics Research19832071-22332402-s2.0-49049127666MossinoJ.Some nonlinear problems involving a free boundary in plasma physicsJournal of Differential Equations19793411141382-s2.0-49249148285NijsingR.EiflerW.A computational analysis of transient heat transfer in fuel rod bundles with single phase liquid metal coolingNuclear Engineering and Design1980621–339682-s2.0-0019300352MakindeO. D.MotsaS. S.Hydromagnetic stability of plane poiseuille flow using chebyshev spectral collocation methodJournal of Institute of Mathematics and Computer Sciences2001122175183MakindeO. D.Magneto-hydrodynamic stability of plane-Poiseuille flow using multi-deck asymptotic techniqueMathematical and Computer Modelling2003373-42512592-s2.0-0037371525HaradaN.IkewadaJ.TerasakiY.Basic studies on an MHD acceleratorThe American Institute of Aeronautics and Astronautics20022175AnwariM.HaradaN.TakahashiS.Performance study of a magnetohydrodynamic accelerator using air-plasma as working gasEnergy Conversion and Management20054615-16260526132-s2.0-1774440029810.1016/j.enconman.2004.12.003JangJ.LeeS. S.Theoretical and experimental study of MHD (magnetohydrodynamic) micropumpSensors and Actuators200080184892-s2.0-003388255410.1016/S0924-4247(99)00302-7HomsyA.KosterS.EijkelJ. C. T.Van Den BergA.LucklumF.VerpoorteE.De RooijN. F.A high current density DC magnetohydrodynamic (MHD) micropumpThe Royal Society of Chemistry, Lab Chip2005544664712-s2.0-1714442812010.1039/b417892kAdomianG.A review of the decomposition method in applied mathematicsJournal of Mathematical Analysis and Applications198813525015442-s2.0-0041185368JafariH.Daftardar-GejjiV.Revised Adomian decomposition method for solving systems of ordinary and fractional differential equationsApplied Mathematics and Computation200618115986082-s2.0-3374951870210.1016/j.amc.2005.12.049GhoshS.RoyA.RoyD.royd@civil.iisc.ernet.inAn adaptation of Adomian decomposition for numeric-analytic integration of strongly nonlinear and chaotic oscillatorsComputer Methods in Applied Mechanics and Engineering20071964–61133115310.1016/j.cma.2006.08.010AdomianG.A review of the decomposition method in applied mathematicsJournal of Mathematical Analysis and Applications198813525015442-s2.0-0041185368JafariH.Daftardar-GejjiV.Revised Adomian decomposition method for solving a system of nonlinear equationsApplied Mathematics and Computation20061751172-s2.0-3364489657910.1016/j.amc.2005.07.010BulutH.ErgutM.AsilV.BokorR. H.Numerical solution of a viscous incompressible flow problem through an orifice by Adomian decomposition methodApplied Mathematics and Computation200415337337412-s2.0-294256960810.1016/S0096-3003(03)00667-2AllanF. M.SyamM. I.m.syam@uaeu.ac.aeOn the analytic solutions of the nonhomogeneous Blasius problemJournal of Computational and Applied Mathematics2005182236237110.1016/j.cam.2004.12.017WangL. A new algorithm for solving classical Blasius equationApplied Mathematics and Computation200415711910.1016/j.amc.2003.06.011SohC. W.wafosoh@yahoo.comNon-perturbative semi-analytical source-type solutions of thin-film equationApplied Mathematics and Computation200617421576158510.1016/j.amc.2005.07.006HashimI.Adomian decomposition method for solving BVPs for fourth-order integro-differential equationsJournal of Computational and Applied Mathematics200619326586642-s2.0-3364624761110.1016/j.cam.2005.05.034HashimI.ishak_h@ukm.myComments on "a new algorithm for solving classical Blasius equationThe Journal of Computational and Applied Mathematics200518236237110.1016/j.amc.2005.10.016KechilS. A.HashimI.Non-perturbative solution of free-convective boundary-layer equation by Adomian decomposition methodPhysics Letters A20073631-21101142-s2.0-3384664078710.1016/j.physleta.2006.11.054ChangM. H.A decomposition solution for fins with temperature dependent surface heat fluxInternational Journal of Heat and Mass Transfer2005489181918242-s2.0-1764439308810.1016/j.ijheatmasstransfer.2004.07.049ArslanturkC.A decomposition method for fin efficiency of convective straight fins with temperature-dependent thermal conductivityInternational Communications in Heat and Mass Transfer20053268318412-s2.0-1814442545910.1016/j.icheatmasstransfer.2004.10.006AllanF. M.SyamM. I.On the analytic solutions of the nonhomogeneous Blasius problemJournal of Computational and Applied Mathematics200518223623712-s2.0-1964439486510.1016/j.cam.2004.12.017KayaD.YokusA.A decomposition method for finding solitary and periodic solutions for a coupled higher-dimensional Burgers equationsApplied Mathematics and Computation200516438578642-s2.0-1634437905310.1016/j.amc.2004.06.012PamukS.Solution of the porous media equation by Adomian's decomposition methodPhysics Letters A20053442–41841882-s2.0-2404447846210.1016/j.physleta.2005.06.068M AllanF.Al-KhaledK.An approximation of the analytic solution of the shock wave equationJournal of Computational and Applied Mathematics200619223013092-s2.0-3364608600010.1016/j.cam.2005.05.009Daftardar-GejjiV.JafariH.An iterative method for solving nonlinear functional equationsJournal of Mathematical Analysis and Applications200631627537632-s2.0-2984444230410.1016/j.jmaa.2005.05.009LesnicD.Decomposition methods for non-linear, non-characteristic Cauchy heat problemsCommunications in Nonlinear Science and Numerical Simulation20051065815962-s2.0-1114426590710.1016/j.cnsns.2004.02.002ZhuY.ChangQ.WuS.A new algorithm for calculating Adomian polynomialsApplied Mathematics and Computation200516914024162-s2.0-2714446947010.1016/j.amc.2004.09.082LuoX. G.A two-step Adomian decomposition methodApplied Mathematics and Computation200517015705832-s2.0-2684457298310.1016/j.amc.2004.12.010ZhangX.A modification of the Adomian decomposition method for a class of nonlinear singular boundary value problemsJournal of Computational and Applied Mathematics200518023773892-s2.0-1864437282010.1016/j.cam.2004.11.007KayaD.YokusA.A comparison of partial solutions in the decomposition method for linear and nonlinear partial differential equationsMathematics and Computers in Simulation20026065075122-s2.0-003710789910.1016/S0378-4754(01)00438-4AxfordW. I.The mahnetohydrodynamic Jeffery-Hamel problem for a weakly conducting fluidThe Quarterly Journal of Mechanics and Applied Mathematics196114335351EsmaeilpourM.GanjiD. D.Solution of the Jeffery-Hamel flow problem by optimal homotopy asymptotic methodComputers and Mathematics with Applications20105911340534112-s2.0-7795312892410.1016/j.camwa.2010.03.024GanjiZ. Z.GanjiD. D.EsmaeilpourM.Study on nonlinear Jeffery-Hamel flow by He's semi-analytical methods and comparison with numerical resultsComputers and Mathematics with Applications20095811-12210721162-s2.0-7035057407810.1016/j.camwa.2009.03.044MakindeO. D.MhoneP. Y.Temporal stability of small disturbances in MHD Jeffery-Hamel flowsComputers and Mathematics with Applications20075311281362-s2.0-3404727266210.1016/j.camwa.2006.06.014GanjiZ. Z.GanjiD. D.EsmaeilpourM.Study on nonlinear Jeffery-Hamel flow by He's semi-analytical methods and comparison with numerical resultsComputers and Mathematics with Applications20095811-12210721162-s2.0-7035057407810.1016/j.camwa.2009.03.044