Research article Special Issues

A neural networks technique for analysis of MHD nano-fluid flow over a rotating disk with heat generation/absorption

  • Received: 03 August 2024 Revised: 12 October 2024 Accepted: 04 November 2024 Published: 14 November 2024
  • MSC : 34G20, 35A20, 35A22, 35R11

  • In this paper, the neural network domain with the backpropagation Levenberg-Marquardt scheme (NNB-LMS) is novel with a convergent stability and generates a numerical solution of the impact of the magnetohydrodynamic (MHD) nanofluid flow over a rotating disk (MHD-NRD) with heat generation/absorption and slip effects. The similarity variation in the MHD flow of a viscous liquid through a rotating disk is explained by transforming the original non-linear partial differential equations (PDEs) to an equivalent non-linear ordinary differential equation (ODEs). Varying the velocity slip parameter, Hartman number, thermal slip parameter, heat generation/absorption parameter, and concentration slip parameter, generates a Prandtl number using the Runge-Kutta 4th order method (RK4) numerical technique, which is a dataset for the suggested (NNB-LMS) for numerous MHD-NRD scenarios. The validity of the data is tested, and the data is processed and properly tabulated to test the exactness of the suggested model. The recommended model was compared for verification, and the estimation solutions for particular instances were assessed using the NNB-LMS training, testing, and validation procedures. A regression analysis, a mean squared error (MSE) assessment, and a histogram analysis were used to further evaluate the proposed NNB-LMS. The NNB-LMS technique has various applications such as disease diagnosis, robotic control systems, ecosystem evaluation, etc. Some statistical data such as the gradient, performance, and epoch of the model were analyzed. This recommended method differs from the reference and suggested results, and has an accuracy rating ranging from 1009to 1012.

    Citation: Yousef Jawarneh, Humaira Yasmin, Wajid Ullah Jan, Ajed Akbar, M. Mossa Al-Sawalha. A neural networks technique for analysis of MHD nano-fluid flow over a rotating disk with heat generation/absorption[J]. AIMS Mathematics, 2024, 9(11): 32272-32298. doi: 10.3934/math.20241549

    Related Papers:

    [1] Fahad Alsharari, Ahmed O. M. Abubaker, Islam M. Taha . On r-fuzzy soft γ-open sets and fuzzy soft γ-continuous functions with some applications. AIMS Mathematics, 2025, 10(3): 5285-5306. doi: 10.3934/math.2025244
    [2] Jinguo Zhang, Dengyun Yang, Yadong Wu . Existence results for a Kirchhoff-type equation involving fractional p(x)-Laplacian. AIMS Mathematics, 2021, 6(8): 8390-8403. doi: 10.3934/math.2021486
    [3] Cenap Ozel, M. A. Al Shumrani, Aynur Keskin Kaymakci, Choonkil Park, Dong Yun Shin . On δb-open continuous functions. AIMS Mathematics, 2021, 6(3): 2947-2955. doi: 10.3934/math.2021178
    [4] Ibtesam Alshammari, Islam M. Taha . On fuzzy soft β-continuity and β-irresoluteness: some new results. AIMS Mathematics, 2024, 9(5): 11304-11319. doi: 10.3934/math.2024554
    [5] Sang-Eon Han . Semi-topological properties of the Marcus-Wyse topological spaces. AIMS Mathematics, 2022, 7(7): 12742-12759. doi: 10.3934/math.2022705
    [6] Samirah Alzahrani, A. A. Nasef, N. Youns, A. I. EL-Maghrabi, M. S. Badr . Soft topological approaches via soft γ-open sets. AIMS Mathematics, 2022, 7(7): 12144-12153. doi: 10.3934/math.2022675
    [7] Nicky K. Tumalun, Philotheus E. A. Tuerah, Marvel G. Maukar, Anetha L. F. Tilaar, Patricia V. J. Runtu . An application of generalized Morrey spaces to unique continuation property of the quasilinear elliptic equations. AIMS Mathematics, 2023, 8(11): 26007-26020. doi: 10.3934/math.20231325
    [8] Dina Abuzaid, Samer Al-Ghour . Supra soft Omega-open sets and supra soft Omega-regularity. AIMS Mathematics, 2025, 10(3): 6636-6651. doi: 10.3934/math.2025303
    [9] Syafrizal Sy, Rinovia Simanjuntak, Tamaro Nadeak, Kiki Ariyanti Sugeng, Tulus Tulus . Distance antimagic labeling of circulant graphs. AIMS Mathematics, 2024, 9(8): 21177-21188. doi: 10.3934/math.20241028
    [10] Mesfer H. Alqahtani, Alaa M. Abd El-latif . Separation axioms via novel operators in the frame of topological spaces and applications. AIMS Mathematics, 2024, 9(6): 14213-14227. doi: 10.3934/math.2024690
  • In this paper, the neural network domain with the backpropagation Levenberg-Marquardt scheme (NNB-LMS) is novel with a convergent stability and generates a numerical solution of the impact of the magnetohydrodynamic (MHD) nanofluid flow over a rotating disk (MHD-NRD) with heat generation/absorption and slip effects. The similarity variation in the MHD flow of a viscous liquid through a rotating disk is explained by transforming the original non-linear partial differential equations (PDEs) to an equivalent non-linear ordinary differential equation (ODEs). Varying the velocity slip parameter, Hartman number, thermal slip parameter, heat generation/absorption parameter, and concentration slip parameter, generates a Prandtl number using the Runge-Kutta 4th order method (RK4) numerical technique, which is a dataset for the suggested (NNB-LMS) for numerous MHD-NRD scenarios. The validity of the data is tested, and the data is processed and properly tabulated to test the exactness of the suggested model. The recommended model was compared for verification, and the estimation solutions for particular instances were assessed using the NNB-LMS training, testing, and validation procedures. A regression analysis, a mean squared error (MSE) assessment, and a histogram analysis were used to further evaluate the proposed NNB-LMS. The NNB-LMS technique has various applications such as disease diagnosis, robotic control systems, ecosystem evaluation, etc. Some statistical data such as the gradient, performance, and epoch of the model were analyzed. This recommended method differs from the reference and suggested results, and has an accuracy rating ranging from 1009to 1012.



    Since shunting inhibitory cellular neural networks were proposed by Bouzerdoum and Pinter [1] as a new type of neural networks, they have received more and more attention and have been widely applied in optimisation, psychophysics, speech and other fields. At the same time, since time delays are ubiquitous, many research results have been obtained on the dynamics of shunting inhibitory cellular neural networks with time delays [2,3,4,5,6].

    On the one hand, the quaternion is a generalization of real and complex numbers [7]. The skew field of quaternions is defined by

    H:={q=qR+iqI+jqJ+kqK},

    where qR,qI,qJ,qKR and the elements i,j and k obey the Hamilton's multiplication rules:

    ij=jk=k,jk=kj=i,ki=ik=j,i2=j2=k2=1.

    For q=qR+iqI+jqJ+kqK, we denote ˇq=iqI+jqJ+kqK and qR=qˇq. The norm of q is defined by qH=(qR)2+(qI)2+(qJ)2+(qK)2. For y=(y1,y2,,yn)THn, we define yHn=max1pn{yH}, then (Hn,Hn) is a Banach space. As we all know, quaternion-valued neural networks include real-valued neural networks and complex-valued neural networks as their special cases. Compared with complex-valued neural networks, quaternion-valued neural networks only needs half of the connection weight parameters of complex-valued neural networks when dealing with multi-level information [8]. In recent years, quaternion-valued neural networks have attracted the attention of many researchers, and their various dynamic behaviors, including fractional-order and stochastic quaternion-valued neural networks, have been extensively studied [9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24].

    On the other hand, because periodic and almost periodic oscillations are important dynamics of neural networks, the periodic and almost periodic oscillations of neural networks have been studied a lot in the past few decades [25,26,27,28,29,30,31,32,33]. Weyl almost periodicity is a generalization of Bohr almost periodicity and Stepanov almost periodicity [34,35,36,37]. It is a more complex recurrent oscillation. Because the spaces composed of Bohr almost periodic functions and Stepanov almost periodic functions are Banach spaces, it brings some convenience to study the existence of almost periodic solutions in these two senses of differential equations. Therefore, many results have been obtained on the Bohr almost periodic oscillation and Stepanov almost periodic oscillation of neural networks. However, the space composed of Weyl almost periodic functions is incomplete [38]. Therefore, the results of Weyl almost periodic solutions of neural networks are still very rare. Therefore, it is a meaningful and challenging work to study the existence of Weyl almost periodic solutions of neural networks.

    Motivated by the above, in this paper, we consider the following shunting inhibitory cellular neural networks with time-varying delays:

    ˙xij(t)=aij(t)xij(t)CklNr(i,j)Bklij(t)fij(xkl(t))xij(t)CklNs(i,j)Cklij(t)gij(xkl(tτkl(t)))xij(t)+Iij(t), (1.1)

    where ij{11,12,,1n,,m1,m2,,mn}:=Λ, Cij denotes the cell at the (i,j) position of the lattice. The r-neighborhood Nr(i,j) of Cij is given as

    Nr(i,j)={Ckl:max(|ki|,|lj|)r,1km,1ln},

    and Ns(i,j) is similarly specified; xij(t)H denotes the activity of the cell of Cij, Iij(t)H is the external input to Cij, aij(t)H is the coefficient of the leakage term, which represents the passive decay rate of the activity of the cell Cij, Bklij(t)0 and Cklij(t)0 represent the connection or coupling strength of postsynaptic of activity of the cell transmitted to the cell Cij, the activity functions fij,gij:HH are continuous functions representing the output or firing rate of the cell Cij, τkl(t) corresponds to the transmission delay and satisfies 0τkl(t)τ.

    The purpose of this paper is to use the fixed point theorem and a variant of Gronwall inequality to establish the existence and global exponential stability of Weyl almost periodic solutions for a class of quaternion-valued shunting inhibitory cellular neural networks whose coefficients of the leakage terms are quaternions. This is the first paper to study the existence and global exponential stability of Weyl almost periodic solutions of system (1.1) by using the fixed point theorem and a variant of Gronwall inequality. Our result of this paper is new, and our method can be used to study other types of quaternion-valued neural networks.

    For convenience, we introduce the following notations:

    am=minijΛ{inftR{aij(t)}},aM=maxijΛ{suptR{aij(t)}},ˇaMij=suptRˇaij(t)H,
    BklMij=suptR{Bklij(t)},CklMij=suptR{Cklij(t)},τMij=suptR{τij(t)},τ=maxklΛ{suptR{τkl(t)}}.

    The initial condition of system (1.1) is given by

    xij(s)=φij(s),s[τ,0],

    where φijC(R,H),ijΛ.

    Throughout this paper, we assume that:

    (H1) For ij,klΛ, functions aRij,Bklij,CklijAP(R,R+), ˇaijAP(R,H), IijAPWp(R,H), τklAP(R,R+)C1(R,R) and τ<1.

    (H2) For ijΛ, there exist positive constants Lfij and Lgij such that for all x,yH,

    fij(x)fij(y)HLfijxyH,gij(x)gij(y)HLgijxyH,

    and fij(0)=gij(0)=0.

    The rest of this paper is arranged as follows. In Section 2, we introduce some definitions and lemmas. In Section 3, we study the existence and global exponential stability of Weyl almost periodic solutions of (1.1). In Section 4, an example is given to verify the theoretical results. This paper ends with a brief conclusion in Section 5.

    Let (X,X) be a Banach space and BC(R,X) be the set of all bounded continuous functions from R to X.

    Definition 2.1. [38] A function fBC(R,X) is said to be almost periodic, if for every ϵ>0, there exists a constant l=l(ϵ)>0 such that in every interval of length l(ϵ) contains at least one σ such that

    f(t+σ)f(t)X<ϵ,tR.

    Denote by AP(R,X) the set of all such functions.

    For p[1,), we denote by Lploc(R,X) the space of all functions from R into X which are locally p-integrable. For fLploc(R,X), we define the following seminorm:

    fWp=limr+supβR(1rβ+rβf(t)pXdt)1p.

    Definition 2.2. [38] A function fLploc(R,X) is said to be p-th Weyl almost periodic (Wp-almost periodic for short), if for every ϵ>0, there exists a constant l=l(ϵ)>0 such that in every interval of length l(ϵ) contains at least one σ such that

    f(t+σ)f(t)Wp<ϵ.

    This σ is called on ϵ-translation number of f. The set of all such functions will be denoted by APWp(R,X).

    Remark 2.1. By Definitions 2.1 and 2.2, it is easy to see that if fAP(R,X), then fAPWp(R,X).

    Similar to the proofs of the lemma on page 83 and the lemma on page 84 of [39], it is not difficult to prove the following two lemmas.

    Lemma 2.1. If fAPWp(R,X), then f is bounded and uniformly continuous on R withrespect to the seminorn Wp.

    Using the argumentation contained in the proof of Proposition 3.21 in [38], one can easily prove the following.

    Lemma 2.2. If fkAPWp(R,X), k=1,2,,n. Then, for every ϵ>0, there exist common ϵ-translation numbers for these functions.

    Lemma 2.3. [40]Let g:RR be a continuous function such that, for every tR,

    0g(t)ρ(t)+γ1teη1(ts)g(s)ds++γnteηn(ts)g(s)ds (2.1)

    for some locally integrable function ρ:RR, and for some constantsγ1,,γn0, and some constants η1,,ηnγ, whereγ=np=1γp. We assume that the integrals in the right hand side of (2.1)are convergent. Let η=min{η1,,ηn}. Then, for every ξ(0,ηγ]such that 0eξsρ(s)ds converges, we have, for every tR,

    g(t)ρ(t)+γteξ(ts)ρ(s)ds.

    In particular, if ρ(t) is constant, we have

    g(t)ρηηγ.

    Let BUC(R,Hm×n) be a collection of bounded and uniformly continuous functions from R to Hm×n, then, the space BUC(R,Hm×n) with the norm x=suptRx(t)Hm×n is a Banach space, where xBUC(R,Hm×n).

    Denote ϕ0=(ϕ011,,ϕ01n,ϕ021,,ϕ02n,,ϕ0m1,,ϕ0mn)T, where

    ϕ0ij(t)=tetsaRij(v)dvIij(s)ds,ijΛ.

    We will show that ϕ0 is well defined under assumption (H1). In fact, by IijAPWp(R,H) and Lemma 2.1, there exists a constant M>0 such that IijWpM for all ijΛ. According to the Hölder inequality, one has

    ϕ0ij(t)Hteam(ts)Iij(s)dsHr=0(trt(r+1)eamq(ts)ds)1q(trt(r+1)Iij(s)pHds)1pr=0eamrM<+, (3.1)

    where 1p+1q=1, which means that ϕ0 is well defined.

    Take a positive constant αϕ0. Let

    Ω={ϕBUC(R,Hm×n)|ϕϕ0α}.

    Then, for every ϕΩ, one has

    ϕϕϕ0+ϕ02α.

    Theorem 3.1. Assume (H1)(H2) hold. Furthermore, suppose that

    (H3)

    κ=maxijΛ{2am[ˇaMij+CklNr(i,j)2BklMijLfijα+CklNs(i,j)2CklMijLgijα]}<1,

    (H4)for p>2,

    maxijΛ{24(2p4amp)p2(4amp)2[2(ˇaMij)p+2(CklNr(i,j)2BklMijLfijα)p+(1+2ep4amτ1τ)(CklNs(i,j)2CklMijLgijα)p]}<1,

    and, for p=2,

    maxijΛ{24(2am)2[2(ˇaMij)2+2(CklNr(i,j)2BklMijLfijα)2+(1+2e12amτ1τ)(CklNs(i,j)2CklMijLgijα)2]}<1,

    then system (1.1) has a unique Wp-almost periodic solution in Ω.

    Proof. It is easy to check that if x=(x11,,x1n,x21,,x2n,,xm1,,xmn)TΩ is a solution of the integral equation

    xij(t)=tetsaRij(v)dv[ˇaij(s)xij(s)CklNr(i,j)Bklij(s)fij(xkl(s))xij(s)CklNs(i,j)Cklij(s)gij(xkl(sτkl(s)))xij(s)+Iij(s)]ds,ijΛ, (3.2)

    then x is a solution of system (1.1).

    Define an operator T:ΩHm×n by

    (Tϕ)(t)=((T11ϕ)(t),,(T1nϕ)(t),(T21ϕ)(t),,(T2nϕ)(t),,(Tm1ϕ)(t),,(Tmnϕ)(t))T,

    where

    (Tijϕ)(t)=tetsaRij(v)dv[ˇaij(s)ϕij(s)CklNr(i,j)Bklij(s)fij(ϕkl(s))ϕij(s)CklNs(i,j)Cklij(s)gij(ϕkl(sτkl(s)))ϕij(s)+Iij(s)]ds,ijΛ.

    Now, we will prove that Tϕ is well defined. Actually, by (H1)(H3) and (3.1), for ijΛ, one deduces that

    (Tijϕ)(t)HtetsaRij(v)dvˇaij(s)ϕij(s)CklNr(i,j)Bklij(s)(fij(ϕkl(s))fij(0))ϕij(s)CklNs(i,j)Cklij(s)(gij(ϕkl(sτkl(s)))gij(0))ϕij(s)Hds+tetsaRij(v)dvIij(s)Hds1am(ˇaMij+CklNr(i,j)BklMijLfijϕ+CklNs(i,j)CklMijLgijϕ)ϕ+tetsaRij(v)dvIij(s)Hds<+. (3.3)

    That is, Tϕ is well defined.

    We will divide the rest of the proof into four steps.

    Step 1, we will prove that TϕBUC(R,Hm×n), for every ϕΩ.

    In fact, by (3.3), we see that Tϕ is bounded on R. So, we only need to show that Tϕ is uniformly continuous on R. Based on the Hölder inequality for 0h1 and q1 with 1p+1q=1, one has

    (Tijϕ)(t+h)(Tijϕ)(t)H=t+het+hsaRij(v)dv(ˇaij(s)ϕij(s)CklNr(i,j)Bklij(s)fij(ϕkl(s))ϕij(s)CklNs(i,j)Cklij(s)×gij(ϕkl(sτkl(s)))ϕij(s)+Iij(s))dstetsaRij(v)dv(ˇaij(s)ϕij(s)CklNr(i,j)Bklij(s)×fij(ϕkl(s))ϕij(s)CklNs(i,j)Cklij(s)gij(ϕkl(sτkl(s)))ϕij(s)+Iij(s))dsHt|et+hsaRij(v)dvetsaRij(v)dv|ˇaij(s)ϕij(s)+CklNr(i,j)Bklij(s)fij(ϕkl(s))ϕij(s)+CklNs(i,j)Cklij(s)×gij(ϕkl(sτkl(s)))ϕij(s)Hds+t|et+hsaRij(v)dvetsaRij(v)dv|Iij(s)Hds+t+htet+hsaRij(v)dvˇaij(s)ϕij(s)+CklNr(i,j)Bklij(s)fij(ϕkl(s))ϕij(s)+CklNs(i,j)Cklij(s)×gij(ϕkl(sτkl(s)))ϕij(s)Hds+t+htet+hsaRij(v)dvIij(s)Hds(ˇaMij+CklNr(i,j)BklMijLfijϕ+CklNs(i,j)CklMijLgijϕ)ϕt(etsaRij(v)dvt+htaRij(v)dv)ds+t(etsaRij(v)dvt+htaRij(v)dv)Iij(s)Hds+(ˇaMij+CklNr(i,j)BklMijLfijϕ+CklNs(i,j)CklMijLgijϕ)ϕt+htetsaRij(v)dvds+t+htetsaRij(v)dvIij(s)HdsaMh(ˇaMij+CklNr(i,j)BklMijLfijϕ+CklNs(i,j)CklMijLgijϕ)ϕteam(ts)ds+aMhtetsaRij(v)dvIij(s)Hds+(ˇaMij+CklNr(i,j)BklMijLfijϕ+CklNs(i,j)CklMijLgijϕ)ϕ×t+hteam(ts)ds+(t+hteqam(ts)ds)1q(t+htIij(s)pHds)1paMam(ˇaMij+CklNr(i,j)BklMijLfijϕ+CklNs(i,j)CklMijLgijϕ)ϕh+aMhtetsaRij(v)dvIij(s)Hds+eamh(ˇaMij+CklNr(i,j)BklMijLfijϕ+CklNs(i,j)CklMijLgijϕ)ϕh+eamhIijWph,

    where ijΛ. Hence, letting h0+, by (3.1), we have

    (Tijϕ)(t+h)(Tijϕ)(t)H0,

    which means that (Tijϕ) is uniformly continuous on R, ijΛ. Therefore, TϕBUC(R,Hm×n).

    Step 2, we will prove that T is a self-mapping from Ω to Ω.

    Actually, for arbitrary ϕΩ, from (H2)(H3), we have

    Tϕϕ0suptR{maxijΛ[tetsaRij(v)dvˇaij(s)ϕij(s)+CklNr(i,j)Bklij(s)(fij(ϕkl(s))f(0))ϕij(s)+CklNs(i,j)Cklij(s)(gij(ϕkl(sτkl(s)))gij(0))ϕij(s)Hds]}maxijΛ{1am[ˇaMij+CklNr(i,j)2BklMijLfijα+CklNs(i,j)2CklMijLgijα]}ϕκαα,

    which implies that TϕΩ. Consequently, T is a self-mapping from Ω to Ω.

    Step 3, we will prove T is a contraction mapping.

    As a matter of fact, in view of (H1)(H2), for any ϕ,νΩ, we can get

    TϕTνsuptR{maxijΛ[team(ts)(ˇaMijϕij(s)νij(s)H+CklNr(i,j)BklMijfij(ϕkl(s))(ϕij(s)νij(s))+(fij(ϕkl(s))fij(νkl(s)))νij(s)H+CklNs(i,j)CklMijgij(ϕkl(sτkl(s)))(ϕij(s)νij(s))+(gij(ϕkl(sτkl(s)))gij(νkl(sτkl(s))))νij(s)H)ds]}suptR{maxijΛ[team(ts)[ˇaMijϕij(s)νij(s)H+CklNr(i,j)BklMij(2Lfijαϕij(s)νij(s)H+2Lfijαϕkl(s)νkl(s)H)+CklNs(i,j)CklMij(2Lgijαϕij(s)νij(s)H2Lgijαϕkl(sτkl(s))νkl(sτkl(s))H)]ds]}maxijΛ{1am[ˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)4CklMijLgijα]}ϕν.

    From this and (H3), one has

    ||TϕTν||κϕν.

    Noticing that κ<1, T is a contraction mapping. Consequently, system (1.1) has a unique solution x in Ω.

    Step 4, we will prove that the unique solution xΩ is Wp-almost periodic.

    Indeed, since x=(x11,,x1n,x21,,x2n,,xm1,,xmn)TΩ, x is bounded and uniformly continuous. Hence, for every ϵ>0, there exists a δ(0,ϵ) such that for any t1,t2R with |t1t2|<δ and ijΛ, we have

    xij(t1)xij(t2)H<ϵ. (3.4)

    Also, for this δ, in view of (H1) and Lemma 2.2, we see that there exists a common δ-translation number σ such that

    limr+supβR(1rβ+rβIij(t+σ)Iij(t)pHdt)1p<δ<ϵ, (3.5)
    |Bklij(t+σ)Bklij(t)|<ϵ, (3.6)
    |Cklij(t+σ)Cklij(t)|<ϵ, (3.7)
    |aRij(t+σ)aRij(t)|<ϵ,ˇaij(t+σ)ˇaij(t)H<ϵ (3.8)

    and

    |τij(t+σ)τij(t)|<δ, (3.9)

    where ij,klΛ. Consequently, from (3.4) and (3.9), we get

    xij(tτij(t+σ))xij(tτij(t))H<ϵ. (3.10)

    Since x is a solution of system (1.1), by (3.2), for ijΛ, we have

    xij(t+σ)xij(t)HtetsaRij(v+σ)dv(ˇaij(s+σ)xij(s+σ)ˇaij(s)xij(s))dsH+tetsaRij(v+σ)dvCklNr(i,j)[Bklij(s+σ)fij(xkl(s+σ))xij(s+σ)Bklij(s)fij(xkl(s))xij(s)]dsH+tetsaRij(v+σ)dvCklNs(i,j)[Cklij(s+σ)×gij(xkl(s+στkl(s+σ)))xij(s+σ)Cklij(s)gij(xkl(sτkl(s)))xij(s)]dsH+tetsaRij(v+σ)dv[Iij(s+σ)Iij(s)]dsH+t(etsaRij(v+σ)dvetsaRij(v)dv)ˇaij(s)xij(s)dsH+t(etsaRij(v+σ)dvetsaRij(v)dv)CklNr(i,j)Bklij(s)fij(xkl(s))xij(s)dsH+t(etsaRij(v+σ)dvetsaRij(v)dv)CklNs(i,j)Cklij(s)gij(xkl(sτkl(s)))xij(s)dsH+t(etsaRij(v+σ)dvetsaRij(v)dv)Iij(s)dsH:=8l=1Blij(t). (3.11)

    When p>2, it follows from Hölder's inequality (2p,p2p), Hölder's inequality (12,12) and (H2) that

    B2ij(t)team(ts)CklNr(i,j)Bklij(s+σ)fij(xkl(s+σ))(xij(s+σ)xij(s))Hds+team(ts)CklNr(i,j)Bklij(s+σ)(fij(xkl(s+σ))fij(xkl(s)))xij(s)Hds+team(ts)CklNr(i,j)(Bklij(s+σ)Bklij(s))fij(xkl(s))xij(s)Hds(tep2p4am(ts)ds)p2p[tep4am(ts)(CklNr(i,j)2BklMijLfijαxij(s+σ)xij(s))H)p2ds]2p+(tep2p4am(ts)ds)p2p[tep4am(ts)(CklNr(i,j)2BklMijLfijαxkl(s+σ)xkl(s)H)p2ds]2p+(tep2p4am(ts)ds)p2p[tep4am(ts)(CklNr(i,j)4Lfijα2|Bklij(s+σ)Bklij(s)|ds)p2ds]2p(2p4amp)p2p{[tep4am(ts)(CklNr(i,j)2BklMijLfijαxij(s+σ)xij(s))H)p2ds]2p+[tep4am(ts)(CklNr(i,j)2BklMijLfijαxkl(s+σ))xkl(s)H)p2ds]2p+[tep4am(ts)(CklNr(i,j)4Lfijα2|Bklij(s+σ)Bklij(s)|ds)p2ds]2p}(2p4amp)p2p{[(tep4am(ts)ds)12(tep4am(ts)(CklNr(i,j)2BklMijLfijαxij(s+σ)xij(s))H)pds)12]2p+[(tep4am(ts)ds)12(tep4am(ts)(CklNr(i,j)2BklMijLfijα×xkl(s+σ))xkl(s)H)pds)12]2p+[(tep4am(ts)ds)12(tep4am(ts)×(CklNr(i,j)4Lfijα2|Bklij(s+σ)Bklij(s)|)pds)12]2p}(2p4amp)p2p(4amp)1p{[tep4am(ts)(CklNr(i,j)2BklMijLfijαxij(s+σ)xij(s))H)pds]1p+[tep4am(ts)(CklNr(i,j)2BklMijLfijαxkl(s+σ))xkl(s)H)pds]1p+[tep4am(ts)(CklNr(i,j)4Lfijα2|Bklij(s+σ)Bklij(s)|)pds]1p}, (3.12)

    for ijΛ. Similarly, we have

    B1ij(t)(2p4amp)p2p(4amp)1p{ˇaMij[tep4am(ts)xij(s+σ)xij(s)pHds]1p+2α[tep4am(ts)ˇaij(s+σ)˘aij(s)pHds]1p}, (3.13)
    B3ij(t)(2p4amp)p2p(4amp)1p{[tep4am(ts)(CklNs(i,j)2CklMijLgijαxij(s+σ)xij(s))H)pds]1p+[tep4am(ts)(CklNs(i,j)2CklMijLgijαxkl(s+στkl(s+σ))xkl(sτkl(s))H)pds]1p+[tep4am(ts)(CklNs(i,j)4Lgijα2|Cklij(s+σ)Cklij(s))|)pds]1p} (3.14)

    and

    B4ij(t)(2p4amp)p2p(4amp)1p(tep4am(ts)Iij(s+σ)Iij(s)pHds)1p, (3.15)

    for ijΛ.

    Besides, combining with Hölder's inequality (2p,p2p), Hölder's inequality (12,12) and (H2) that

    B5ij(t)2ˇaMijαteam(ts)(ts|aRij(v+σ)aRij(v)|dv)ds2ˇaMijα(tep2p4am(ts)ds)p2p[tep4am(ts)(ts|aRij(v+σ)aRij(v)|dv)p2ds]2p(2p4amp)p2p2ˇaMijα[(tep4am(ts)ds)12(tep4am(ts)(ts|aRij(v+σ)aRij(v)|dv)pds)12]2p(2p4amp)p2p(4amp)1p2ˇaMijα[tep4am(ts)(ts|aRij(v+σ)aRij(v)|dv)pds]1p, (3.16)

    for ijΛ. In a similar way, one can get that

    B6ij(t)(2p4amp)p2p(4amp)1pCklNr(i,j)4BklMijLfijα2[tep4am(ts)(ts|aRij(v+σ)aRij(v)|dv)pds]1p, (3.17)
    B7ij(t)(2p4amp)p2p(4amp)1pCklNs(i,j)4CklMijLgijα2[tep4am(ts)(ts|aRij(v+σ)aRij(v)|dv)pds]1p (3.18)

    and

    B8ij(t)(2p4amp)p2p(4amp)1p[tep4am(ts)(ts|aRij(v+σ)aRij(v)|dv)pIij(s)pHds]1p, (3.19)

    for ijΛ. Hence, together with a change of variables, Fubini's theorem, Hölder's inequality, (3.8) and (3.13), we derive that

    1rβ+rβBp1ij(t)dt(2p4amp)p24amp{1rβ+rβ[[ˇaMijtep4am(ts)xij(s+σ)xij(s)pHds]1p+2α[tep4am(ts)ˇaij(s+σ)˘aij(s)pHds]1p]pdt}(2p4amp)p28amp{(ˇaMij)p1rβ+rβtep4am(ts)xij(s+σ)xij(s)pHdsdt+(2α)p1rβ+rβtep4am(ts)ˇaij(s+σ)˘aij(s)pHdsdt}(2p4amp)p28amp{(ˇaMij)pβep4am(βs)(1rs+rsxij(t+σ)xij(t)pHdt)ds+(2α)pβep4am(βs)(1rs+rsˇaij(t+σ)˘aij(t)pHdt)ds}=(ˇaMij)p(2p4amp)p28ampβep4am(βs)Θσ,r(s)ds+ρ1ij,

    where

    Θσ,r(s):=1rs+rsx(t+σ)x(t)pHndt

    and

    ρ1ij:=2(2p4amp)p2(4amp)2(2αϵ)p, (3.20)

    and, together with a change of variables, Fubini's theorem, Hölder's inequality, (3.6) and (3.12), we derive that

    1rβ+rβBp2ij(t)dt(2p4amp)p24amp{1rβ+rβ[(tep4am(ts)(CklNr(i,j)2BklMijLfijαxij(s+σ)xij(s))H)pds)1p+(tep4am(ts)(CklNr(i,j)2BklMijLfijαxkl(s+σ)xkl(s)H)pds)1p+(tep4am(ts)(CklNr(i,j)4Lfijα2|Bklij(s+σ)Bklij(s)|)pds)1p]pdt}(2p4amp)p212amp{1rβ+rβtep4am(ts)(CklNr(i,j)2BklMijLfijαxij(s+σ)xij(s)H)pdsdt+1rβ+rβtep4am(ts)(CklNr(i,j)2BklMijLfijαxkl(s+σ)xkl(s)H)pdsdt+1rβ+rβtep4am(ts)(CklNr(i,j)4Lfijα2|Bklij(s+σ)Bklij(s)|)pdsdt}(2p4amp)p212amp[(CklNr(i,j)2BklMijLfijα)p1rβ+rβtep4am(ts)x(s+σ)x(s)pHm×ndsdt+(CklNr(i,j)2BklMijLfijα)p1rβ+rβtep4am(ts)x(s+σ)x(s)Hm×ndsdt+1rβ+rβtep4am(ts)(4mnLfijα2ϵ)pdsdt](2p4amp)p212amp[(CklNr(i,j)2BklMijLfijα)pβep4am(βs)(1rs+rsx(t+σ)x(t)pHm×ndt)ds+(CklNr(i,j)2BklMijLfijα)pβep4am(βs)(1rs+rsx(t+σ)x(t)Hm×ndt)ds+4amp(4mnLfijα2ϵ)p]=(2p4amp)p224amp(CklNr(i,j)2BklMijLfijα)pβep4am(βs)Θσ,r(s)ds+ρ2ij,

    where

    ρ2ij:=3(2p4amp)p2(4amp)2(4mnLfijα2ϵ)p. (3.21)

    Moreover, based on a change of variables, Fubini's theorem, Hölder's inequality, (3.7), (3.10) and (3.14), we deduce that

    1rβ+rβBp3ij(t)dt(2p4amp)p24amp{1rβ+rβ[(tep4am(ts)(CklNs(i,j)2CklMijLgijαxij(s+σ)xij(s))H)pds)1p+(tep4am(ts)(CklNs(i,j)2CklMijLgijαxkl(s+στkl(s+σ))xkl(sτkl(s))H)pds)1p+(tep4am(ts)(CklNs(i,j)4Lgijα2(Cklij(s+σ)Cklij(s))H)pds)1p]pdt}(2p4amp)p212amp{1rβ+rβtep4am(ts)(CklNs(i,j)2CklMijLgijαxij(s+σ)xij(s))H)pdsdt+1rβ+rβtep4am(ts)(CklNs(i,j)2CklMijLgijαxkl(s+στkl(s+σ))xkl(sτkl(s))H)pdsdt+1rβ+rβtep4am(ts)(CklNs(i,j)4Lgijα2|Cklij(s+σ)Cklij(s))|)pdsdt}(2p4amp)p212amp{1rβ+rβtep4am(ts)(CklNs(i,j)2CklMijLgijαxij(s+σ)xij(s))H)pdsdt+1rβ+rβ[2tep4am(ts)(CklNs(i,j)2CklMijLgijαxkl(s+στkl(s+σ))xkl(sτkl(s+σ))H)pds+2tep4am(ts)(CklNs(i,j)4CklMijLgijα2xkl(sτkl(s+σ))xkl(sτkl(s))H)pds]dt+1rβ+rβtep4am(ts)(4mnLgijα2ϵ)pdsdt}(2p4amp)p212amp{1rβ+rβtep4am(ts)(CklNs(i,j)2CklMijLgijαxij(s+σ)xij(s))H)pdsdt+1rβ+rβ[21τtτkl(t+σ)ep4am(tuτ)(CklNs(i,j)2CklMijLgijαxkl(u+σ)xkl(u)H)pdu+2(CklNs(i,j)2CklMijLgijαϵ)ptep4am(ts)ds]dt+4amp(4mnLgijα2ϵ)p}(2p4amp)p212amp{1rβ+rβtep4am(ts)(CklNs(i,j)2CklMijLgijαxij(s+σ)xij(s))H)pdsdt+1rβ+rβ[2ep4amτ1τtep4am(ts)(CklNs(i,j)2CklMijLgijαxkl(s+σ)xkl(s)H)pds]dt+8amp(CklNs(i,j)2CklMijLgijαϵ)p+4amp(4mnLgijα2ϵ)p}(2p4amp)p212amp[(CklNs(i,j)2CklMijLgijα)pβep4am(βs)(1rs+rsx(t+σ)x(t))Hm×ndt)ds+2ep4amτ1τ(CklNs(i,j)2CklMijLgijα)pβep4am(βs)(1rs+rsx(t+σ)x(t)pHm×ndt)ds+8amp(CklNs(i,j)2CklMijLgijαϵ)p+4amp(4mnLgijα2ϵ)p]=(2p4amp)p212amp(1+2ep4amτ1τ)(CklNs(i,j)2CklMijLgijα)pβep4am(βs)Θσ,r(s)ds+ρ3ij,

    where

    ρ3ij:=3(2p4amp)p2(4amp)2[2(CklNs(i,j)2CklMijLgijα)p+(4mnLgijα2)p]ϵp. (3.22)

    In view of (3.5), (3.8) and (3.15)–(3.19), we can easily obtain that

    1rβ+rβBp4ij(t)dt(2p4amp)p24ampβep4am(βs)(1rs+rsIij(t+σ)Iij(t)pHdt)ds(2p4amp)p2(4amp)2ϵp:=ρ4ij, (3.23)
    1rβ+rβBp5ij(t)dt(2p4amp)p24amp(2ˇaMijα)p[1rβ+rβtep4am(ts)(ts|aRij(v+σ)aRij(v)|dv)pdsdt](2p4amp)p24amp(2ˇaMijαϵ)p0ep4amsspds:=ρ5ij, (3.24)
    1rβ+rβBp6ij(t)dt(2p4amp)p24amp(CklNr(i,j)4BklMijLfijα2)p[1rβ+rβtep4am(ts)×(ts|aRij(v+σ)aRij(v)|dv)pdsdt](2p4amp)p24amp(CklNr(i,j)4BklMijLfijα2ϵ)p0ep4amsspds:=ρ6ij, (3.25)
    1rβ+rβBp7ij(t)dt(2p4amp)p24amp(CklNs(i,j)4CklMijLgijα2)p[1rβ+rβtep4am(ts)×(ts|aRij(v+σ)aRij(v)|dv)pdsdt](2p4amp)p24amp(CklNs(i,j)4CklMijLgijα2ϵ)p0ep4amsspds:=ρ7ij (3.26)

    and

    1rβ+rβBp8ij(t)dt(2p4amp)p24amp[1rβ+rβtep4am(ts)(ts|aRij(v+σ)aRij(v)|dv)pIij(s)pHdsdt]ϵp(2p4amp)p24amp[1rβ+rβtep4am(ts)(ts)pIij(s)pHdsdt](2p4amp)p24ampIijpWpϵp0ep4amsspds:=ρ8ij, (3.27)

    for ijΛ. Consequently, combining with (3.11) and (3.20)–(3.27), we obtain

    Θσ,r(β)maxijΛ{88l=11rβ+rβBplij(t)dt}ρ+γβeη(βs)Θσ,r(s)ds,

    where η=p4am,

    ρ=86l=1maxijΛ{ρlij}=8maxijΛ{(2p4amp)p24amp[8amp(2α)p+12amp(4mnLfijα2)p+24amp(CklNs(i,j)2CklMijLgijα)P+12amp(4mnLgijα2)p+4amp+[(2ˇaMijα)p+(CklNr(i,j)4BklMijLfijα2)p+(CklNs(i,j)4CklMijLgijα2)p+IijpWp]0ep4amsspds]}ϵp

    and

    γ=8maxijΛ{(2p4amp)p212amp[23(ˇaMij)p+2(CklNr(i,j)2BklMijLfijα)p+(1+2ep4amτ1τ)(CklNs(i,j)2CklMijLgijα)p]}.

    By (H4), we have γ<η. Thus, it follows from Lemma 2.3 that

    1rβ+rβx(t+σ)x(t)pHm×ndtρηηγ.

    Hence, xAPWp(R,Hm×n).

    When p=2, similar to the proof of the case of p>2, one can obtain

    Θσ,r(β)˜ρ+˜γβe˜η(βs)Θσ,r(s)ds,

    where ˜η=am2,

    ˜ρ=88l=1maxijΛ{˜ρlij}=maxijΛ{16am[4am(2α)2+6am(4mnLfijα2ϵ)2+12am(2CklNs(i,j)CklMijLgijα)2+6am(4mnLgijα2)2+2am+[(2ˇaMijα)2+(CklNr(i,j)4BklMijLfijα2)2+(CklNs(i,j)4CklMijLgijα2)2+Iij2W2]0e12amss2ds]}ϵ2

    and

    ˜γ=maxijΛ{48am[23(ˇaMij)2+2(CklNr(i,j)2BklMijLfijα)2+(1+2e12amτ1τ)(CklNs(i,j)2CklMijLgijα)2]}.

    By (H4), we have ˜γ<˜η. Thus, it follows from Lemma 2.3 that

    1rβ+rβx(t+σ)x(t)pHm×ndt˜ρ˜η˜η˜γ,

    which means that xAPW2(R,Hm×n). The proof is complete.

    Definition 3.1. [14] Let x be a solution of system (1.1) with the initial value φ and y be an arbitrary solution of system (1.1) with the initial value ψ, respectively. If there exist positive constants λ and M such that

    x(t)y(t)Hm×nMeλtφψτ,tR+,

    where φψτ=supt[τ,0]φ(t)ψ(t)Hm×n. Then the solution x of system (1.1) is said to be globally exponentially stable.

    Theorem 3.2. Assume that (H1)(H3) hold, then system (1.1) has a unique Wp-almost periodic solution that is globallyexponentially stable.

    Proof. Let x(t) be the Wp-almost periodic solution with the initial value φ(t) and y(t) be an arbitrary solution with the initial value ψ(t). Taking

    zij(t)=xi(t)yij(t),ϕij(t)=φij(t)ψij(t),ijΛ,

    we have

    ˙zij(t)=aij(t)zij(t)CklNr(i,j)Bklij(t)(fij(xkl(t))xij(t)fij(ykl(t))yij(t)CklNs(i,j)Cklij(t)(gij(xkl(tτkl(t)))xij(t)gij(ykl(tτkl(t)))yij(t)),ijΛ. (3.28)

    For ijΛ, we define the following functions:

    Πij(u)=amu(ˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)2CklijLgijαeuτMij+CklNs(i,j)2CklijLgijα).

    From (H3), we get

    Πij(0)=am(ˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)4CklMijLgijα)>0,ijΛ.

    Since Πij(u) is continuous on [0,+) and Πij(u) as u, there exists ζij>0 such that Πij(ζij)=0 and Πij(u)>0, for u(0,ζij), ijΛ. Let ς=minijΛ{ζij}, then we have Πij(ς)0,ijΛ. Hence, we can choose a positive constant λ such that 0<λ<min{ς,am} and Πij(λ)>0. Thus, one has

    1amλ(ˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)2CklijLgijαeλτMij+CklNs(i,j)2CklijLgijα)<1, (3.29)

    where ijΛ. Take a constant M=maxiΛ{amˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)4CklMijLgijα}, then by (H3), we have M>1. Thus,

    1M1amλ(ˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)2CklijLgijαeλτMij+CklNs(i,j)2CklijLgijα)<0,ijΛ. (3.30)

    From (3.28), we have

    ˙zij(t)+aRij(t)zij(t)=ˇaij(t)zij(t)CklNr(i,j)Bklij(t)(fij(xkl(t))xij(t)fij(ykl(t))yij(t))CklNs(i,j)Cklij(t)(gij(xkl(tτkl(t)))xij(t)gij(ykl(tτkl(t)))yij(t)),ijΛ. (3.31)

    Multiplying both sides of (3.31) by et0aRij(v)dv and integrating over [0,t], we have

    zij(t)=ϕij(0)et0aRij(v)dv+t0etsaRij(v)dv[ˇaij(t)zij(s)CklNr(i,j)Bklij(t)(fij(xkl(t))xij(t)fij(ykl(t))yij(t))CklNs(i,j)Cklij(t)(gij(xkl(tτkl(t)))xij(t)gij(ykl(tτkl(t)))yij(t))]ds,ijΛ.

    Hence, for any ϵ>0, it is easy to see that

    z(t)Hm×n<(ϕτ+ϵ)eλt<M(ϕτ+ϵ)eλt,t(τ,0].

    We claim that

    z(t)Hm×n<M(ϕτ+ϵ)eλt,t[0,+). (3.32)

    Otherwise, there exists t>0 such that

    z(t)Hm×n=M(ϕτ+ϵ)eλt (3.33)

    and

    z(t)Hm×n<M(ϕτ+ϵ)eλt,t<t. (3.34)

    From (3.29), (3.30) and (3.34), we get

    zij(t)Hϕij(0)Hetam+t0etam[ˇaMijzij(s)H+CklNr(i,j)Bklij(t)×(fij(xkl(t))(xij(t)yij(t))H+(fij(xkl(t))fij(ykl(t)))yij(t)H)+CklNs(i,j)Cklij(t)(gij(xkl(tτkl(t)))(xij(t)yij(t))H+(gij(xkl(tτkl(t)))gij(ykl(tτkl(t))))yij(t)H)]ds(ϕτ+ϵ)eamt+M(ϕτ+ϵ)t0e(ts)am(ˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)2CklijLgijαeλτMij+CklNs(i,j)2CklijLgijα)eλsds(ϕτ+ϵ)eλte(λam)t+M(ϕτ+ϵ)eλt(1e(λam)t)amλ(ˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)2CklijLgijαeλτMij+CklNs(i,j)2CklijLgijα)=M(ϕτ+ϵ)eλt[e(λam)tM+1e(λam)tamλ(ˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)2CklijLgijαeλτMij+CklNs(i,j)2CklijLgijα)]=M(ϕτ+ϵ)eλt[(1M1amλ(ˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)2CklijLgijαeλτMij+CklNs(i,j)2CklijLgijα)e(λam)t+1amλ(ˇaMij+CklNr(i,j)4BklMijLfijα+CklNs(i,j)2CklijLgijαeλτMij+CklNs(i,j)2CklijLgijα)]<M(ϕτ+ϵ)eλt,ijΛ,

    which contradicts the Eq (3.33). Hence, (3.32) holds. Letting ϵ0+, from (3.32), we have

    z(t)Hm×nMϕτeλt,t>0.

    Therefore, the Wp-almost periodic solution of system (1.1) is globally exponentially stable. This completes the proof.

    Example 4.1. In system (1.1), let i,j=1,2, r=s=1 and take

    xij(t)=xRij(t)+ixIij(t)+jxJij(t)+kxKij(t)H,
    a11(t)=32sin2(3t)+i25cos(2t)j25sin(5t)+k1100cos2(3t),
    a21(t)=39|cos(2t)|i210sin(2t)+j310sin2(7t)+k3100cos(t),
    a12(t)=33cos2(t)i36sin(5t)j14cos3(5t)+k16cos2(3t)
    a22(t)=37sin4(3t)+i14cos(7t)j2916|sin(11t)|+k18cos2(5t),
    τ11(t)=232sin8(2t),τ12(t)=116cos2(3t),τ21(t)=141sin2(25t),τ22(t)=129sin4(37t),
    B11(t)=13|cos(2t)|,B12(t)=16sin(4t)+12B21(t)=34sin(2t)+1,B22(t)=14sin2(5t),
    C11(t)=cos(3t)+3,C12(t)=|sin(5t)|,C21(t)=2sin2(7t),C22(t)=14sin(2t)+74,
    f11(x)=f12(x)=15sin(14xR+36xK)i15|324xI+35xJ|+k14sin(15xR),
    f21(x)=f22(x)=14|23xI+37xK|i28sin(15xR+25xK)+k23sin(15xIj),
    g11(x)=g12(x)=120sin(922xR)i340sin(75xK)+j220sin(13xJ+34xI),
    g21(x)=g22(x)=225|13xJ+23xI|i340sin(75xK+xR)+j121sin(42xR),
    I11(t)=I21(t)=2sint+i43e|t|+k23cos(12t),
    I12(t)=I22(t)=i23sin2tj15cos2t+ke|t|.

    By computing, we obtain

    Lf11=Lf12=310,Lf21=Lf22=16,Lg11=Lg12=920,Lg21=Lg22=821,
    am=32,ˇaM11=12,ˇaM12=25,ˇaM21=512,ˇaM22=716,τ=116,τ=12,
    CklN1(1,1)BklM11=CklN1(1,2)BklM12=CklN1(2,1)BklM21=CklN1(2,2)BklM22=4,
    CklN1(1,1)CklM11=CklN1(1,2)CklM12=CklN1(2,1)CklM21=CklN1(2,2)CklM22=8.

    Choose a constant α=116ϕ0, for i,j=1,2, we obtain that

    κ=maxijΛ{2am[ˇaMij+CklNr(i,j)2BklMijLfijα+CklNs(i,j)2CklMijLgijα]}=0.06875<1.

    For i,j=1,2, take p=3, it is easy to obtain that

    maxijΛ{96(23am)3[23(ˇaMij)3+2(CklNr(i,j)2BklMijLfijα)3+(1+2e34amτ1τ)(CklNs(i,j)2CklMijLgijα)3]}=0.0016<1.

    Take p=2, we have

    maxijΛ{24(2am)2[23(ˇaMij)2+2(CklNr(i,j)2BklMijLfijα)2+(1+2e12amτ1τ)(CklNs(i,j)2CklMijLgijα)2]}=0.2832<1.

    Thus, conditions (H1)(H4) of Theorems 3.1 and Theorems 3.2 are satisfied. Hence, system (1.1) has a unique Wp-almost periodic solution that is globally exponentially stable (see Figures 1, 2).

    Figure 1.  Curves of (xl11(t),xl12(t))T of system (1.1) with the initial values (xl11(0),xl12(0))T=(1,2)T,(3,4)T,(5,5)T,(7,7)T,(9,9)T, l = R, I, J, K.
    Figure 2.  Curves of (xl21(t),xl22(t))T of system (1.1) with the initial values (xl21(0),xl22(0))T=(8,1)T,(9,3)T,(5,5)T,(7,2)T,(9,4)T, l = R, I, J, K.

    Remark 4.1. No known results are available to give the results of Example 4.1.

    In this paper, the existence and global exponential stability of Weyl almost periodic solutions for a class of quaternion-valued neural networks with time-varying delays are established. Even when the system we consider is a real-valued system, our results are brand-new. In addition, the method in this paper can be used to study the existence of Weyl almost periodic solutions for other types of neural networks.

    This work is supported by the National Natural Science Foundation of China under Grant 11861072.

    The authors declare no conflict of interest in this paper.



    [1] B. M. Wilamowski, H. Yu, Improved computation for Levenberg-Marquardt training, IEEE Trans. Neur. Net., 21 (2010), 930–937. https://doi.org/10.1109/TNN.2010.2045657 doi: 10.1109/TNN.2010.2045657
    [2] D. B. Parker, A comparison of algorithms for neuron‐like cells, AIP Conf. Proc., 151 (1986), 327–332. https://doi.org/10.1063/1.36233 doi: 10.1063/1.36233
    [3] A. Akbar, H. Ullah, M. A. Z. Raja, K. S. Nisar, S. Islam, M. Shoaib, 2022. A design of neural networks to study MHD and heat transfer in two phase model of nano-fluid flow in the presence of thermal radiation, Waves Random Complex Media, 2022, 1–24. https://doi.org/10.1080/17455030.2022.2152905 doi: 10.1080/17455030.2022.2152905
    [4] M. Shoaib, M. A. Z. Raja, M. T. Sabir, S. Islam, Z. Shah, P. Kumam, et al., Numerical investigation for rotating flow of MHD hybrid nanofluid with thermal radiation over a stretching sheet, Sci. Rep., 10 (2020), 18533. https://doi.org/10.1038/s41598-020-75254-8 doi: 10.1038/s41598-020-75254-8
    [5] R. A. Khan, H. Ullah, M. A. Z. Raja, M. A. R. Khan, S. Islam, M. Shoaib, Heat transfer between two porous parallel plates of steady nano fludis with Brownian and Thermophoretic effects: a new stochastic numerical approach, Int. Commun. Heat Mass Transfer, 126 (2021), 105436. https://doi.org/10.1016/j.icheatmasstransfer.2021.105436 doi: 10.1016/j.icheatmasstransfer.2021.105436
    [6] Z. Sabir, M. A. Z. Raja, J. L. Guirao, M. Shoaib, Integrated intelligent computing with neuro-swarming solver for multi-singular fourth-order nonlinear Emden-Fowler equation, Comp. Appl. Math., 39 (2020), 307. https://doi.org/10.1007/s40314-020-01330-4 doi: 10.1007/s40314-020-01330-4
    [7] H. Yasmin, L. A. AL-Essa, A. M. Mahnashi, W. Hamali, A. Saeed, A magnetohydrodynamic flow of a water-based hybrid nanofluid past a convectively heated rotating disk surface: a passive control of nanoparticles, Rev. Adv. Mater. Sci., 63 (2024), 20240054. https://doi.org/10.1515/rams-2024-0054 doi: 10.1515/rams-2024-0054
    [8] M. M. Al-Sawalha, H. Yasmin, S. Muhammad, Y. Khan, R. Shah, Optimal power management of a stand-alone hybrid energy management system: hydro-photovoltaic-fuel cell, Ain Shams Eng. J., 2024, 103089. https://doi.org/10.1016/j.asej.2024.103089 doi: 10.1016/j.asej.2024.103089
    [9] K. B. Pavlov, Magnetohydrodynamic flow of an incompressible viscous fluid caused by deformation of a plane surface, Magn. Gidrodin., 4 (1974), 146–147.
    [10] A. A. Aldhafeeri, H. Yasmin, A numerical analysis of the rotational flow of a hybrid nanofluid past a unidirectional extending surface with velocity and thermal slip conditions, Rev. Adv. Mater. Sci., 63 (2024), 20240052. https://doi.org/10.1515/rams-2024-0052 doi: 10.1515/rams-2024-0052
    [11] A. Chakrabarti, A. S. Gupta, Hydromagnetic flow and heat transfer over a stretching sheet, Q. Appl. Math., 37 (1979), 73–78.
    [12] M. R. Khan, S. Mao, Numerical solution of magnetohydrodynamics radiative flow of Oldroyd-B nanofluid toward a porous stretched surface containing gyrotactic microorganisms, ZAMM‐J. Appl. Math. Mech./Z. Angew. Math. Mech., 102 (2022), e202100388. https://doi.org/10.1002/zamm.202100388 doi: 10.1002/zamm.202100388
    [13] H. Yasmin, Numerical investigation of heat and mass transfer study on MHD rotatory hybrid nanofluid flow over a stretching sheet with gyrotactic microorganisms, Ain Shams Eng. J., 15 (2024), 102918. https://doi.org/10.1016/j.asej.2024.102918 doi: 10.1016/j.asej.2024.102918
    [14] S. Heysiattalab, A. Malvandi, D. D. Ganji, Anisotropic behavior of magnetic nanofluids (MNFs) at filmwise condensation over a vertical plate in presence of a uniform variable-directional magnetic field, J. Mol. Liq., 219 (2016), 875–882. https://doi.org/10.1016/j.molliq.2016.04.004 doi: 10.1016/j.molliq.2016.04.004
    [15] H. Yasmin, Analytical investigation of convective phenomena with nonlinearity characteristics in nanostratified liquid film above an inclined extended sheet, Nanotechnol. Rev., 13 (2024), 20240064. https://doi.org/10.1515/ntrev-2024-0064 doi: 10.1515/ntrev-2024-0064
    [16] T. Hayat, A. Aziz, T. Muhammad, A. Alsaedi, On magnetohydrodynamic three-dimensional flow of nanofluid over a convectively heated nonlinear stretching surface, Int. J. Heat Mass Transfer, 100 (2016), 566–572. https://doi.org/10.1016/j.ijheatmasstransfer.2016.04.113 doi: 10.1016/j.ijheatmasstransfer.2016.04.113
    [17] T. Hayat, M. Waqas, M. I. Khan, A. Alsaedi, Analysis of thixotropic nanomaterial in a doubly stratified medium considering magnetic field effects, Int. J. Heat Mass Transfer, 102 (2016), 1123–1129. https://doi.org/10.1016/j.ijheatmasstransfer.2016.06.090 doi: 10.1016/j.ijheatmasstransfer.2016.06.090
    [18] M. R. Khan, S. Mao, W. Deebani, A. M. A. Elsiddieg, Numerical analysis of heat transfer and friction drag relating to the effect of Joule heating, viscous dissipation and heat generation/absorption in aligned MHD slip flow of a nanofluid, Int. Commun. Heat Mass Transfer, 131 (2022), 105843. https://doi.org/10.1016/j.icheatmasstransfer.2021.105843 doi: 10.1016/j.icheatmasstransfer.2021.105843
    [19] A. Malvandi, A. Ghasemi, D. D. Ganji, Thermal performance analysis of hydromagnetic Al2O3-water nanofluid flows inside a concentric microannulus considering nanoparticle migration and asymmetric heating, Int. J. Thermal Sci., 109 (2016), 10–22. https://doi.org/10.1016/j.ijthermalsci.2016.05.023 doi: 10.1016/j.ijthermalsci.2016.05.023
    [20] A. A. Aldhafeeri, H. Yasmin, Thermal analysis of the water-based micropolar hybrid nanofluid flow comprising diamond and copper nanomaterials past an extending surface, Case Stud. Thermal Eng., 59 (2024), 104466. https://doi.org/10.1016/j.csite.2024.104466 doi: 10.1016/j.csite.2024.104466
    [21] T. Hayat, T. Muhammad, S. A. Shehzad, A. Alsaedi, An analytical solution for magnetohydrodynamic Oldroyd-B nanofluid flow induced by a stretching sheet with heat generation/absorption, Int. J. Thermal Sci., 111 (2017), 274–288. https://doi.org/10.1016/j.ijthermalsci.2016.08.009 doi: 10.1016/j.ijthermalsci.2016.08.009
    [22] W. Yu, D. M. France, S. U. S. Choi, J. L. Routbort, Review and assessment of nanofluid technology for transportation and other applications, Argonne National Lab. (ANL), Argonne, IL (United States), 2007. https://doi.org/10.2172/919327
    [23] H. Yasmin, L. A. Al-Essa, R. Bossly, H. Alrabaiah, S. A. Lone, A. Saeed, A homotopic analysis of the blood-based bioconvection Carreau-Yasuda hybrid nanofluid flow over a stretching sheet with convective conditions, Nanotechnol. Rev., 13 (2024), 20240031. https://doi.org/10.1515/ntrev-2024-0031 doi: 10.1515/ntrev-2024-0031
    [24] H. Yasmin, L. A. AL-Essa, R. Bossly, H. Alrabaiah, S. A. Lone, A. Saeed, A numerical investigation of the magnetized water-based hybrid nanofluid flow over an extending sheet with a convective condition: active and passive controls of nanoparticles, Nanotechnol. Rev., 13 (2024), 20240035. https://doi.org/10.1515/ntrev-2024-0035 doi: 10.1515/ntrev-2024-0035
    [25] S. U. S. Choi, J. A. Eastman, Enhancing thermal conductivity of fluids with nanoparticles, Argonne National Lab. (ANL), Argonne, IL (United States), 1995.
    [26] M. R. Khan, V. Puneeth, M. K. Alaoui, R. Alroobaea, M. M. M. Abdou, Heat transfer in a dissipative nanofluid passing by a convective stretching/shrinking cylinder near the stagnation point, ZAMM-J. Appl. Math. Mech./Z. Angew. Math. Mech., 104 (2024), e202300733. https://doi.org/10.1002/zamm.202300733 doi: 10.1002/zamm.202300733
    [27] L. Ali, Y. J. Wu, B. Ali, S. Abdal, S. Hussain, The crucial features of aggregation in TiO2-water nanofluid aligned of chemically comprising microorganisms: a FEM approach, Comput. Math. Appl., 123 (2022), 241–251. https://doi.org/10.1016/j.camwa.2022.08.028 doi: 10.1016/j.camwa.2022.08.028
    [28] A. Raza, A. Ghaffari, S. U. Khan, A. U. Haq, M. I. Khan, M. R. Khan, Non-singular fractional computations for the radiative heat and mass transfer phenomenon subject to mixed convection and slip boundary effects, Chaos Soliton. Fract., 155 (2022), 111708. https://doi.org/10.1016/j.chaos.2021.111708 doi: 10.1016/j.chaos.2021.111708
    [29] M. Sheikholeslami, T. Hayat, A. Alsaedi, MHD free convection of Al2O3-water nanofluid considering thermal radiation: a numerical study, Int. J. Heat Mass Transfer, 96 (2016), 513–524. https://doi.org/10.1016/j.ijheatmasstransfer.2016.01.059 doi: 10.1016/j.ijheatmasstransfer.2016.01.059
    [30] A. M. Alqahtani, M. R. Khan, N. Akkurt, V. Puneeth, A. Alhowaity, H. Hamam, Thermal analysis of a radiative nanofluid over a stretching/shrinking cylinder with viscous dissipation, Chem. Phys. Lett., 808 (2022), 140133. https://doi.org/10.1016/j.cplett.2022.140133 doi: 10.1016/j.cplett.2022.140133
    [31] W. U. Jan, M. Farooq, R. A. Shah, A. Khan, M. S. Zobaer, R. Jan, Flow dynamics of the homogeneous and heterogeneous reactions with an internal heat generation and thermal radiation between two squeezing plates, Mathematics, 9 (2021), 2309. https://doi.org/10.3390/math9182309 doi: 10.3390/math9182309
    [32] L. Ali, P. Kumar, H. Poonia, S. Areekara, R. Apsari, The significant role of Darcy-Forchheimer and thermal radiation on Casson fluid flow subject to stretching surface: a case study of dusty fluid, Mod. Phys. Lett. B, 38 (2024), 2350215. https://doi.org/10.1142/S0217984923502159 doi: 10.1142/S0217984923502159
    [33] M. M. Bhatti, M. M. Rashidi, Numerical simulation of entropy generation on MHD nanofluid towards a stagnation point flow over a stretching surface, Int. J. Appl. Comput. Math., 3 (2017), 2275–2289. https://doi.org/10.1007/s40819-016-0193-4 doi: 10.1007/s40819-016-0193-4
    [34] W. Ibrahim, B. Shankar, M. M. Nandeppanavar, MHD stagnation point flow and heat transfer due to nanofluid towards a stretching sheet, Int. J. Heat Mass Transfer, 56 (2013), 1–9. https://doi.org/10.1016/j.ijheatmasstransfer.2012.08.034 doi: 10.1016/j.ijheatmasstransfer.2012.08.034
    [35] L. Ali, P. Kumar, Z. Iqbal, S. E. Alhazmi, S. Areekara, M. M. Alqarni, et al., The optimization of heat transfer in thermally convective micropolar-based nanofluid flow by the influence of nanoparticle's diameter and nanolayer via stretching sheet: sensitivity analysis approach, J. Non-Equil. Thermody., 48 (2023), 313–330. https://doi.org/10.1515/jnet-2022-0064 doi: 10.1515/jnet-2022-0064
    [36] L. Ahmad, M. Irfan, S. Javed, M. I. Khan, M. R. Khan, U. M. Niazi, et al., Influential study of novel microorganism and nanoparticles during heat and mass transport in Homann flow of visco-elastic materials, Int. Commun. Heat Mass Transfer, 131 (2022), 105871. https://doi.org/10.1016/j.icheatmasstransfer.2021.105871 doi: 10.1016/j.icheatmasstransfer.2021.105871
    [37] L. Ali, B. Ali, M. B. Ghori, Melting effect on Cattaneo-Christov and thermal radiation features for aligned MHD nanofluid flow comprising microorganisms to leading edge: FEM approach, Comput. Math. Appl., 109 (2022), 260–269. https://doi.org/10.1016/j.camwa.2022.01.009 doi: 10.1016/j.camwa.2022.01.009
    [38] S. Nadeem, S. Ahmad, N. Muhammad, Computational study of Falkner-Skan problem for a static and moving wedge, Sensors Actuat. B: Chem., 263 (2018), 69–76. https://doi.org/10.1016/j.snb.2018.02.039 doi: 10.1016/j.snb.2018.02.039
    [39] L. Ali, R. Apsari, A. Abbas, P. Tak, Entropy generation on the dynamics of volume fraction of nano-particles and coriolis force impacts on mixed convective nanofluid flow with significant magnetic effect, Numer. Heat Transfer, Part A, 2024, 1–16. https://doi.org/10.1080/10407782.2024.2360652 doi: 10.1080/10407782.2024.2360652
    [40] Z. Hu, W. Lu, M. D. Thouless, Slip and wear at a corner with Coulomb friction and an interfacial strength, Wear, 338-339 (2015), 242–251. https://doi.org/10.1016/j.wear.2015.06.010
    [41] Z. Hu, W. Lu, M. D. Thouless, J. R. Barber, Effect of plastic deformation on the evolution of wear and local stress fields in fretting, Int. J. Solids Struct., 82 (2016), 1–8. https://doi.org/10.1016/j.ijsolstr.2015.12.031 doi: 10.1016/j.ijsolstr.2015.12.031
    [42] H. Wang, Z. Hu, W. Lu, M. D. Thouless, The effect of coupled wear and creep during grid-to-rod fretting, Nucl. Eng. Des., 318 (2017), 163–173. https://doi.org/10.1016/j.nucengdes.2017.04.018 doi: 10.1016/j.nucengdes.2017.04.018
    [43] T. Von Karman, Uberlaminare und turbulente Reibung, Z. Angew. Math Mech., 1 (1921), 233–252.
    [44] W. G. Cochran, The flow due to a rotating disk, Math. Proc. Cambridge Philos. Soc., 30 (1934), 365–375. https://doi.org/10.1017/S0305004100012561 doi: 10.1017/S0305004100012561
    [45] J. A. D. Ackroyd, On the steady flow produced by a rotating disk with either surface suction or injection, J. Eng. Math., 12 (1978), 207–220. https://doi.org/10.1007/BF00036459 doi: 10.1007/BF00036459
    [46] K. Millsaps, K. Pohlhausen, Heat transfer by laminar flow from a rotating plate, J. Aeronaut. Sci., 19 (1952), 120–126. https://doi.org/10.2514/8.2175 doi: 10.2514/8.2175
    [47] M. Miklavčič, C. Y. Wang, The flow due to a rough rotating disk, Z. Angew. Math. Phys., 55 (2004), 235–246. https://doi.org/10.1007/s00033-003-2096-6 doi: 10.1007/s00033-003-2096-6
    [48] R. Alizadeh, N. Karimi, R. Arjmandzadeh, A. Mehdizadeh, Mixed convection and thermodynamic irreversibilities in MHD nanofluid stagnation-point flows over a cylinder embedded in porous media, J. Therm. Anal. Calorim., 135 (2019), 489–506. https://doi.org/10.1007/s10973-018-7071-8 doi: 10.1007/s10973-018-7071-8
    [49] Z. Shah, H. Babazadeh, P. Kumam, A. Shafee, P. Thounthong, Numerical simulation of magnetohydrodynamic nanofluids under the influence of shape factor and thermal transport in a porous media using CVFEM, Front. Phys., 7 (2019), 164. https://doi.org/10.3389/fphy.2019.00164 doi: 10.3389/fphy.2019.00164
    [50] H. A. Attia, Steady flow over a rotating disk in porous medium with heat transfer, Nonlinear Anal., 14 (2009), 21–26. https://doi.org/10.15388/NA.2009.14.1.14527 doi: 10.15388/NA.2009.14.1.14527
    [51] M. Turkyilmazoglu, P. Senel, Heat and mass transfer of the flow due to a rotating rough and porous disk, Int. J. Thermal Sci., 63 (2013), 146–158. https://doi.org/10.1016/j.ijthermalsci.2012.07.013 doi: 10.1016/j.ijthermalsci.2012.07.013
    [52] Z. Shah, A. Dawar, I. Khan, S. Islam, D. L. C. Ching, A. Z. Khan, Cattaneo-Christov model for electrical magnetite micropoler Casson ferrofluid over a stretching/shrinking sheet using effective thermal conductivity model, Case Stud. Therm. Eng., 13 (2019), 100352. https://doi.org/10.1016/j.csite.2018.11.003 doi: 10.1016/j.csite.2018.11.003
    [53] M. Mustafa, J. A. Khan, T. Hayat, A. Alsaedi, On Bö dewadt flow and heat transfer of nanofluids over a stretching stationary disk, J. Mol. Liq., 211 (2015), 119–125. https://doi.org/10.1016/j.molliq.2015.06.065 doi: 10.1016/j.molliq.2015.06.065
    [54] M. M. Rashidi, N. Kavyani, S. Abelman, Investigation of entropy generation in MHD and slip flow over a rotating porous disk with variable properties, Int. J. Heat Mass Transfer, 70 (2014), 892–917. https://doi.org/10.1016/j.ijheatmasstransfer.2013.11.058 doi: 10.1016/j.ijheatmasstransfer.2013.11.058
    [55] T. Hayat, T. Muhammad, S. A. Shehzad, A. Alsaedi, On magnetohydrodynamic flow of nanofluid due to a rotating disk with slip effect: a numerical study, Comput. Methods Appl. Mech. Eng., 315 (2017), 467–477. https://doi.org/10.1016/j.cma.2016.11.002 doi: 10.1016/j.cma.2016.11.002
    [56] T. Hayat, F. Haider, T. Muhammad, A. Alsaedi, On Darcy-Forchheimer flow of carbon nanotubes due to a rotating disk, Int. J. Heat Mass Transfer, 112 (2017), 248–254. https://doi.org/10.1016/j.ijheatmasstransfer.2017.04.123 doi: 10.1016/j.ijheatmasstransfer.2017.04.123
    [57] H. Waqas, S. U. Khan, M. Hassan, M. M. Bhatti, M. Imran, Analysis on the bioconvection flow of modified second-grade nanofluid containing gyrotactic microorganisms and nanoparticles, J. Mol. Liq., 291 (2019), 111231. https://doi.org/10.1016/j.molliq.2019.111231 doi: 10.1016/j.molliq.2019.111231
    [58] I. Siddique, M. Nadeem, R. Ali, F. Jarad, Bioconvection of MHD second-grade fluid conveying nanoparticles over an exponentially stretching sheet: a biofuel applications, Arab. J. Sci. Eng., 48 (2023), 3367–3380. https://doi.org/10.1007/s13369-022-07129-1 doi: 10.1007/s13369-022-07129-1
    [59] A. Shafiq, G. Rasool, C. M. Khalique, S. Aslam, Second grade bioconvective nanofluid flow with buoyancy effect and chemical reaction, Symmetry, 12 (2020), 621. https://doi.org/10.3390/sym12040621 doi: 10.3390/sym12040621
    [60] S. Zuhra, N. S. Khan, Z. Shah, S. Islam, E. Bonyah, Simulation of bioconvection in the suspension of second grade nanofluid containing nanoparticles and gyrotactic microorganisms, Aip Adv., 8 (2018), 105210. https://doi.org/10.1063/1.5054679 doi: 10.1063/1.5054679
    [61] S. Abbas, S. F. F. Gilani, M. Nazar, M. Fatima, M. Ahmad, Z. U. Nisa, Bio-convection flow of fractionalized second grade fluid through a vertical channel with Fourier's and Fick's laws, Mod. Phys. Lett. B, 37 (2023), 2350069. https://doi.org/10.1142/S0217984923500690 doi: 10.1142/S0217984923500690
    [62] T. Hayat, Inayatullah, K. Muhammad, A. Alsaedi, Heat transfer analysis in bio-convection second grade nanofluid with Cattaneo-Christov heat flux model, Proc. Inst. Mech. Eng., Part E, 237 (2023), 1117–1124. https://doi.org/10.1177/09544089221097684 doi: 10.1177/09544089221097684
    [63] N. M. Sarif, M. Z. Salleh, R. Nazar, Numerical solution of flow and heat transfer over a stretching sheet with Newtonian heating using the Keller box method, Procedia Eng., 53 (2013), 542–554. https://doi.org/10.1016/j.proeng.2013.02.070 doi: 10.1016/j.proeng.2013.02.070
    [64] S. Liao, On the homotopy analysis method for nonlinear problems, Appl. Math. Comput., 147 (2004), 499–513. https://doi.org/10.1016/S0096-3003(02)00790-7 doi: 10.1016/S0096-3003(02)00790-7
    [65] J. Wang, X. Ye, A weak Galerkin finite element method for second-order elliptic problems, J. Comput. Appl. Math., 241 (2013), 103–115. https://doi.org/10.1016/j.cam.2012.10.003 doi: 10.1016/j.cam.2012.10.003
    [66] S. S. Motsa, P. G. Dlamini, M. Khumalo, Spectral relaxation method and spectral quasilinearization method for solving unsteady boundary layer flow problems, Adv. Math. Phys., 2014 (2014), 341964. https://doi.org/10.1155/2014/341964 doi: 10.1155/2014/341964
    [67] T. Hayat, T. Muhammad, A. Alsaedi, M. S. Alhuthali, Magnetohydrodynamic three-dimensional flow of viscoelastic nanofluid in the presence of nonlinear thermal radiation, J. Magn. Magn. Mater., 385 (2015), 222–229. https://doi.org/10.1016/j.jmmm.2015.02.046 doi: 10.1016/j.jmmm.2015.02.046
  • This article has been cited by:

    1. Adnène Arbi, Najeh Tahri, New results on time scales of pseudo Weyl almost periodic solution of delayed QVSICNNs, 2022, 41, 2238-3603, 10.1007/s40314-022-02003-0
    2. Marko Kostić, Weyl ρ-Almost Periodic Functions in General Metric, 2023, 73, 1337-2211, 465, 10.1515/ms-2023-0035
    3. Qi Shao, Yongkun Li, Almost periodic solutions for Clifford-valued stochastic shunting inhibitory cellular neural networks with mixed delays, 2024, 9, 2473-6988, 13439, 10.3934/math.2024655
    4. Adnène Arbi, Najeh Tahri, Synchronization analysis of novel delayed dynamical Clifford-valued neural networks on timescales, 2024, 18, 1748-3018, 10.1177/17483026241241492
  • Reader Comments
  • © 2024 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(688) PDF downloads(63) Cited by(0)

Figures and Tables

Figures(11)  /  Tables(2)

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog