Processing math: 62%
Research article Special Issues

Assessment of the BTEX concentrations and health risk in urban nursery schools in Gliwice, Poland

  • Indoor air quality (IAQ) in nursery school is believed to be different from elementary school. Moreover, younger children are more vulnerable to air pollution than higher grade children because they spend more time indoors, and their immune systems and bodies are less mature. The purpose of this study was to compare the concentrations of the monoaromatic volatile benzene, toluene, ethylbenzene m,p-xylene and o-xylene (BTEX) in urban nursery schools located in Gliwice, Poland. The nursery schools were chosen to include areas with different urbanization and traffic density characteristics in order to gather a more diverse picture of exposure risks in the various regions of the city. BTEX were sampled during winter and spring seasons in older and younger children classrooms. The samples were thermally desorbed (TD) and then analyzed with use of gas chromatography (GC). In addition, outdoor measurements were carried out in the playground at each nursery school. BTEX quantification, indoor/outdoor concentration, and correlation coefficients were used to identify pollutant sources. Elevated levels of o-xylene and ethylbenzene were found in all monitored classrooms during the winter season. Outdoor concentrations were lower than indoors for each classroom. Indicators based on health risk assessment for chronic health effects associated with carcinogenic benzene or non-carcinogenic BTEX were proposed to rank sites according to their hazard level.

    Citation: Anna Mainka, Barbara Kozielska. Assessment of the BTEX concentrations and health risk in urban nursery schools in Gliwice, Poland[J]. AIMS Environmental Science, 2016, 3(4): 858-870. doi: 10.3934/environsci.2016.4.858

    Related Papers:

    [1] Wenjun Liu, Zhijing Chen, Zhiyu Tu . New general decay result for a fourth-order Moore-Gibson-Thompson equation with memory. Electronic Research Archive, 2020, 28(1): 433-457. doi: 10.3934/era.2020025
    [2] Siyi Luo, Yinghui Zhang . Space-time decay rate for the 3D compressible quantum magnetohydrodynamic model. Electronic Research Archive, 2025, 33(7): 4184-4204. doi: 10.3934/era.2025189
    [3] Yi Cheng, Ying Chu . A class of fourth-order hyperbolic equations with strongly damped and nonlinear logarithmic terms. Electronic Research Archive, 2021, 29(6): 3867-3887. doi: 10.3934/era.2021066
    [4] Mohammad M. Al-Gharabli, Adel M. Al-Mahdi . Existence and stability results of a plate equation with nonlinear damping and source term. Electronic Research Archive, 2022, 30(11): 4038-4065. doi: 10.3934/era.2022205
    [5] Chahrazed Messikh, Soraya Labidi, Ahmed Bchatnia, Foued Mtiri . Energy decay for a porous system with a fractional operator in the memory. Electronic Research Archive, 2025, 33(4): 2195-2215. doi: 10.3934/era.2025096
    [6] Jincheng Shi, Shuman Li, Cuntao Xiao, Yan Liu . Spatial behavior for the quasi-static heat conduction within the second gradient of type Ⅲ. Electronic Research Archive, 2024, 32(11): 6235-6257. doi: 10.3934/era.2024290
    [7] Jie Qi, Weike Wang . Global solutions to the Cauchy problem of BNSP equations in some classes of large data. Electronic Research Archive, 2024, 32(9): 5496-5541. doi: 10.3934/era.2024255
    [8] Xu Liu, Jun Zhou . Initial-boundary value problem for a fourth-order plate equation with Hardy-Hénon potential and polynomial nonlinearity. Electronic Research Archive, 2020, 28(2): 599-625. doi: 10.3934/era.2020032
    [9] Qin Ye . Space-time decay rate of high-order spatial derivative of solution for 3D compressible Euler equations with damping. Electronic Research Archive, 2023, 31(7): 3879-3894. doi: 10.3934/era.2023197
    [10] Huafei Di, Yadong Shang, Jiali Yu . Existence and uniform decay estimates for the fourth order wave equation with nonlinear boundary damping and interior source. Electronic Research Archive, 2020, 28(1): 221-261. doi: 10.3934/era.2020015
  • Indoor air quality (IAQ) in nursery school is believed to be different from elementary school. Moreover, younger children are more vulnerable to air pollution than higher grade children because they spend more time indoors, and their immune systems and bodies are less mature. The purpose of this study was to compare the concentrations of the monoaromatic volatile benzene, toluene, ethylbenzene m,p-xylene and o-xylene (BTEX) in urban nursery schools located in Gliwice, Poland. The nursery schools were chosen to include areas with different urbanization and traffic density characteristics in order to gather a more diverse picture of exposure risks in the various regions of the city. BTEX were sampled during winter and spring seasons in older and younger children classrooms. The samples were thermally desorbed (TD) and then analyzed with use of gas chromatography (GC). In addition, outdoor measurements were carried out in the playground at each nursery school. BTEX quantification, indoor/outdoor concentration, and correlation coefficients were used to identify pollutant sources. Elevated levels of o-xylene and ethylbenzene were found in all monitored classrooms during the winter season. Outdoor concentrations were lower than indoors for each classroom. Indicators based on health risk assessment for chronic health effects associated with carcinogenic benzene or non-carcinogenic BTEX were proposed to rank sites according to their hazard level.


    The Moore-Gibson-Thompson (MGT) equation is one of the equations of nonlinear acoustics describing acoustic wave propagation in gases and liquids [13,15,30] and arising from modeling high frequency ultrasound waves [9,18] accounting for viscosity and heat conductivity as well as effect of the radiation of heat on the propagation of sound. This research field is highly active due to a wide range of applications such as the medical and industrial use of high intensity ultrasound in lithotripsy, thermotherapy, ultraound cleaning, etc. The classical nonlinear acoustics models include the Kuznetson's equation, the Westervelt equation and the Kokhlov-Zabolotskaya-Kuznetsov equation.

    In order to gain a better understanding of the nonlinear MGT equation, we shall begin with the linearized model. In [15], Kaltenbacher, Lasiecka and Marchand investigated the following linearized MGT equation

    τuttt+αutt+c2Au+bAut=0. (1.1)

    For equation (1.1), they disclosed a critical parameter γ=αc2τb and showed that when γ>0, namely in the subcritical case, the problem is well-posed and its solution is exponentially stable; while γ=0, the energy is conserved. Since its appearance, an increasing interest has been developed to study the MGT equation, see [4,5,8,10,14]. Caixeta, Lasiecka and Cavalcanti [4] considered the following nonlinear equation

    τuttt+αutt+c2Au+bAut=f(u,ut,utt). (1.2)

    They proved that the underlying PDE generates a well-posed dynamical system which admits a global and finite dimensional attractor. They also overcomed the difficulty of lacking the Lyapunov function and the lack of compactness of the trajectory.

    Now, we concentrate on the stabilization of MGT equation with memory which has received a considerable attention recently. For instance, Lasiecka and Wang [17] studied the following equation:

    τuttt+αutt+bAut+c2Aut0g(ts)Aw(s)ds=0, (1.3)

    where αc2τb0 and the form of w classifies the memory into three types. By imposing the assumption on the relaxation function g, for a positive constant c0, as

    g(t)c0g(t), (1.4)

    they discussed the effect of memory described by three types on decay rates of the energy when αc2τb>0. Moreover, in the critical case αc2τb=0, they proved an exponential rate of decay for the solution of (1.3) under "the right mixture" of memory. Lasiecka and Wang [18] showed the general decay result of the equation (1.3) when w=u, and established their result under weaker condition on g. In [9], Filippo et al. investigated the critical case of equation (1.3) (that is αbc2τ=0) for w=u and g satisfies (1.4), and obtained an exponential decay result if and only if A is a bounded operator. When t0 is replaced by 0, (1.3) turns to

    τuttt+αutt+bAut+c2Au0g(s)Aw(ts)ds=0. (1.5)

    Alves et al. [1] investigated the uniform stability of equation (1.5) encompassing three different types of memory in a history space set by the linear semigroup theory. Moreover, we refer the reader to [3,6,7,12,24,25,26,28] for other works of the equation(s) with memory.

    More recently, Filippo and Vittorino [11] considered the fourth-order MGT equation

    utttt+αuttt+βutt+γAutt+δAut+ϱAu=0. (1.6)

    They investigated the stability properties of the related solution semigroup. And, according to the values of certain stability numbers depending on the strictly positive parameters α, β, γ, δ, ϱ, they established the necessary and sufficient condition for exponential stability. For other related results on the higher-order equations, please see [20,27,34,35,36,37] and the references therein.

    Motivated by the above results, we intend to study the following abstract version of the fourth-order Moore-Gibson-Thompson (MGT) equation with a memory term

    utttt+αuttt+βutt+γAutt+δAut+ϱAut0g(ts)Au(s)ds=0, (1.7)

    where α, β, γ, δ, ϱ are strictly positive constants, A is a strictly positive self-adjoint linear operator defined in a real Hilbert space H where the (dense) embedding D(A)H need not to be compact. And we consider the following initial conditions

    u(0)=u0,ut(0)=u1,utt(0)=u2,uttt(0)=u3. (1.8)

    A natural question that arised in dealing with the general decay of fourth-order MGT equation with memory:

    ● Can we get a general decay result for a class of relaxation functions satisfying g(t)ξ(t)M(g(t)) for M to be increasing and convex function near the origin and ξ(t) to be a nonincreasing function?

    Mustafa answered this question for viscoelastic wave equations in [31,32]. Messaoudi and Hassan [29] considered the similar question for memory-type Timoshenko system in the cases of equal and non-equal speeds of wave propagation. Moreover, they extended the range of polynomial decay rate optimality from p[1,32) to p[1,2) when g satisfies g(t)ξ(t)gp(t). We refer to [19] for the non-equal wave speeds case. And, Liu et al. [22,23] also concerned with the similar question for third-order MGT equations with memory term.

    The aim of this paper is to establish the well-posedness and answer the above mention question for fourth-order MGT equation with memory (1.7). We first use the Faedo-Galerkin method to prove the well-posedness result. We then use the idea developed by Mustafa in [31,32], taking into consideration the nature of fourth-order MGT equation, to prove new general decay results for the case γδα>0 and βαϱδ>0, based on the perturbed energy method and on some properties of convex functions. Our result substantially improves and generalizes the earlier related results in previous literature.

    The rest of our paper is organized as follows. In Section 2, we give some assumptions and state our main results. In Section 3, we give the proof of well-posedness. In Section 4, we state and prove some technical lemmas that are relevant in the entire work. In Section 5, we prove the general decay result.

    In this section, we consider the following assumptions and state our main results. We use c>0 to denote a positive constant which does not depend on the initial data.

    First, we consider the following assumptions as in [11] for (A1), in [18] for (A3), (A5) and in [31] for (A2), (A4):

    (A1) γδα>0 and βαϱδ>0.

    (A2) g:R+R+  is a non-increasing differentiable function such that

    0<g(0)<2αϱδ(αγδ),    ϱ+0g(s)ds=l>0.

    (A3) g(t)0 almost everywhere.

    (A4) There exists a non-increasing differentiable function ξ : R+R+ and a C1 function M: [0,)[0,) which is either linear or strictly increasing and strictly convex C2 function on (0,r], rg(0), with M(0)=M(0)=0, such that

    g(t)ξ(t)M(g(t)), t0. (2.1)

    (A5) There exists λ0>0 such that A satisfies u2λ0A12u2 for all uH.

    Remark 1. ([31,Remark 2.8]) (1) From assumption (A2), we deduce that

    g(t)0ast+andg(t)ϱlt, t>0.

    Furthermore, from the assumption (A4), we obtain that there exists t00 large enough such that

    g(t0)=randg(t)r, tt0.

    The non-increasing property of g(t) and ξ(t) gives

    0<g(t0)g(t)g(0)and0<ξ(t0)ξ(t)ξ(0), t[0,t0].

    A combination of these with the continuity of H, for two constants a,d>0, yields

    aξ(t)M(g(t))d, t[0,t0].

    Consequently, for any t[0,t0], we get

    g(t)ξ(t)M(g(t))a=ag(0)g(0)ag(0)g(t)

    and, hence,

    g(t)g(0)ag(t), t[0,t0]. (2.2)

    (2) If M is a strictly increasing and strictly convex C2 function on (0,r], with M(0)=M(0)=0, then it has an extension ¯M, which is strictly increasing and strictly convex C2 function on (0,). For example, if we set M(r)=A, M(r)=B, M(r)=C, we can define ¯M, for any t>r, by

    ¯M=C2t2+(BCr)t+(A+C2r2Br).

    Then, inspired by the notations in [11], we define the Hilbert spaces

    Hr:=D(Ar2),rR.

    In order to simplify the notation, we denote the usual space H0 by H. The phase space of our problem is

    H=D(A12)×D(A12)×D(A12)×H.

    Moreover, we will denote the inner product of H by (,) and its norm by .

    After that, we introduce the following energy functional

    E(t)=12[uttt+αutt+αϱδut2+δα(ϱG(t)ϱ)A12utt+αA12ut+αϱδA12u2+δαϱG(t)A12utt+αA12ut2+(γδα)A12utt2+(γδα)αϱδA12ut2+2t0g(ts)(A12u(t)A12u(s),A12utt+αA12ut)ds+αϱδ(gA12u)(t)α(gA12u)(t)+αg(t)A12u2+(βαϱδ)utt2+αϱδ(βαϱδ)ut2],

    where G(t)=t0g(s)ds and for any vL2loc(R+;L2(Ω)),

    (gv)(t):=Ωt0g(ts)(v(t)v(s))2dsdx.

    As in [31], we set, for any 0<ν<1,

    Cν=0g2(s)νg(s)g(s)dsandh(t)=νg(t)g(t).

    The following lemmas play an important role in the proof of our main results.

    Lemma 2.1. ([31] ) Assume that condition (A2) holds. Then for any uL2loc(R+; L2(Ω)), we have

    Ω(t0g(ts)(A12u(s)A12u(t))ds)2dxCν(hA12u)(t), t0.

    Lemma 2.2. (Jensen's inequality) Let P: [b,c]R be a convex function. Assume that functions f:Ω[b,c] and h:ΩR are integrable such that h(x)0, for any xΩ and Ωh(x)dx=k>0. Then

    P(1kΩf(x)h(x)dx)1kΩP(f(x))h(x)dx.

    Lemma 2.3. ([2])(The generalized Young inequality) If f is a convex function defined on a real vector space X and ite convex conjugate is denoted by f, then

    ABf(A)+f(B), (2.3)

    where

    f(s)=s(f)1(s)f[(f)1(s)]. (2.4)

    We are now in a position to state the well-posedness and the general decay result for problem (1.7)-(1.8).

    Theorem 2.4. (Well-posedness) Assume that (A1)(A5) hold. Then, for given (u0,u1,u2,u3)H and T>0, there exists a unique weak solution u of problem (1.7)-(1.8) such that

    uC([0,T];D(A12))C1([0,T];H).

    Theorem 2.5. (General decay) Let (u0,u1,u2,u3)H. Assume that (A1)-(A5) hold. Then there exist positive constants k1 and k2 such that, along the solution of problem (1.7)-(1.8), the energy functional satisfies

    E(t)k2M11(k1tg1(r)ξ(s)ds), tg1(t), (2.5)

    where M1(t)=rt1sM(s)ds and M1 is strictly decreasing and convex on (0,r], with limt0M1(t)=+.

    Remark 2. Assume that M(s)=sp, 1p<2 in (A4), then by simple calculations, we see that the decay rate of E(t) is given by, for constants ¯k, ˜k and C,

    E(t){Cexp(˜kt0ξ(s)ds),ifp=1,¯k(1+t0ξ(s)ds)1p1,if1<p<2. (2.6)

    In this section, we will prove the global existence and uniqueness of the solution of problem (1.7)-(1.8). Firstly, we give the following lemmas.

    Lemma 3.1. If 0<g(0)<2αϱδ(αγδ), then there is σ>0 such that α(γδα)δg(0)2(ασ)ϱ>0.

    Proof. Since g(0)<2αϱδ(αγδ), all we need is to show

    2αϱδ(ασ)(γδα)2αϱδ(αγδ)asσ0,

    which is trivially true.

    Lemma 3.2. Assume that (A1)-(A5) hold. Then, the energy functional E(t) satisfies, for all t0,

    12[uttt+αutt+αϱδut2+δα(ϱG(t)ϱ)A12utt+αA12ut+αϱδA12u2+(γδα)A12utt2+(γδα)αϱδA12ut2α(gA12u)(t)+αg(t)A12u2+(βαϱδ)utt2+αϱδ(βαϱδ)ut2]E(t)12[uttt+αutt+αϱδut2+δα(ϱG(t)ϱ)A12utt+αA12ut+αϱδA12u2+(γδα)A12utt2+(γδα)αϱδA12ut2α(gA12u)(t)+2αϱδ(gA12u)(t)+αg(t)A12u2+(βαϱδ)utt2+αϱδ(βαϱδ)ut2+2δαϱG(t)A12utt+αA12ut2].

    Proof. From the definition of E(t), we have

    E(t)=12[uttt+αutt+αϱδut2+δα(ϱG(t)ϱ)A12utt+αA12ut+αϱδA12u2+δαϱG(t)A12utt+αA12ut2+(γδα)A12utt2+(γδα)αϱδA12ut2+2t0g(ts)(A12u(t)A12u(s),A12utt+αA12ut)ds+αϱδ(gA12u)(t)α(gA12u)(t)+αg(t)A12u2+(βαϱδ)utt2+αϱδ(βαϱδ)ut2].

    Then, we estimate the sixth term of the above equality

    2|t0g(ts)(A12u(t)A12u(s),A12utt+αA12ut)ds|t0g(ts)[αϱδA12u(t)A12u(s)2+δαϱA12utt+αA12ut2]ds=αϱδ(gA12u)(t)+δαϱG(t)A12utt+αA12ut2.

    A combination of the above results, we complete the proof of lemma.

    Now, we prove the well-posedness result of problem (1.7)-(1.8).

    Proof of Theorem 2.1. The proof is given by Faedo-Galerkin method and combines arguments from [16,39,38]. We present only the main steps.

    Step 1. Approximate problem

    We construct approximations of the solution u by the Faedo-Galerkin method as follows. For every m1, let Wm=span{w1,,wm} be a Hilbertian basis of the space H10(Ω). We choose four sequences (um0), (um1), (um2) and (um3) in Wm such that um0u0 strongly in D(A12), um1u1 strongly in D(A12), um2u2 strongly in D(A12) and um3u3 strongly in H. We define now the approximations:

    um(t)=mj=1amj(t)wj(x), (3.1)

    where um(t) are solutions to the finite dimensional Cauchy problem (written in normal form):

    Ωumtttt(t)wjdx+αΩumttt(t)wjdx+βΩumtt(t)wjdx+γΩA12umtt(t)A12wjdx+δΩA12umt(t)A12wjdx+ϱΩA12um(t)A12wjdxt0g(ts)ΩA12um(t)A12wjdxds=0 (3.2)

    with initial conditions

    (um(0),umt(0),umtt(0),umttt(0))=(um0,um1,um2,um3). (3.3)

    According to the standard theory of ordinary differential equation, the finite dimensional problem (3.2)-(3.3) has a local solution (um(t),umt(t),umtt(t),umttt(t)) in some interval [0,Tm) with 0<TmT, for every mN. Next, we present some estimates that allow us to extend the local solutions to the interval [0,T], for any given T>0.

    Step 2. Weak solutions

    Multiplying equation (3.2) by amjttt+αamjtt+αϱδamjt and integrating over Ω, we have

    ddtEm(t)+α(βαϱδ)umtt2αg(t)2A12um2+δ2αϱg(t)A12umtt+αϱδA12um2+α2(gA12um)(t)=[α(γδα)δg(t)2αϱ]A12umtt2+t0g(ts)(A12um(t)A12um(s),A12umtt)ds+αϱ2δ(gA12um)(t),

    where

    Em(t)=12[umttt+αumtt+αϱδumt2+δα(ϱG(t)ϱ)A12umtt+αA12umt+αϱδA12um2+δαϱG(t)A12umtt+αA12umt2+(γδα)A12umtt2+(γδα)αϱδA12umt2+2t0g(ts)(A12um(t)A12um(s),A12umtt+αA12umt)ds+αϱδ(gA12um)(t)α(gA12um)(t)+αg(t)A12um2+(βαϱδ)umtt2+αϱδ(βαϱδ)umt2]. (3.4)

    From assumptions (A1)(A3) and Lemma 3.1, we get, for ε(0,α),

    t0g(ts)(A12um(t)A12um(s),A12umtt)ds(αε)ϱ2δ(gA12um)(t)δ2(αε)ϱA12umtt2t0g(ts)ds=(αε)ϱ2δ(gA12um)(t)+δ(g(0)g(t))2(αε)ϱA12umtt2

    and so

    [α(γδα)δg(t)2αϱ]A12umtt2+t0g(ts)(A12um(t)A12um(s),A12umtt)ds+αϱ2δ(gA12um)(t)[α(γδα)δg(t)2αϱ]A12umtt2(αε)ϱ2δ(gA12um)(t)+δ(g(0)g(t))2(αε)ϱA12umtt2+αϱ2δ(gA12um)(t)=[α(γδα)δg(0)2(αε)ϱ]A12umtt2+εϱ2δ(gA12um)(t)[δg(t)2(αε)ϱδg(t)2αϱ]A12umtt20. (3.5)

    Therefore, we have

    ddtEm(t)+α(βαϱδ)umtt2αg(t)2A12um2+δ2αϱg(t)A12umtt+αϱδA12um2+α2(gA12um)(t)0. (3.6)

    Integrating (3.6) from 0 to tTm, one has

    Em(t)+t0[α(βαϱδ)umtt2αg(τ)2A12um2+δ2αϱg(τ)A12umtt+αϱδA12um2+α2(gA12um)(τ)]dτEm(0). (3.7)

    Now, since the sequences (um0)mN, (um1)mN, (um2)mN and (um3)mN converge and using (A1)(A3), we can find a positive constant C independent of m such that

    Em(t)C. (3.8)

    Therefore, using the fact ϱt0g(s)dsl, the last estimate (3.8) together with (3.4) give us, for all mN,Tm=T, we deduce that

    (um)mNisboundedinL(0,T;D(A12))(umt)mNisboundedinL(0,T;D(A12))(umtt)mNisboundedinL(0,T;D(A12))(umttt)mNisboundedinL(0,T;H). (3.9)

    Consequently, we may conclude that

    umuweakinL(0,T;D(A12))umtutweakinL(0,T;D(A12))umttuttweakinL(0,T;D(A12))umtttutttweakinL(0,T;H).

    From (3.9), we get that (um)mN is bounded in L(0,T;D(A12)). Then, (um)mN is bounded in L2(0,T;D(A12)). Since (umt)mN is bounded in L(0,T;D(A12)), (umt)mN is bounded in L2(0,T;D(A12)). Consequently, (umtt)mN is bounded in L2(0,T;D(A12)) and (umttt)mN is bounded in L2(0,T;H). Moreover, (um)mN is bounded in H3(0,T;H1(Ω)).

    Since the embedding H3(0,T;H1(Ω))L2(0,T;H(Ω)) is compact, using Aubin-Lions theorem [21], we can extract a subsequence (un)nN of (um)mN such that

    unustronglyinL2(0,T;H(Ω)).

    Therefore,

    unustronglyanda.e.on(0,T)×Ω.

    The proof now can be completed arguing as in [21].

    Step 3. Uniqueness

    It is sufficient to show that the only weak solution of (1.7)-(1.8) with u0=u1=u2=u3=0 is

    u0. (3.10)

    According to the energy estimate (3.8) and noting that E(u(0))=0, we obtain

    E(u(t))=0,t[0,T].

    So, we have

    uttt+αutt+αϱδut2=A12utt+αA12ut+αϱδA12u2=A12utt2=A12ut2=A12u2=utt2=0,t[0,T].

    And this implies (3.10). Thus, we conclude that problem (1.7)-(1.8) has at most one solution.

    In this section, we state and prove some lemmas needed to establish our general decay result.

    Lemma 4.1. Let (u,ut,utt,uttt) be the solution of (1.7). Assume that (A1)-(A3) hold. Then, we have

    ddtE(t)α(βαϱδ)utt2[α(γδα)δg(0)2(αε)ϱ]A12utt2+αg(t)2A12u2[δg(t)2(αε)ϱδg(t)2αϱ]A12utt2δ2αϱg(t)A12utt+αϱδA12u2α2(gA12u)(t)+εϱ2δ(gA12u)0.

    Proof. Multiplying (1.7) by uttt+αutt+αϱδut and integrating over Ω yield

    ddtE(t)=α(βαϱδ)utt2[α(γδα)δg(t)2αϱ]A12utt2+αg(t)2A12u2δ2αϱg(t)A12utt+αϱδA12u2+t0g(ts)(A12u(t)A12u(s),A12utt)dsα2(gA12u)(t)+αϱ2δ(gA12u)(t). (4.1)

    We proceed to show that, for a constant ε(0,α),

    |t0g(ts)(A12u(t)A12u(s),A12utt)ds|(αε)ϱ2δ(gA12u)(t)+δ(g(0)g(t))2(αε)ϱA12utt2. (4.2)

    Then, combining (4.1) and (4.2), we can obtain

    ddtE(t)α(βαϱδ)utt2[α(γδα)δg(0)2(αε)ϱ]A12utt2+αg(t)2A12u2[δg(t)2(αε)ϱδg(t)2αϱ]A12utt2δ2αϱg(t)A12utt+αϱδA12u2α2(gA12u)(t)+εϱ2δ(gA12u).

    According to (A1)-(A3) and Lemma 3.1, we complete the proof of lemma.

    Lemma 4.2. Assume that (A1)-(A5) hold. Then, the functional F1(t) defined by

    F1(t)=Ω(utt+αut+αϱδu)(uttt+αutt+αϱδut)dx

    satisfies the estimate

    F1(t)δ2αA12utt+αA12ut+αϱδA12u2+2αλ0δ(βαϱδ)2utt2+2αδ(γδα)2A12utt2+uttt+αutt+αϱδut2+2α(ϱl)2δA12u2+2αδCν(hA12u)(t). (4.3)

    Proof. Taking the derivative of F1(t) with respect to t, exploiting (1.7) and integrating by parts, we get

    F1(t)=Ω[(βαϱδ)utt](utt+αut+αϱδu)dxδαA12utt+αA12ut+αϱδA12u2+uttt+αutt+αϱδut2Ω(γδα)A12utt(A12utt+αA12ut+αϱδA12u)dx+Ω(t0g(ts)A12u(s)ds)(A12utt+αA12ut+αϱδA12u)dx.

    Using Young's inequality, Lemma 2.1, (A5) and the fact γδα>0 and βαϱδ>0, we have

    Ω[(βαϱδ)utt](utt+αut+αϱδu)dx2αλ0δ(βαϱδ)2utt2+δ8αλ0utt+αut+αϱδu22αλ0δ(βαϱδ)2utt2+δ8αA12utt+αA12ut+αϱδA12u2

    and

    Ω(t0g(ts)A12u(s)ds)(A12utt+αA12ut+αϱδA12u)dx=Ω(t0g(ts)(A12u(s)A12u(t))ds)(A12utt+αA12ut+αϱδA12u)dx+Ω(t0g(ts)A12u(t)ds)(A12utt+αA12ut+αϱδA12u)dx2αδCν(hA12u)(t)+δ4αA12utt+αA12ut+αϱδA12u2+2α(ϱl)2δA12u2.

    Also, we have

    Ω(γδα)A12utt(A12utt+αA12ut+αϱδA12u)dx2αδ(γδα)2A12utt2+δ8αA12utt+αA12ut+αϱδA12u2.

    Then, combining the above inequalities, we complete the proof of (4.3).

    Lemma 4.3. Assume that (A1)-(A5) hold. Then the functional F2(t) defined by

    F2(t)=Ω(uttt+αutt+αϱδut)t0g(ts)[(utt+αut+αϱδu)(t)αϱδu(s)]dsdx

    satisfies the estimate

    F2(t)G(t)4uttt+αutt+αϱδut2+[(ϱl)2α22ε1+4λ0g2(0)α2G(t)]A12ut2+[λ0(ϱl)22+(δ2α2+ϱ2)+3(ϱl)2]ε1A12utt+αA12ut+αϱδA12u2+[α2ϱ2(ϱl)2λ0ε12δ2+(ϱl)22ε1+3(ϱl)2(αϱδ)2ε1+(ϱl)22(αϱδ)2]A12u2+[(ϱl)22ε1+12ε1(γδα)2+12(αϱδ)2(γδα)2]A12utt2+[(ε1α2ϱ2λ04δ2+34ε1+1+αϱδ+2α4ϱ2λ0G(t)δ2)Cν+2α2ϱ2λ0G(t)δ2](hA12u)(t)+[2(βαϱδ)2ε1+4g2(0)G(t)]utt2, (4.4)

    where 0<\varepsilon_{1}<1 .

    Proof. By differentiating F_{2}(t) with respect to t , using (1.7) and integrating by parts, we obtain

    \begin{align*} &F'_{2}(t)\nonumber\\ = &\int_{\Omega}\left[\beta u_{tt}+\gamma\mathcal{A}u_{tt}+\delta\mathcal{A}u_{t}+\varrho\mathcal{A}u-\int_{0}^{t}g(t-s)\mathcal{A}u(s)ds -\frac{\alpha\varrho}{\delta}u_{tt}\right] \nonumber\\ &\times\int_{0}^{t}g(t-s)\left[\left(u_{tt}+\alpha u_{t} +\frac{\alpha\varrho}{\delta}u\right)(t)-\frac{\alpha\varrho}{\delta}u(s)\right]dsdx\nonumber\\ &-g(0)\int_{\Omega}\left(u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right)\left(u_{tt}+\alpha u_{t}\right)dx \nonumber\\ &-\int_{\Omega}\left(u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right)\int_{0}^{t}g'(t-s)\left[\left(u_{tt}+\alpha u_{t}+\frac{\alpha\varrho}{\delta}u\right)(t)\right. \nonumber\\ &\left.-\frac{\alpha\varrho}{\delta}u(s)\right]dsdx-\int_{0}^{t}g(s)ds\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2} \nonumber\\ = &\int_{\Omega}\left[\left(\beta-\frac{\alpha\varrho}{\delta}\right)u_{tt}+\frac{\delta}{\alpha}\left(\mathcal{A}u_{tt} +\alpha\mathcal{A}u_{t}+\frac{\alpha\varrho}{\delta}\mathcal{A}u\right) +\left(\gamma-\frac{\delta}{\alpha}\right)\mathcal{A}u_{tt}\right. \nonumber\\ &\left.-\int_{0}^{t}g(s)ds\mathcal{A}u(t)\right] \int_{0}^{t}g(t-s)\left[\left(u_{tt}+\alpha u_{t}+\frac{\alpha\varrho}{\delta}u\right)(t)- \frac{\alpha\varrho}{\delta}u(s)\right]dsdx\nonumber\\ &-\int_{0}^{t}g(s)ds\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2} \nonumber\\ &-\int_{\Omega}\left(u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right)\int_{0}^{t}g'(t-s)\left[\left(u_{tt}+\alpha u_{t}+\frac{\alpha\varrho}{\delta}u\right)(t)-\frac{\alpha\varrho}{\delta}u(s)\right]dsdx \nonumber\\ &+\left(\int_{0}^{t}g(s)ds\right)\int_{\Omega}\int_{0}^{t}g(t-s)\left(\mathcal{A}u(t)-\mathcal{A}u(s)\right)ds(u_{tt}+\alpha u_{t})dx \nonumber\\ &+\frac{\alpha\varrho}{\delta}\int_{\Omega}\left(\int_{0}^{t}g(t-s)\left(\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(s)\right)ds\right)^{2}dx \nonumber\\ &-g(0)\int_{\Omega}\left(u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right)\left(u_{tt}+\alpha u_{t}\right)dx. \end{align*}

    Now, we estimate the terms in the right-hand side of the above identity.

    Using Young's inequality, we obtain, for 0<\varepsilon_{1}<1 ,

    \begin{align*} &\int_{\Omega}\left[\left(\beta-\frac{\alpha\varrho}{\delta}\right)u_{tt}+\frac{\delta}{\alpha}\left(\mathcal{A}u_{tt} +\alpha\mathcal{A}u_{t}+\frac{\alpha\varrho}{\delta}\mathcal{A}u\right)+\left(\gamma-\frac{\delta}{\alpha}\right)\mathcal{A}u_{tt} \right. \nonumber\\ &\left.-\int_{0}^{t}g(s)ds\mathcal{A}u(t)\right] \int_{0}^{t}g(t-s)\left[\left(u_{tt}+\alpha u_{t}+\frac{\alpha\varrho}{\delta}u\right)-\frac{\alpha\varrho}{\delta}u(s)\right]dsdx \nonumber\\ \leq&\left[\frac{\lambda_{0}(\varrho-l)^{2}}{2}+\left(\frac{\delta^{2}}{\alpha^{2}}+\varrho^{2}\right) +2(\varrho-l)^{2}\right]\varepsilon_{1}\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}+\alpha\mathcal{A}^{\frac{1}{2}}u_{t} +\frac{\alpha\varrho}{\delta}\mathcal{A}^{\frac{1}{2}}u\right\|^{2}\nonumber\\ &+\frac{2\left(\beta-\frac{\alpha\varrho}{\delta}\right)^{2}} {\varepsilon_{1}}\|u_{tt}\|^{2}+\left[\frac{\alpha^{2}\varrho^{2}(\varrho-l)^{2}\lambda_{0}\varepsilon_{1}}{2\delta^{2}} +\frac{(\varrho-l)^{2}}{2\varepsilon_{1}}\right. \nonumber\\ &\left. +2(\varrho-l)^{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\varepsilon_{1}+\frac{(\varrho-l)^{2}}{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\right]\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2} \nonumber\\ &+\left[\frac{(\varrho-l)^{2}}{2\varepsilon_{1}}+\frac{1}{2\varepsilon_{1}}\left(\gamma -\frac{\delta}{\alpha}\right)^{2}+\frac{1}{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\left(\gamma -\frac{\delta}{\alpha}\right)^{2}\right]\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}\right\|^{2} \nonumber\\ &+\frac{(\varrho-l)^{2} \alpha^{2}}{2\varepsilon_{1}}\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2}+\left(\frac{\varepsilon_{1}\alpha^{2}\varrho^{2}\lambda_{0}}{4\delta^{2}}+\frac{1}{4\varepsilon_{1}}+1\right)C_{\nu}\left(h\circ\mathcal{A}^{\frac{1}{2}}u\right)(t) \end{align*}

    and

    \begin{align*} &\left(\int_{0}^{t}g(s)ds\right)\int_{\Omega}\int_{0}^{t}g(t-s)\left(\mathcal{A}u(t)-\mathcal{A}u(s)\right)ds(u_{tt}+\alpha u_{t})dx \nonumber\\ = &\left(\int_{0}^{t}g(s)ds\right)\int_{\Omega}\int_{0}^{t}g(t-s)\left(\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(s)\right)ds(\mathcal{A}^{\frac{1}{2}}u_{tt}+\alpha \mathcal{A}^{\frac{1}{2}}u_{t})dx \nonumber\\ \leq&\frac{1}{2\varepsilon_{1}}C_{\nu}\left(h\circ\mathcal{A}^{\frac{1}{2}}u\right)(t)+\frac{(\varrho-l)^{2}}{2}\varepsilon_{1}\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}+\alpha\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2} \nonumber\\ \leq&\frac{1}{2\varepsilon_{1}}C_{\nu}\left(h\circ\mathcal{A}^{\frac{1}{2}}u\right)(t)+(\varrho-l)^{2}\varepsilon_{1} \left\|\mathcal{A}^{\frac{1}{2}}u_{tt}+\alpha\mathcal{A}^{\frac{1}{2}}u_{t}+\frac{\alpha\varrho}{\delta}\mathcal{A}^{\frac{1}{2}}u\right\|^{2} \nonumber\\ &+(\varrho-l)^{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\varepsilon_{1}\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}. \end{align*}

    Also, we have

    \begin{align*} \frac{\alpha\varrho}{\delta}\int_{\Omega}\left(\int_{0}^{t}g(t-s)\left(\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(s)\right)ds\right)^{2}dx\leq&\frac{\alpha\varrho} {\delta}C_{\nu}\left(h\circ\mathcal{A}^{\frac{1}{2}}u\right)(t) \end{align*}

    and

    \begin{align*} &-g(0)\int_{\Omega}\left(u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right)\left(u_{tt}+\alpha u_{t}\right)dx \nonumber\\ \leq&\frac{G(t)}{4}\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2}+\frac{g^{2}(0)}{G(t)}\left\|u_{tt}+\alpha u_{t}\right\|^{2} \nonumber\\ \leq&\frac{G(t)}{4}\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2}+\frac{2g^{2}(0)}{G(t)}\left\|u_{tt}\right\|^{2}+\frac{2\lambda_{0}g^{2}(0)\alpha^{2}}{G(t)}\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2}. \end{align*}

    Exploiting Young's inequality and (A5) , we get

    \begin{align*} &-\int_{\Omega}\left(u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right)\int_{0}^{t}g'(t-s)\left[\left(u_{tt}+\alpha u_{t}+\frac{\alpha\varrho}{\delta}u\right)(t)-\frac{\alpha\varrho}{\delta}u(s)\right]dsdx \nonumber\\ = &-\int_{\Omega}\left(u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right)\int_{0}^{t}g'(t-s)\left(u_{tt}+\alpha u_{t}\right)(t)dsdx \nonumber\\ &-\frac{\alpha\varrho}{\delta}\int_{\Omega}\left(u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right)\int_{0}^{t}g'(t-s)(u(t)-u(s))dsdx \nonumber\\ \leq&\frac{G(t)}{2}\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2}+\frac{2g^{2}(0)}{G(t)}\|u_{tt}\|^{2}+\frac{2\lambda_{0}\alpha^{2}g^{2}(0)}{G(t)}\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2} \nonumber\\ &+\frac{2\alpha^{2}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}\left(\alpha^{2} C_{\nu}+1\right)\left(h\circ\mathcal{A}^{\frac{1}{2}}u\right)(t). \end{align*}

    A combination of all the above estimates gives the desired result.

    As in [11], we introduce the following auxiliary functional

    \begin{align*} F_{3}(t) = \int_{\Omega}\left(u_{ttt}+\alpha u_{tt}\right)u_{t}dx+\frac{\varrho}{2}\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}. \end{align*}

    Lemma 4.4. Assume that (A1) - (A5) hold. Then the functional F_{3}(t) satisfies the estimate

    \begin{align} F'_{3}(t)\leq&-\left(\frac{3\delta}{8}-\frac{\varepsilon_{2}\delta}{4}\right)\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2}+\frac{\varepsilon_{2}\delta^{3}}{8\alpha^{2}\varrho^{2}\lambda_{0}}\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2} \\ &+\frac{2\gamma^{2}}{\delta}\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}\right\|^{2}+\left(\frac{\varepsilon_{2}\delta^{3}}{4\varrho^{2}\lambda_{0}} +\frac{4\alpha^{2}\varrho^{2}\lambda_{0}}{\varepsilon_{2}\delta^{3}} +\frac{2\beta^{2}\lambda_{0}}{\delta}\right)\|u_{tt}\|^{2}\\ &+\frac{1}{\delta}C_{\nu}\left(h\circ\mathcal{A}^{\frac{1}{2}}u\right)(t)+\frac{2(\varrho-l)^{2}}{\delta}\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}, \end{align} (4.5)

    where 0<\varepsilon_{2}<1 .

    Proof. Using the equation (1.7), a direct computation leads to the following identity

    \begin{align} F'_{3}(t) = &\int_{\Omega}(u_{ttt}+\alpha u_{tt})u_{tt}dx+\int_{\Omega}(u_{tttt}+\alpha u_{ttt})u_{t}dx+\varrho\left(\mathcal{A}^{\frac{1}{2}}u,\mathcal{A}^{\frac{1}{2}}u_{t}\right) \\ = &(u_{ttt},u_{tt})+\alpha\|u_{tt}\|^{2}-\beta(u_{tt},u_{t})-\gamma\left(\mathcal{A}^{\frac{1}{2}}u_{tt},\mathcal{A}^{\frac{1}{2}}u_{t}\right)-\delta\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2} \\ &+\left(\int_{0}^{t}g(t-s)\mathcal{A}^{\frac{1}{2}}u(s)ds,\mathcal{A}^{\frac{1}{2}}u_{t}\right). \end{align} (4.6)

    Now, the first and third terms in the right-hand side of (4.6) can be estimated as follows:

    \begin{align*} & (u_{ttt},u_{tt}) \nonumber\\ \leq& \frac{\varepsilon_{2}\delta^{3}}{16\alpha^{2}\varrho^{2}\lambda_{0}}\|u_{ttt}\|^{2} +\frac{4\alpha^{2}\varrho^{2}\lambda_{0}}{\varepsilon_{2}\delta^{3}}\|u_{tt}\|^{2} \nonumber\\ \leq& \frac{\varepsilon_{2}\delta^{3}}{8\alpha^{2}\varrho^{2}\lambda_{0}}\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2} +\frac{\varepsilon_{2}\delta^{3}}{8\alpha^{2}\varrho^{2}\lambda_{0}}\left\|\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2} +\frac{4\alpha^{2}\varrho^{2}\lambda_{0}}{\varepsilon_{2}\delta^{3}}\|u_{tt}\|^{2} \nonumber\\ \leq& \frac{\varepsilon_{2}\delta^{3}}{8\alpha^{2}\varrho^{2}\lambda_{0}}\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2} +\alpha^{2}\left(\frac{\varepsilon_{2}\delta^{3}}{4\alpha^{2}\varrho^{2}\lambda_{0}}+\frac{4\varrho^{2}\lambda_{0}}{\varepsilon_{2}\delta^{3}}\right)\|u_{tt}\|^{2} +\frac{\varepsilon_{2}\delta}{4}\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2} \end{align*}

    and

    \begin{align*} -\beta(u_{tt},u_{t})\leq\frac{2\beta^{2}\lambda_{0}}{\delta}\|u_{tt}\|^{2}+\frac{\delta}{8\lambda_{0}}\|u_{t}\|^{2}\leq\frac{2\beta^{2}\lambda_{0}}{\delta}\|u_{tt}\|^{2}+\frac{\delta}{8}\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2}, \end{align*}

    where 0<\varepsilon_{2}<1 .

    Using Young's inequality and Lemma 2.1, we get

    \begin{align*} -\gamma\left(\mathcal{A}^{\frac{1}{2}}u_{tt},\mathcal{A}^{\frac{1}{2}}u_{t}\right)\leq\frac{2\gamma^{2}}{\delta}\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}\right\|^{2}+\frac{\delta}{8}\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2} \end{align*}

    and

    \begin{align*} & \left(\int_{0}^{t}g(t-s)\mathcal{A}^{\frac{1}{2}}u(s)ds,\mathcal{A}^{\frac{1}{2}}u_{t}\right) \nonumber\\ = & \left(\int_{0}^{t}g(t-s)\left(\mathcal{A}^{\frac{1}{2}}u(s)-\mathcal{A}^{\frac{1}{2}}u(t)+\mathcal{A}^{\frac{1}{2}}u(t)\right)ds,\mathcal{A}^{\frac{1}{2}}u_{t}\right) \nonumber\\ \leq& \frac{1}{\delta}\int_{\Omega}\left(\int_{0}^{t}g(t-s)\left(\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(s)\right)ds\right)^{2}dx \nonumber\\ &+\frac{\delta}{4}\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2} +\frac{2(\varrho-l)^{2}}{\delta}\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2} +\frac{\delta}{8}\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2} \nonumber\\ \leq& \frac{1}{\delta}C_{\nu}\left(h\circ\mathcal{A}^{\frac{1}{2}}u\right)(t) +\frac{3\delta}{8}\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2} +\frac{2(\varrho-l)^{2}}{\delta}\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}. \end{align*}

    Then, combining the above inequalities, we obtain the desired result.

    Lemma 4.5. Assume that (A1) - (A5) hold. Then the functional F_{4}(t) defined by

    \begin{align*} F_{4}(t) = \int_{\Omega}\int_{0}^{t}f(t-s)\left|\mathcal{A}^{\frac{1}{2}}u(s)\right|^{2}dsdx \end{align*}

    satisfies the estimate

    \begin{align} F'_{4}(t)\leq-\frac{1}{2}\left(g\circ\mathcal{A}^{\frac{1}{2}}u\right)(t)+3(\varrho-l)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}, \end{align} (4.7)

    where f(t) = \int_{t}^{\infty}g(s)ds .

    Proof. Noting that f'(t) = -g(t) , we see that

    \begin{align*} &F'_{4}(t)\nonumber\\ = &f(0)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}-\int_{\Omega}\int_{0}^{t}g(t-s)\left|\mathcal{A}^{\frac{1}{2}}u(s)\right|^{2}dsdx \nonumber\\ = &f(0)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}-\int_{\Omega}\int_{0}^{t}g(t-s)\left|\mathcal{A}^{\frac{1}{2}}u(s)-\mathcal{A}^{\frac{1}{2}}u(t)\right|^{2}dsdx \nonumber\\ &-2\int_{\Omega}\mathcal{A}^{\frac{1}{2}}u\int_{0}^{t}g(t-s)\left(\mathcal{A}^{\frac{1}{2}}u(s)-\mathcal{A}^{\frac{1}{2}}u(t)\right)dsdx-\left(\int_{0}^{t}g(s)ds\right)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2} \nonumber\\ = &-\left(g\circ\mathcal{A}^{\frac{1}{2}}u\right)(t)-2\int_{\Omega}\mathcal{A}^{\frac{1}{2}}u\int_{0}^{t}g(t-s)\left(\mathcal{A}^{\frac{1}{2}}u(s)-\mathcal{A}^{\frac{1}{2}}u(t)\right)dsdx+f(t)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}. \end{align*}

    Exploiting Young's inequality and the fact \int_{0}^{t}g(s)ds\leq \varrho-l , we obtain

    \begin{align*} &-2\int_{\Omega}\mathcal{A}^{\frac{1}{2}}u\int_{0}^{t}g(t-s)\left(\mathcal{A}^{\frac{1}{2}}u(s)-\mathcal{A}^{\frac{1}{2}}u(t)\right)dsdx \nonumber\\ \leq&2(\varrho-l)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}\nonumber\\ &+\frac{1}{2(\varrho-l)}\left(\int_{0}^{t}g(t-s)ds\right)\int_{\Omega}\int_{0}^{t}g(t-s)\left(\mathcal{A}^{\frac{1}{2}}u(s)-\mathcal{A}^{\frac{1}{2}}u(t)\right)^{2}dsdx \nonumber\\ \leq&2(\varrho-l)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}+\frac{1}{2}\left(g\circ\mathcal{A}^{\frac{1}{2}}u\right)(t). \end{align*}

    Moreover, taking account of f(t)\leq f(0) = \varrho-l , we have

    \begin{align*} f(t)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}\leq(\varrho-l)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}. \end{align*}

    Combining the above estimates, we arrive at the desired result.

    Lemma 4.6. Assume that (A1)-(A5) hold. The functional \mathcal{L}(t) defined by

    \begin{align*} \mathcal{L}(t) = NE(t)+F_{1}(t)+N_{2}F_{2}(t)+N_{3}F_{3}(t) \end{align*}

    satisfies, for a suitable choice of N, N_{2}, N_{3} ,

    \begin{align*} \mathcal{L}(t)\sim E(t) \end{align*}

    and the estimate, for all t\geq t_{0} ,

    \begin{align} \mathcal{L'}(t)\leq&-c\left[\|u_{tt}\|^{2}+\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}\right\|^{2}+\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2}\right. \\ &\left. +\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}+\alpha\mathcal{A}^{\frac{1}{2}}u_{t}+\frac{\alpha\varrho}{\delta}u\right\|^{2}\right]-4(\varrho-l)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}+\frac{1}{8}\left(g\circ\mathcal{A}^{\frac{1}{2}}u\right)(t), \end{align} (4.8)

    where t_{0} has been introduced in Remark 2.1 .

    Proof. Combining Lemmas 4.1-4.4 and recalling that g' = \nu g-h , we obtain, for all t\geq t_{0} ,

    \begin{align*} &\mathcal{L'}(t)\nonumber\\ \leq&-\left[\alpha\left(\beta-\frac{\alpha\varrho}{\delta}\right)N-\frac{2\alpha\lambda_{0}}{\delta}\left(\beta-\frac{\alpha\varrho}{\delta}\right)^{2}-\left(\frac{2\left(\beta-\frac{\alpha\varrho}{\delta}\right)^{2}}{\varepsilon_{1}}+\frac{4g^{2}(0)}{G(t)}\right)N_{2}\right. \nonumber\\ &\left.-\left(\frac{\varepsilon_{2}\delta^{3}}{4\varrho^{2}\lambda_{0}}+\frac{4\alpha^{2}\varrho^{2}\lambda_{0}}{\varepsilon_{2}\delta^{3}}+\frac{2\beta^{2}\lambda_{0}}{\delta}\right)N_{3}\right]\|u_{tt}\|^{2}-\left[\left(\alpha\left(\gamma-\frac{\delta}{\alpha}\right)-\frac{\delta g(0)}{2(\alpha-\varepsilon)\varrho}\right)N\right. \nonumber\\ &\left.+\left(\frac{\delta g(t)}{2(\alpha-\varepsilon)\varrho}-\frac{\delta g(t)}{2\alpha\varrho}\right)N-\frac{2\alpha}{\delta}\left(\gamma-\frac{\delta}{\alpha}\right)^{2}-\left(\frac{(\varrho-l)^{2}}{2\varepsilon_{1}}+\frac{1}{2\varepsilon_{1}}\left(\gamma-\frac{\delta}{\alpha}\right)^{2}\right.\right. \nonumber\\ &\left.\left.+\frac{1}{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\left(\gamma-\frac{\delta}{\alpha}\right)^{2}\right)N_{2}-\frac{2\gamma^{2}}{\delta}N_{3}\right]\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}\right\|^{2}-\frac{\alpha}{2}N\left(g''\circ\mathcal{A}^{\frac{1}{2}}u\right)(t) \nonumber\\ &-\left[\left(\frac{3\delta}{8}-\frac{\varepsilon_{2}\delta}{4}\right)N_{3}-\left(\frac{(\varrho-l)^{2}\alpha^{2}}{2\varepsilon_{1}}+\frac{4\lambda_{0}g^{2}(0)\alpha^{2}}{G(t)}\right)N_{2}\right]\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2} \nonumber\\ &-\left(\frac{G(t)}{4}N_{2}-1-\frac{\varepsilon_{2}\delta^{3}N_{3}}{8\alpha^{2}\varrho^{2}\lambda_{0}}\right)\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2}+\frac{\varepsilon\varrho\nu}{2\delta}N\left(g\circ\mathcal{A}^{\frac{1}{2}}u\right)(t) \nonumber\\ &-\left[\frac{\delta}{2\alpha}-\left(\frac{\lambda_{0}(\varrho-l)^{2}}{2} +\left(\frac{\delta^{2}}{\alpha^{2}}+\varrho^{2}\right) +3(\varrho-l)^{2}\right)\varepsilon_{1}N_{2}\right]\nonumber\\ &\times\left\|\mathcal{A}^{\frac{1}{2}}u_{tt} +\alpha\mathcal{A}^{\frac{1}{2}}u_{t}+\frac{\alpha\varrho}{\delta}\mathcal{A}^{\frac{1}{2}}u\right\|^{2} \nonumber\\ &-\left[-\frac{\alpha g'(t)}{2}N-\frac{2\alpha(\varrho-l)^{2}}{\delta} -\left(\frac{\alpha^{2}\varrho^{2}(\varrho-l)^{2}\lambda_{0}\varepsilon_{1}}{2\delta^{2}} +\frac{(\varrho-l)^{2}}{2\varepsilon_{1}}\right.\right. \nonumber\\ &\left.\left.+3(\varrho-l)^{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\varepsilon_{1}+\frac{(\varrho-l)^{2}}{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\right)N_{2} -\frac{2(\varrho-l)^{2}}{\delta}N_{3}\right]\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2} \nonumber\\ &-\left[\frac{\varepsilon\varrho}{2\delta}N-\frac{2\alpha}{\delta}C_{\nu} -\left(\left(\frac{\varepsilon_{1}\alpha^{2}\varrho^{2}\lambda_{0}}{4\delta^{2}} \right.\right.\right.\left.+\frac{3}{4\varepsilon_{1}}+1+\frac{\alpha\varrho}{\delta} +\frac{2\alpha^{4}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}\right)C_{\nu} \nonumber\\ &\left.\left. +\frac{2\alpha^{2}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}\right)N_{2} -\frac{N_{3}}{\delta}C_{\nu}\right]\left(h\circ\mathcal{A}^{\frac{1}{2}}u\right)(t). \end{align*}

    At this point, we need to choose our constants very carefully. First, we choose

    \begin{align*} \varepsilon_{1} = \frac{\alpha\delta}{2N_{2}\left[\lambda_{0}\alpha^{2}(\varrho-l)^{2}+2(\delta^{2}+\alpha^{2}\varrho^{2})+6\alpha^{2}(\varrho-l)^{2}\right]}\quad\quad and \quad\quad \varepsilon_{2} = \frac{1}{N_{3}}. \end{align*}

    The above choice yields

    \begin{align*} &\mathcal{L'}(t)\nonumber\\ &\leq-\left[\alpha\left(\beta-\frac{\alpha\varrho}{\delta}\right)N-\frac{2\alpha\lambda_{0}}{\delta}\left(\beta-\frac{\alpha\varrho}{\delta}\right)^{2}-\left(\frac{2\left(\beta-\frac{\alpha\varrho}{\delta}\right)^{2}}{\varepsilon_{1}}+\frac{4g^{2}(0)}{G(t)}\right)N_{2}\right. \nonumber\\ &\left.-\frac{\delta^{3}}{4\varrho^{2}\lambda_{0}}-\frac{4\alpha^{2}\varrho^{2}\lambda_{0}}{\delta^{3}}N_{3}^{2}-\frac{2\beta^{2}\lambda_{0}}{\delta}N_{3}\right]\|u_{tt}\|^{2}-\left[\left(\alpha\left(\gamma-\frac{\delta}{\alpha}\right)-\frac{\delta g(0)}{2(\alpha-\varepsilon)\varrho}\right)N\right. \nonumber\\ &\left.+\left(\frac{\delta g(t)}{2(\alpha-\varepsilon)\varrho}-\frac{\delta g(t)}{2\alpha\varrho}\right)N-\frac{2\alpha}{\delta}\left(\gamma-\frac{\delta}{\alpha}\right)^{2}-\left(\frac{(\varrho-l)^{2}}{2\varepsilon_{1}}+\frac{1}{2\varepsilon_{1}}\left(\gamma-\frac{\delta}{\alpha}\right)^{2}\right.\right. \nonumber\\ &\left.\left.+\frac{1}{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\left(\gamma-\frac{\delta}{\alpha}\right)^{2}\right)N_{2}-\frac{2\gamma^{2}}{\delta}N_{3}\right]\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}\right\|^{2}-\frac{\alpha}{2}N\left(g''\circ\mathcal{A}^{\frac{1}{2}}u\right)(t) \nonumber\\ &-\left[\frac{3\delta}{8}N_{3}-\frac{\delta}{4}-\left(\frac{(\varrho-l)^{2}\alpha^{2}}{2\varepsilon_{1}}+\frac{4\lambda_{0}g^{2}(0) \alpha^{2}}{G(t)}\right)N_{2}\right]\left\|\mathcal{A}^{\frac{1}{2}}u_{t}\right\|^{2} \nonumber\\ &+\frac{\varepsilon\varrho\nu}{2\delta}N\left(g\circ\mathcal{A}^{\frac{1}{2}}u\right)(t)-\left(\frac{G(t)}{4}N_{2}-1-\frac{\delta^{3}}{8\alpha^{2}\varrho^{2}\lambda_{0}}\right)\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2}\nonumber\\ & -\frac{\delta}{4\alpha}\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}+\alpha\mathcal{A}^{\frac{1}{2}}u_{t}+\frac{\alpha\varrho}{\delta}\mathcal{A}^{\frac{1}{2}}u\right\|^{2} \nonumber\\ &-\left[-\frac{\alpha g'(t)}{2}N-\frac{2\alpha(\varrho-l)^{2}}{\delta}-\left(\frac{\alpha^{2}\varrho^{2}(\varrho-l)^{2} \lambda_{0}\varepsilon_{1}}{2\delta^{2}}+\frac{(\varrho-l)^{2}}{2\varepsilon_{1}}+3(\varrho-l)^{2} \left(\frac{\alpha\varrho}{\delta}\right)^{2}\varepsilon_{1}\right.\right. \nonumber\\ &\left.\left.+\frac{(\varrho-l)^{2}}{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\right)N_{2} -\frac{2(\varrho-l)^{2}}{\delta}N_{3}\right]\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2} -\left[\frac{\varepsilon\varrho}{2\delta}N-\frac{2\alpha}{\delta}C_{\nu} \right. \nonumber\\ &-\left(\left(\frac{\varepsilon_{1}\alpha^{2}\varrho^{2}\lambda_{0}}{4\delta^{2}}\right.\right.\left. \left.\left.+\frac{3}{4\varepsilon_{1}}+1+\frac{\alpha\varrho}{\delta}+\frac{2\alpha^{4}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}\right) C_{\nu}+\frac{2\alpha^{2}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}\right)N_{2}\right. \nonumber\\ &\left.-\frac{N_{3}}{\delta}C_{\nu}\right]\left(h\circ\mathcal{A}^{\frac{1}{2}}u\right)(t). \end{align*}

    Then, we choose N_{2} large enough so that

    \begin{align*} \frac{G(t)}{4}N_{2}-1-\frac{\delta^{3}}{8\alpha^{2}\varrho^{2}\lambda_{0}} > 0. \end{align*}

    Next, we choose N_{3} large enough so that

    \begin{align*} \frac{3\delta}{8}N_{3}-\frac{\delta}{4}-\left(\frac{(\varrho-l)^{2}\alpha^{2}}{2\varepsilon_{1}}+\frac{4\lambda_{0}g^{2}(0)\alpha^{2}}{G(t)}\right)N_{2} > 0. \end{align*}

    Now, as \frac{\nu^{2}g(s)}{\nu g(s)-g'(s)}<g(s) , it is easy to show, using the Lebesgue dominated convergence theorem, that

    \begin{align*} \nu C_{\nu} = \int_{0}^{\infty}\frac{\nu^{2}g(s)}{\nu g(s)-g'(s)}ds\rightarrow 0, \quad\quad as \quad\quad \nu\rightarrow 0. \end{align*}

    Hence, there is 0<\nu_{0}<1 such that if \nu<\nu_{0} , then

    \begin{align*} \nu C_{\nu} < \frac{1}{16\left(\frac{2\alpha}{\delta}+\left(\frac{\varepsilon_{1}\alpha^{2}\varrho^{2}\lambda_{0}}{4\delta^{2}}+\frac{3}{4\varepsilon_{1}}+1+\frac{\alpha\varrho}{\delta}+\frac{2\alpha^{4}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}\right)N_{2}+\frac{N_{3}}{\delta}\right)}. \end{align*}

    Now, let us choose N large enough and choose \nu satisfying

    \begin{align*} \frac{\varepsilon\varrho}{4\delta}N-\frac{2\alpha^{2}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}N_{2} > 0 \quad\quad and \quad\quad \nu = \frac{\delta}{4\varepsilon\varrho N} < \nu_{0}, \end{align*}

    which means

    \begin{align*} &\frac{\varepsilon\varrho}{2\delta}N-\frac{2\alpha^{2}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}N_{2}\nonumber\\ &-C_{\nu}\left(\frac{2\alpha}{\delta}+\left(\frac{\varepsilon_{1}\alpha^{2}\varrho^{2}\lambda_{0}}{4\delta^{2}}+\frac{3}{4\varepsilon_{1}} +1+\frac{\alpha\varrho}{\delta}+\frac{2\alpha^{4}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}\right)N_{2}+\frac{N_{3}}{\delta}\right) \nonumber\\ > &\frac{\varepsilon\varrho}{2\delta}N-\frac{2\alpha^{2}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}N_{2}-\frac{1}{16\nu} = \frac{\varepsilon\varrho}{4\delta}N-\frac{2\alpha^{2}\varrho^{2}\lambda_{0}}{G(t)\delta^{2}}N_{2} > 0 \end{align*}

    and

    \begin{align*} &\alpha\left(\beta-\frac{\alpha\varrho}{\delta}\right)N-\frac{2\alpha\lambda_{0}}{\delta}\left(\beta-\frac{\alpha\varrho}{\delta}\right)^{2}-\left(\frac{2\left(\beta-\frac{\alpha\varrho}{\delta}\right)^{2}}{\varepsilon_{1}}+\frac{4g^{2}(0)}{G(t)}\right)N_{2} \nonumber\\ &-\frac{\delta^{3}}{4\varrho^{2}\lambda_{0}}-\frac{4\alpha^{2}\varrho^{2}\lambda_{0}}{\delta^{3}}N_{3}^{2}-\frac{2\beta^{2}\lambda_{0}}{\delta}N_{3} > 0, \nonumber\\ &\left(\alpha\left(\gamma-\frac{\delta}{\alpha}\right)-\frac{\delta g(0)}{2(\alpha-\varepsilon)\varrho}\right)N +\left(\frac{\delta g(t)}{2(\alpha-\varepsilon)\varrho}-\frac{\delta g(t)}{2\alpha\varrho}\right)N-\frac{2\alpha}{\delta}\left(\gamma-\frac{\delta}{\alpha}\right)^{2} \nonumber\\ &-\left(\frac{(\varrho-l)^{2}}{2\varepsilon_{1}}+\frac{1}{2\varepsilon_{1}}\left(\gamma-\frac{\delta}{\alpha}\right)^{2} +\frac{1}{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\left(\gamma-\frac{\delta}{\alpha}\right)^{2}\right)N_{2}-\frac{2\gamma^{2}}{\delta}N_{3} > 0, \nonumber\\ &-\frac{\alpha g'(t)}{2}N-\frac{2\alpha(\varrho-l)^{2}}{\delta}-\left(\frac{\alpha^{2}\varrho^{2}(\varrho-l)^{2}\lambda_{0}\varepsilon_{1}}{2\delta^{2}}+\frac{(\varrho-l)^{2}}{2\varepsilon_{1}}+3(\varrho-l)^{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\varepsilon_{1}\right. \nonumber\\ &\left.+\frac{(\varrho-l)^{2}}{2}\left(\frac{\alpha\varrho}{\delta}\right)^{2}\right)N_{2}-\frac{2(\varrho-l)^{2}}{\delta}N_{3} > 4(\varrho-l). \end{align*}

    So we arrive at, for positive constant c ,

    \begin{align*} \mathcal{L'}(t)\leq&-c\left[\|u_{tt}\|^{2}+\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}\right\|^{2}+\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2} \right. \nonumber\\ &\left.+\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}+\alpha\mathcal{A}^{\frac{1}{2}}u_{t}+\frac{\alpha\varrho}{\delta}u\right\|^{2}\right]-4(\varrho-l)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}+\frac{1}{8}\left(g\circ\mathcal{A}^{\frac{1}{2}}u\right)(t). \end{align*}

    On the other hand, from Lemma 3.2, we find that

    \begin{align*} &\left|\mathcal{L}(t)-NE(t)\right| \nonumber\\ \leq&\int_{\Omega}\left|u_{tt}+\alpha u_{t}+\frac{\alpha\varrho}{\delta}u\right|\left|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right|dx \nonumber\\ &+N_{2}\int_{\Omega}\left|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right|\int_{0}^{t}g(t-s) \nonumber\\ & \times\left|\left(u_{tt}+\alpha u_{t}+\frac{\alpha\varrho}{\delta}u\right)(t)-\frac{\alpha\varrho}{\delta} u(s)\right|dsdx \nonumber\\ &+N_{3}\int_{\Omega}\left|u_{ttt}+\alpha u_{tt}\right||u_{t}|dx+N_{3}\frac{\varrho}{2}\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2} \nonumber\\ \leq&c E(t). \end{align*}

    Therefore, we can choose N even large (if needed) so that (4.8) is satisfied.

    In this section, we will give an estimate to the decay rate for the problem (1.7)-(1.8).

    Proof of Theorem 2.2. Our proof starts with the observation that, for any t\geq t_{0} ,

    \begin{align*} &\int_{0}^{t_{0}}g(s)\int_{\Omega}\left|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right|^{2}dxds\nonumber\\ \leq&-\frac{g(0)}{a}\int_{0}^{t_{0}}g'(s)\int_{\Omega}\left|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right|^{2}dxds \nonumber\\ \leq&-\frac{g(0)}{a}\int_{0}^{t}g'(s)\int_{\Omega}\left|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right|^{2}dxds \nonumber\\ \leq&-c E'(t), \end{align*}

    which are derived from (2.2) and Lemma 4.1 and can be used in (4.8).

    Taking \mathcal{F}(t) = \mathcal{L}(t)+cE(t) , which is obviously equivalent to E(t) , we get, for all t\geq t_{0} ,

    \begin{align*} &\mathcal{L'}(t)\nonumber\\ \leq&-c\left[\|u_{tt}\|^{2}+\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}\right\|^{2}+\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2}+\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}+\alpha\mathcal{A}^{\frac{1}{2}}u_{t}+\frac{\alpha\varrho}{\delta}u\right\|^{2}\right] \nonumber\\ &-4(\varrho-l)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}+\frac{1}{8}\left(g\circ\mathcal{A}^{\frac{1}{2}}u\right)(t) \nonumber\\ \leq&-mE(t)+c\left(g\circ\mathcal{A}^{\frac{1}{2}}u\right)(t) \nonumber\\ \leq&-mE(t)-cE'(t)+c\int_{t_{0}}^{t}g(s)\int_{\Omega}\left|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right|^{2}dxds, \end{align*}

    where m is a positive constant. Then, we obtain that

    \begin{align} \mathcal{F'}(t)& = \mathcal{L'}(t)+cE'(t)\\ & \leq-mE(t)+c\int_{t_{0}}^{t}g(s)\int_{\Omega}\left|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right|^{2}dxds. \end{align} (5.1)

    We consider the following two cases relying on the ideas presented in [31].

    (ⅰ) M(t) is linear.

    We multiply (5.1) by \xi(t) , then on account of (A1) - (A4) and Lemma 4.1, we obtain, for all t\geq t_{0} ,

    \begin{align*} \xi(t)\mathcal{F'}(t)\leq&-m\xi(t)E(t)+c\xi(t)\int_{t_{0}}^{t}g(s)\int_{\Omega}\left|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right|^{2}dxds \nonumber\\ \leq&-m\xi(t)E(t)+c\int_{t_{0}}^{t}\xi(s)g(s)\int_{\Omega}\left|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right|^{2}dxds \nonumber\\ \leq&-m\xi(t)E(t)-c\int_{t_{0}}^{t}g'(s)\int_{\Omega}\left|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right|^{2}dxds \nonumber\\ \leq&-m\xi(t)E(t)-c\int_{0}^{t}g'(s)\int_{\Omega}\left|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right|^{2}dxds \nonumber\\ \leq&-m\xi(t)E(t)-cE'(t). \end{align*}

    Therefore,

    \begin{align*} \xi(t)\mathcal{F'}(t)+cE'(t)\leq-m\xi(t)E(t). \end{align*}

    As \xi(t) is non-increasing and \mathcal{F}(t)\sim E(t) , we have

    \begin{align*} \xi(t)\mathcal{F}(t)+cE(t)\sim E(t) \end{align*}

    and

    \begin{align*} (\xi\mathcal{F}+cE)'(t)\leq -m\xi(t)E(t), \quad\quad \forall\ t\geq t_{0}. \end{align*}

    It follows immediately that

    \begin{align*} E'(t)\leq -m\xi(t)E(t), \quad\quad \forall\ t\geq t_{0}. \end{align*}

    We may now integrate over (t_{0},t) to conclude that, for two positive constants k_{1} and k_{2}

    \begin{align*} E(t)\leq k_{2}\exp\left(-k_{1}\int_{t_{0}}^{t}\xi(s)ds\right), \quad\quad \forall\ t\geq t_{0}. \end{align*}

    By the continuity of E(t) , we have

    \begin{align*} E(t)\leq k_{2}\exp\left(-k_{1}\int_{0}^{t}\xi(s)ds\right), \quad\quad \forall\ t > 0. \end{align*}

    (ⅱ) M is nonlinear.

    First, we define the functional

    \begin{align*} L(t) = \mathcal{L}(t)+F_{4}(t). \end{align*}

    Obviously, L(t) is nonnegative. And, by Lemma 4.5 and Lemma 4.6, there exists b>0 such that

    \begin{align*} L'(t) = &\mathcal{L}'(t)+F_{4}'(t) \nonumber\\ \leq&-c\left[\|u_{tt}\|^{2}+\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}\right\|^{2}+\left\|u_{ttt}+\alpha u_{tt}+\frac{\alpha\varrho}{\delta}u_{t}\right\|^{2} \right. \nonumber\\ &\left.+\left\|\mathcal{A}^{\frac{1}{2}}u_{tt}+\alpha\mathcal{A}^{\frac{1}{2}}u_{t}+\frac{\alpha\varrho}{\delta}u\right\|^{2}\right]-(\varrho-l)\left\|\mathcal{A}^{\frac{1}{2}}u\right\|^{2}-\frac{3}{8}\left(g\circ\mathcal{A}^{\frac{1}{2}}u\right)(t) \nonumber\\ \leq& -b E(t). \end{align*}

    Therefore, integrating the above inequality over (t_{0},t) , we see at once that

    \begin{align*} -L(t_{0})\leq L(t)-L(t_{0})\leq-b\int_{t_{0}}^{t}E(s){\rm d}s. \end{align*}

    It is sufficient to show that

    \begin{align} \int_{0}^{\infty}E(s){\rm d}s < \infty \end{align} (5.2)

    and

    \begin{align*} E(t)\leq \frac{c}{t-t_{0}}, \quad\quad\quad \forall t > t_{0}. \end{align*}

    Now, we define a functional \lambda(t) by

    \begin{align*} \lambda(t): = -\int_{t_{0}}^{t}g'(s)\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s. \end{align*}

    Clearly, we have

    \begin{align} \lambda(t)\leq&-\int_{0}^{t}g'(s)\left\| \mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}ds \\ \leq&-cE'(t), \quad\quad\quad \forall\ t\geq t_{0}. \end{align} (5.3)

    After that, we define another functional I(t) by

    \begin{align*} I(t): = q\int_{t_{0}}^{t}\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s. \end{align*}

    Now, the following inequality holds under Lemma 4.1 and (5.2) that

    \begin{align} \int_{t_{0}}^{t}\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s \leq & 2\int_{t_{0}}^{t}\left(\left\|\mathcal{A}^{\frac{1}{2}}u(t)\right\|^{2}+\left\|\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}\right)ds \\ \leq& 4\int_{t_{0}}^{t}\left(E(t)+E(t-s)\right)ds \\ \leq& 8\int_{t_{0}}^{t}E(0)ds \\ < & \infty. \end{align} (5.4)

    Then (5.4) allows for a constant 0<q<1 chosen so that, for all t\geq t_{0} ,

    \begin{align} 0 < I(t) < 1; \end{align} (5.5)

    otherwise we get an exponential decay from (5.1). Moreover, recalling that M is strict convex on (0,r] and M(0) = 0 , then

    \begin{align*} M(\theta x)\leq \theta M(x), \quad\quad\quad for \quad\quad 0\leq \theta\leq 1 \quad\quad and \quad\quad x\in(0,r]. \end{align*}

    From assumptions (A2) and (A4) , (5.5) and Lemma 2.2, it follows that

    \begin{align*} \lambda(t) = &-\int_{t_{0}}^{t}g'(s)\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s \nonumber\\ = &\frac{1}{qI(t)}\int_{t_{0}}^{t}I(t)(-g'(s))q\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s \nonumber\\ \geq& \frac{1}{qI(t)}\int_{t_{0}}^{t}I(t)\xi(s)M(g(s))q\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s \nonumber\\ \geq&\frac{\xi(t)}{qI(t)}\int_{t_{0}}^{t}M(I(t)g(s))q\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s \nonumber\\ \geq&\frac{\xi(t)}{q}M\left(\frac{1}{I(t)}\int_{t_{0}}^{t}I(t)g(s)q\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s\right) \nonumber\\ = &\frac{\xi(t)}{q}M\left(q\int_{t_{0}}^{t}g(s)\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s\right). \end{align*}

    According to \overline{M} is an extension of M (see Remark 2.1 (2) ), we also have

    \begin{align*} \lambda(t)\geq&\frac{\xi(t)}{q}\overline{M}\left(q\int_{t_{0}}^{t}g(s)\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s\right). \end{align*}

    In this way,

    \begin{align*} \int^{t}_{t_{0}}g(s)\left\|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right\|^{2}{\rm d}s\leq\frac{1}{q}\overline{M}^{-1}\left(\frac{q\lambda(t)}{\xi(t)}\right) \end{align*}

    and (5.1) becomes

    \begin{align} \mathcal{F'}(t) \leq&-mE(t)+c\int_{t_{0}}^{t}g(s)\int_{\Omega}\left|\mathcal{A}^{\frac{1}{2}}u(t)-\mathcal{A}^{\frac{1}{2}}u(t-s)\right|^{2}dxds \\ \leq&-mE(t)+c\overline{M}^{-1}\left(\frac{q\lambda(t)}{\xi(t)}\right), \quad\quad\quad \forall\ t\geq t_{0}. \end{align} (5.6)

    Let 0<\varepsilon_{0}<r , we define the functional \mathcal{F}_{1}(t) by

    \begin{align*} \mathcal{F}_{1}(t): = \overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)\mathcal{F}(t)+E(t), \quad\quad\quad \forall\ t\geq 0. \end{align*}

    Then, recalling that E'(t)\leq 0 , \overline{M}'>0 and \overline{M}''>0 as well as making use of estimate (5.6), we deduce that \mathcal{F}_{1}(t)\sim E(t) and also, for any t\geq t_{0} , we have

    \begin{align} \mathcal{F}'_{1}(t) \leq& -mE(t)\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)+c\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)\overline{M}^{-1}\left(\frac{q\lambda(t)}{\xi(t)}\right)+E'(t). \end{align} (5.7)

    Taking account of Lemma 2.3, we obtain

    \begin{align} &\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)\overline{M}^{-1}\left(\frac{q\lambda(t)}{\xi(t)}\right) \\ \leq&\overline{M}^{*}\left(\overline{H}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)\right)+\overline{M}\left(\overline{M}^{-1}\left(\frac{q\lambda(t)}{\xi(t)}\right)\right) \\ = &\overline{M}^{*}\left(\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)\right)+\frac{q\lambda(t)}{\xi(t)} \end{align} (5.8)

    where

    \begin{align} &\overline{M}^{*}\left(\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)\right) \\ = &\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)\left(\overline{M}'\right)^{-1}\left(\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)\right)-\overline{M}\left[\left(\overline{M}'\right)^{-1}\left(\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)\right)\right] \\ = &\varepsilon_{0}\frac{E(t)}{E(0)}\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)-\overline{M}\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right) \\ \leq&\varepsilon_{0}\frac{E(t)}{E(0)}\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right). \end{align} (5.9)

    So, combining (5.7), (5.8) and (5.9), we obtain

    \begin{align*} \mathcal{F}'_{1}(t)\leq-(mE(0)-c\varepsilon_{0})\frac{E(t)}{E(0)}\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)+c\frac{q\lambda(t)}{\xi(t)}+E'(t). \end{align*}

    From this, we multiply the above inequality by \xi(t) to get

    \begin{align*} \xi(t)\mathcal{F}'_{1}(t)\leq-(mE(0)-c\varepsilon_{0})\xi(t)\frac{E(t)}{E(0)}\overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)+cq\lambda(t)+\xi(t)E'(t). \end{align*}

    Then, using the fact that, as \varepsilon_{0}\frac{E(t)}{E(0)}<r , \overline{M}'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right) = M'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right) and (5.3), we get

    \begin{align*} \xi(t)\mathcal{F}'_{1}(t) \leq&-(mE(0)-c\varepsilon_{0})\xi(t)\frac{E(t)}{E(0)}M'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right)-cE'(t). \end{align*}

    Consequently, defining \mathcal{F}_{2}(t) = \xi(t)\mathcal{F}_{1}(t)+cE(t) , then since \mathcal{F}_{1}(t)\sim E(t) , we arrive at

    \begin{align} \mathcal{F}_{2}(t)\sim E(t), \end{align} (5.10)

    and with a suitable choice of \varepsilon_{0} , we get, for some positive constant k and for any t\geq t_{0} ,

    \begin{align} \mathcal{F}'_{2}(t)\leq-k\xi(t)\left(\frac{E(t)}{E(0)}\right)M'\left(\varepsilon_{0}\frac{E(t)}{E(0)}\right). \end{align} (5.11)

    Define

    \begin{align*} R(t) = \frac{\lambda_{1}\mathcal{F}_{2}(t)}{E(0)}, \quad\quad\lambda_{1} > 0\quad\quad\quad and \quad\quad\quad M_{2}(t) = tM'(\varepsilon_{0} t). \end{align*}

    Moreover, it suffices to show that M'_{2}(t), M_{2}(t)>0 on (0,1] by the strict convexity of M on (0,r] . And, it is easily seen that

    \begin{align} \mathcal{F}'_{2}(t) \leq -k\xi(t)M_{2}\left(\frac{E(t)}{E(0)}\right). \end{align} (5.12)

    According to (5.10) and (5.12), there exist \lambda_{2}, \lambda_{3}>0 such that

    \begin{align} \lambda_{2}R(t)\leq E(t)\leq\lambda_{3}R(t). \end{align} (5.13)

    Then, it follows that there exists k_{1}>0 such that

    \begin{align} k_{1}\xi(t)\leq-\frac{R'(t)}{M_{2}(R(t))}, \quad\quad\quad \forall\ t\geq t_{0}. \end{align} (5.14)

    Next, we define

    \begin{align*} M_{1}(t): = \int_{t}^{r}\frac{1}{sM'(s)}{\rm d}s. \end{align*}

    And based on the properties of M , we know that M_{1} is strictly decreasing function on (0,r] and {\lim_{t\rightarrow 0}}M_{1}(t) = +\infty .

    Now, we integrate (5.14) over (t_{0},t) to obtain

    \begin{align*} -\int_{t_{0}}^{t}\frac{R'(s)}{M_{2}(R(s))}{\rm d}s\geq k_{1}\int_{t_{0}}^{t}\xi(s){\rm d}s \end{align*}

    so

    \begin{align*} k_{1}\int_{t_{0}}^{t}\xi(s){\rm d}s\leq M_{1}\left(\varepsilon_{0}R(t)\right)-M_{1}\left(\varepsilon_{0}R(t_{0})\right), \end{align*}

    which implies that

    \begin{align*} M_{1}\left(\varepsilon_{0}R(t)\right)\geq k_{1}\int_{t_{0}}^{t}\xi(s){\rm d}s. \end{align*}

    It is easy to obtain that

    \begin{align} R(t)\leq\frac{1}{\varepsilon_{0}}M^{-1}_{1}\left(k_{1}\int_{t_{0}}^{t}\xi(s){\rm d}s\right), \quad\quad\quad \forall\ t\geq t_{0}. \end{align} (5.15)

    A combining of (5.13) and (5.15) gives the proof.

    The authors are grateful to the anonymous referees and the editor for their useful remarks and comments.

    [1] Ashmore MR, Dimitroulopoulou C (2009) Personal exposure of children to air pollution. Atmos Environ 43: 128-141. doi: 10.1016/j.atmosenv.2008.09.024
    [2] Son B, Breysse P, Yang W (2003) Volatile organic compounds concentrations in residential indoor and outdoor and its personal exposure in Korea. Environ Int 29: 79-85. doi: 10.1016/S0160-4120(02)00148-4
    [3] Park KH, Jo WK (2004) Personal volatile organic compound (VOC) exposure of children attending elementary schools adjacent to industrial complex. Atmos Environ 38: 1303-1312. doi: 10.1016/j.atmosenv.2003.11.032
    [4] Salvi S (2007) Health effects of ambient air pollution in children. Paediatr Respir Rev 8: 275-280. doi: 10.1016/j.prrv.2007.08.008
    [5] Daisey JM, Angell WJ, Apte MG, et al. (2003) Indoor air quality, ventilation and health symptoms in schools: an analysis of existing information. Indoor Air 13: 53-64. doi: 10.1034/j.1600-0668.2003.00153.x
    [6] Sofuoglu SC, Aslan G, Inal F, et al. (2011) An assessment of indoor air concentrations and health risks of volatile organic compounds in three primary schools. Int J Hyg Environ Health 214: 36-46. doi: 10.1016/j.ijheh.2010.08.008
    [7] de Gennaro G, Farella G, Marzocca A, et al. (2013) Indoor and outdoor monitoring of volatile organic compounds in school buildings: Indicators based on health risk assessment to single out critical issues. Int J Environ Res Public Health 10: 6273-6291. doi: 10.3390/ijerph10126273
    [8] Godwin C, Batterman S (2007) Indoor air quality in Michigan schools. Indoor Air 17: 109-121. doi: 10.1111/j.1600-0668.2006.00459.x
    [9] Rizk M, Verriele M, Dusanter S, et al. (2016) Fast sorption measurements of volatile organic compounds on building materials: Part 1 - Methodology developed for field applications. Data Br 6: 953-958. doi: 10.1016/j.dib.2016.01.011
    [10] ASHRAE (1999) Standard 62-1999 Ventilation for Acceptable Indoor Air Quality.
    [11] IARC (2015) International Agency for Research on Cancer: Monographs on the Evaluation of Carcinogenic Risks to Humans. Available from: http: //monographs.iarc.fr/ENG/Classification/ index.php
    [12] Mendell MJ, Heath GA (2005) Do indoor pollutants and thermal conditions in schools influence student performance? A critical review of the literature. Indoor Air 15: 27-52.
    [13] Santamouris M, Synnefa A, Asssimakopoulos M, et al. (2008) Experimental investigation of the air flow and indoor carbon dioxide concentration in classrooms with intermittent natural ventilation. Energy Build 40: 1833-1843. doi: 10.1016/j.enbuild.2008.04.002
    [14] Yoon C, Lee K, Park D (2011) Indoor air quality differences between urban and rural preschools in Korea. Environ Sci Pollut Res 18: 333-345. doi: 10.1007/s11356-010-0377-0
    [15] Branco PTBS, Alvim-Ferraz MCM, Martins FG, et al. (2014) Indoor air quality in urban nurseries at Porto city: Particulate matter assessment. Atmos Environ 84: 133-143. doi: 10.1016/j.atmosenv.2013.11.035
    [16] Mainka A, Brągoszewska E, Kozielska B, et al. (2015) Indoor air quality in urban nursery schools in Gliwice, Poland: Analysis of the case study. Atmos Pollut Res 6: 1098-1104. doi: 10.1016/j.apr.2015.06.007
    [17] Mainka A, Zajusz-Zubek E, Kozielska B, et al. (2015) Badanie zanieczyszczeń powietrza oddziałujących na dzieci w przedszkolu miejskim zlokalizowanym przy drodze o dużym natężeniu ruchu. Inżynieria i Ochr Środowiska 18: 119-133.
    [18] Mainka A, Zajusz-Zubek E, Kozielska B, et al. (2014) Badanie zanieczyszczeń powietrza oddziałujących na dzieci w wybranym przedszkolu miejskim. In: Konieczyński J (ed) Ochr powietrza w Teor i Prakt Instytut Podstaw Inżynierii Środowiska Polskiej Akademii Nauk, Zabrze, 1: 115-128.
    [19] Kozielska B (2013) Concentration of benzene and its alkyl derivatives in Gliwice air. Arch Environ Prot Waste Manag Environ Prot 15: 81-88.
    [20] Brągoszewska E, Mainka A, Pastuszka JS (2016) Bacterial aerosols in an urban nursery school in Gliwice, Poland: a case study. Aerobiologia 32: 469-480. doi: 10.1007/s10453-015-9419-x
    [21] SKC (2012) VOC method update SKC appendices to EPA method TO-17 determination of volatile organic compounds (VOCs) in ambient or indoor air.
    [22] Dz.U. (2012) (Journal of Laws) Regulation of the Minister of Environment of 13 September 2012, No. 1031 - The levels of certain substances in the ambient air.
    [23] Demirel G, Özden Ö, Döǧeroǧlu T, et al. (2014) Personal exposure of primary school children to BTEX, NO2 and ozone in Eskişehir, Turkey: Relationship with indoor/outdoor concentrations and risk assessment. Sci Total Environ 473-474: 537-548. doi: 10.1016/j.scitotenv.2013.12.034
    [24] Marć M, Zabiegała B, Simeonov V, et al. (2014) The Relationships Between BTEX, NOx, and O3 Concentrations in Urban Air in Gdansk and Gdynia, Poland. CLEAN Soil Air Water 42: 1326-1336. doi: 10.1002/clen.201300188
    [25] Rogula-Kozłowska W, Pastuszka JS, Talik E (2008) Influence of vehicular traffic on concentration and particle surface composition of PM10 and PM2.5 in Zabrze, Poland. Polish J Environ Stud 17: 539-548.
    [26] Jung KH, Artigas F, Shin JY (2011) Personal, indoor, and outdoor exposure to VOCs in the immediate vicinity of a local airport. Environ Monit Assess 173: 555-567. doi: 10.1007/s10661-010-1404-9
    [27] Buczyńska AJ, Krata A, Stranger M, et al. (2009) Atmospheric BTEX-concentrations in an area with intensive street traffic. Atmos Environ 43: 311-318. doi: 10.1016/j.atmosenv.2008.09.071
    [28] Delgado-Saborit JM, Aquilina NJ, Meddings C, et al. (2011) Relationship of personal exposure to volatile organic compounds to home, work and fixed site outdoor concentrations. Sci Total Environ 409: 478-488. doi: 10.1016/j.scitotenv.2010.10.014
    [29] EPA U (1999) Environmental Technology Protocol Verification Report Emissions of VOCs and Aldehydes from Commercial Furniture.
    [30] Kotzias D (2005) Indoor air and human exposure assessment - Needs and approaches. Exp Toxicol Pathol 57: 5-7. doi: 10.1016/j.etp.2005.05.002
    [31] ATSDR (2015) Agency for Toxic Substances and Diseases Registry. Available from: http://www.atsdr.cdc.gov/toxfaq.html.
    [32] Guo H, Lee SC, Li WM, et al. (2003) Source characterization of BTEX in indoor microenvironments in Hong Kong. Atmos Environ 37: 73-82.
    [33] Hoque RR, Khillare PS, Agarwal T, et al. (2008) Spatial and temporal variation of BTEX in the urban atmosphere of Delhi, India. Sci Total Environ 392: 30-40. doi: 10.1016/j.scitotenv.2007.08.036
    [34] ASHRAE (2009) ASHRAE Handbook - Fundamentals (SI Edition). American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
    [35] WHO (2010) WHO guidelines for indoor air quality: selected pollutants.
    [36] Kim KH, Jahan SA, Kabir E (2011) A review of diseases associated with household air pollution due to the use of biomass fuels. J Hazard Mater 192: 425-431. doi: 10.1016/j.jhazmat.2011.05.087
    [37] Wu XM, Fan ZT, Zhu X, et al. (2012) Exposures to volatile organic compounds (VOCs) and associated health risks of socio-economically disadvantaged population in a “hot spot” in Camden, New Jersey. Atmos Environ 57: 72-79. doi: 10.1016/j.atmosenv.2012.04.029
    [38] Hinwood AL, Rodriguez C, Runnion T, et al. (2007) Risk factors for increased BTEX exposure in four Australian cities. Chemosphere 66: 533-541. doi: 10.1016/j.chemosphere.2006.05.040
    [39] U.S. EPA (2002) Child-Specific Exposure Factors Handbook. EPA/600/R-06/096F.
    [40] Johnson-Restrepo B, Kannan K (2009) An assessment of sources and pathways of human exposure to polybrominated diphenyl ethers in the United States. Chemosphere 76: 542-548.
    [41] U.S. EPA (2004) Risk Assessment: “Supplemental Guidance for Dermal Risk Assessment”, Part E of Risk Assessment Guidance for Superfund, Human Health Evaluation Manual (Volume I), August 16, 2004.
    [42] Ott WR, Steinemann AC, Wallace LA (2006) Exposure Analysis, CRC Press, London.
    [43] Guo H, Lee SC, Chan LY, et al. (2004) Risk assessment of exposure to volatile organic compounds in different indoor environments. Environ Res 94: 57-66. doi: 10.1016/S0013-9351(03)00035-5
    [44] Li H, Qian X, Hu W, et al. (2013) Chemical speciation and human health risk of trace metals in urban street dusts from a metropolitan city, Nanjing, SE China. Sci Total Environ 456-457: 212-221. doi: 10.1016/j.scitotenv.2013.03.094
  • This article has been cited by:

    1. Zhiyu Tu, Wenjun Liu, Well‐posedness and exponential decay for the Moore–Gibson–Thompson equation with time‐dependent memory kernel, 2023, 0170-4214, 10.1002/mma.9133
    2. Marina Murillo‐Arcila, Well‐posedness for the fourth‐order Moore–Gibson–Thompson equation in the class of Banach‐space‐valued Hölder‐continuous functions, 2023, 46, 0170-4214, 1928, 10.1002/mma.8618
    3. Carlos Lizama, Marina Murillo-Arcila, Well-posedness for a fourth-order equation of Moore–Gibson–Thompson type, 2021, 14173875, 1, 10.14232/ejqtde.2021.1.81
    4. Danhua Wang, Wenjun Liu, Well-posedness and decay property of regularity-loss type for the Cauchy problem of the standard linear solid model with Gurtin–Pipkin thermal law, 2021, 123, 18758576, 181, 10.3233/ASY-201631
    5. Yang Liu, Wenke Li, A class of fourth-order nonlinear parabolic equations modeling the epitaxial growth of thin films, 2021, 14, 1937-1632, 4367, 10.3934/dcdss.2021112
    6. Danhua Wang, Wenjun Liu, Kewang Chen, Well-posedness and decay property for the Cauchy problem of the standard linear solid model with thermoelasticity of type III, 2023, 74, 0044-2275, 10.1007/s00033-023-01964-4
    7. Doaa Atta, Ahmed E. Abouelregal, Hamid M. Sedighi, Rasmiyah A. Alharb, Thermodiffusion interactions in a homogeneous spherical shell based on the modified Moore–Gibson–Thompson theory with two time delays, 2023, 1385-2000, 10.1007/s11043-023-09598-9
    8. İbrahim TEKİN, Identification of the time-dependent lowest term in a fourth order in time partial differential equation, 2023, 72, 1303-5991, 500, 10.31801/cfsuasmas.1127250
    9. Husam Alfadil, Ahmed E. Abouelregal, Marin Marin, Erasmo Carrera, Goufo-Caputo fractional viscoelastic photothermal model of an unbounded semiconductor material with a cylindrical cavity, 2023, 1537-6494, 1, 10.1080/15376494.2023.2278181
    10. Danhua Wang, Wenjun Liu, Global Existence and Decay Property for the Cauchy Problem of the Nonlinear MGT Plate Equation, 2024, 89, 0095-4616, 10.1007/s00245-024-10126-5
    11. Flank D. M. Bezerra, Lucas A. Santos, Maria J. M. Silva, Carlos R. Takaessu, A Note on the Spectral Analysis of Some Fourth-Order Differential Equations with a Semigroup Approach, 2023, 78, 1422-6383, 10.1007/s00025-023-01999-z
    12. Carlos Lizama, Marina Murillo-Arcila, On the existence of chaos for the fourth-order Moore–Gibson–Thompson equation, 2023, 176, 09600779, 114123, 10.1016/j.chaos.2023.114123
    13. Wen-jun Liu, Zhi-yu Tu, Equivalence between the internal observability and exponential decay for the Moore-Gibson-Thompson equation, 2024, 39, 1005-1031, 89, 10.1007/s11766-024-4133-5
    14. Ahmed E. Abouelregal, Marin Marin, Holm Altenbach, Thermally stressed thermoelectric microbeam supported by Winkler foundation via the modified Moore–Gibson–Thompson thermoelasticity theory, 2023, 103, 0044-2267, 10.1002/zamm.202300079
  • 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(6940) PDF downloads(1243) Cited by(21)

Figures and Tables

Figures(2)  /  Tables(2)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog