Processing math: 100%

2024年3月4日 星期一

Solution to Partial Differential Equations: Methods and Applications (Robert McOwen) Section 5.1

Solution to Partial Differential Equations: Methods and Applications (Robert McOwen) Section 5.1

  1. A square two-dimensional plate of side length a is heated to a uniform temperature U0>0. Then at t=0 all sides are reduced to zero temperature. (Assume that the top and bottom of the plate are insulated so that the heat flow can be restricted to two dimensions.) Describe the heat diffusion u(x,y,t).
  2. SolutionThe initial/boundary value problem for heat equation can be formulated as

    {ut(x,y,t)=kΔu(x,y,t)=k(uxx(x,y,t)+uyy(x,y,t))for (x,y,t)(0,a)×(0,a)×(0,),u(x,y,0)=U0>0for (x,y)(0,a)×(0,a),u(0,y,t)=u(a,y,t)=0for y[0,a] and t>0,u(x,0,t)=u(x,a,t)=0for x[0,a] and t>0.

    Recall that the eigenvalues and associated eigenfunctions of Δ with zero Dirichlet boundary condition over the two-dimensional square are

    λmn=π2a2(m2+n2),ϕmn(x,y)=sinmπxasinnπyafor m,nN.

    The initial condition can be represented as

    U0=m,n=1Amnϕmn(x,y)=Amnsinmπxasinnπya,

    where Amn is given by

    Amn=4a2a0a0u(x,y,0)sinmπxasinnπyadydx=4a2U0a2mnπ2cosmπxa|a0cosnπya|a0=4(1(1)m)(1(1)n)mnπ2U0={16mnπ2U0if m and n are odd,0otherwise.

    Let us assume that the solution u can be expanded in the eigenfunctions with coefficients depending on t:

    u(x,y,t)=m,n=1vmn(t)ϕmn(x,y)=m,n=1vmn(t)sinmπxasinnπya.

    Assuming sufficient convergence to pass the derivatives inside the summation, this gives

    vmn(t)+kλmnvmn(t)=0for t>0 and m,nN.

    This ordinary differential equation has general solution

    vmn(t)=vmn(0)exp(kλmnt)for t0.

    Hence we get

    u(x,y,t)=m,n=1vmn(0)exp(kλmnt)sinmπxasinnπya.

    By the initial condition, we get vmn(0)=Amn for all m,nN. Therefore, we obtain

    u(x,y,t)=m,n=1Amnexp(kλmnt)sinmπxasinnπya=16U0π2m,n=11(2m1)(2n1)exp(kπ2[(2m1)2+(2n1)2]a2t)sin(2m1)πxasin(2n1)πya.


  3. Describe how you would solve the initial/boundary value problem (3) with the homogeneous Dirichlet condition replaced by u(x,t)=h(x) for xΩ and t>0. What happens to u(x,t) as t?
  4. SolutionConsider the following Laplace equation with the Dirichlet boundary condition

    {Δw(x)=0for xΩ,w(x)=h(x)for xΩ.

    Then we consider the following initial value with homogeneous Dirichlet condition for heat equation:

    {vt(x,t)=Δv(x,t)for xΩ and t>0,v(x,0)=g(x)w(x)for xˉΩ,v(x,t)=0for xΩ and t>0.

    By solving the above two problems, we can find that u(x,t)=v(x,t)+w(x) solves the following initial/boundary value problem

    {ut(x,t)=Δu(x,t)for xΩ and t>0,u(x,0)=g(x)for xˉΩ,u(x,t)=h(x)for xΩ and t>0.

    Now we study the asymptotic behaviour of u as t tends to infinity. Clearly, w is independent of t. Suppose that λn and ϕn are eigenvalues and eigenfunctions for Δ over Ω with the Dirichlet boundary condition, i.e., Δϕn=λnϕn in Ω with ϕn=0 on Ω. Note that λn>0 for all nN and λn as n. Then v can be represented as

    v(x,t)=n=1eλntϕn(x)Ωϕn(y)(gw)(y)dy.

    Since limteλnt=0, we obtain limtv(x,t)=0, which implies

    limtu(x,t)=w(x)+limtv(x,t)=w(x)for xˉΩ.


  5. Suppose f(x,t)=n=1an(t)ϕn(x) converges absolutely and uniformly for (x,t)ˉUT for every T>0. Let g(x)0 and h(x,t)0, and describe the solution of (9).
  6. SolutionSince g0 and h0, equations (9) become

    {ut(x,t)=Δu(x,t)+f(x,t)for xΩ and t>0,u(x,0)=0for xˉΩ,u(x,t)=0for xΩ and t>0.

    Let us assume that the soluion u can be expanded in the eigenfunctions with coefficients depending on t:

    u(x,t)=n=1vn(t)ϕn(x).

    Assuming sufficient convergence to pass the derivatives inside the summation, this gives

    vn(t)+λnvn(t)=an(t)for t>0 and nNwith vn(0)=0,

    where λn is the eigenvalue of Δ. Note that λn>0 for all nN. Multiplying it by eλnt, we get

    (eλntvn(t))=eλntvn(t)+λneλntvn(t)=eλntan(t)for t>0 and nN,

    which implies

    vn(t)=t0eλn(st)an(s)ds.

    Therefore, the solution u is given by

    u(x,t)=n=1ϕn(x)t0eλn(st)an(s)ds.


  7. More generally than in Exercise 2, describe how you would solve the initial boundary value problem (3) with the homogeneous Dirichlet condition replaced by u(x,t)=h(x,t) for xΩ and t>0.
  8. SolutionSuppose that h(x,t) can be extends to xˉΩ such that h has the eigenfunction expansion:

    h(x,t)=n=1an(t)ϕn(x),

    which converges absolutely and uniformly on ˉUT for every T>0. Let f(x,t)=ht(x,t)Δh(x,t) for xΩ and t>0. Clearly, f admits the eigenfunction expansion:

    f(x,t)=n=1(an(t)+λnan(t))ϕn(x),

    which also converges absolutely and uniformly on ˉUT for every T>0. Then we consider the following initial/boundary value problem with homogeneous Dirichlet condition and forcing term:

    {vt(x,t)=Δv(x,t)+f(x,t)for xΩ and t>0,v(x,0)=g(x)h(x,0)for xˉΩ,v(x,t)=0for xˉΩ and t>0.

    Using the solution of v, we may find that u(x,t)=v(x,t)+h(x,t) solves

    {ut(x,t)=Δu(x,t)for xΩ and t>0,u(x,0)=g(x)for xˉΩ,u(x,t)=h(x,t)for xΩ and t>0,

    which is desired.

    Now we derive the expansion solution of u as follows. By the results of Exercise 2 and Exercise 3, we can get

    v(x,t)=n=1eλntϕn(x)[Ωϕn(y)[g(y)h(y,0)]dy+t0eλns[an(s)+λnan(s)]ds].

    From u(x,t)=v(x,t)+h(x,t), we obtain

    u(x,t)=n=1ϕn(x)[an(t)+eλntΩϕn(y)[g(y)h(y,0)]dy+t0eλn(ts)[an(s)+λnan(s)]ds].


  9. Suppose the square plate of Exercise 1 is given an initial temperature distribution u(x,y,0)=g(x,y) and then insulated so that heat cannot flow across Ω (i.e., u/ν=0). Use an expansion in appropriate eigenfunctions to describe the heat diffusion u(x,y,t).
  10. SolutionThe initial/boundary value problem for heat equation can be formulated as

    {ut(x,y,t)=kΔu(x,y,t)=k(uxx(x,y,t)+uyy(x,y,t))for (x,y,t)(0,a)×(0,a)×(0,),u(x,y,0)=g(x,y)for (x,y)(0,a)×(0,a),ux(0,y,t)=ux(a,y,t)=0for y[0,a] and t>0,uy(x,0,t)=uy(x,a,t)=0for x[0,a] and t>0.

    Recall that the eigenvalues and associated eigenfunctions of Δ with zero Neumann boundary condition over the two-dimensional square are

    λmn=π2a2(m2+n2),ϕmn(x,y)=cosmπxacosnπyafor m,nN{0}.

    Using the eigenfunction expansion, the initial data can be represetned as

    g(x,y)=m,n=0Gmnϕmn(x,y)=m,n=0Gmncosmπxacosnπya,

    where

    Gmn={1a2a0a0g(x,y)dydxif m=n=0,2a2a0a0g(x,y)cosmπxadydxif m0, n=0,2a2a0a0g(x,y)cosnπyadydxif m=0, n0,4a2a0a0g(x,y)cosmπxacosnπyadydxif m0, n0.

    Let us assume that the solution u can be expanded in the eigenfunctions with coefficients depending on t:

    u(x,y,t)=m,n=0vmn(t)ϕmn(x,y)=m,n=0vmn(t)sinmπxasinnπya.

    Assuming sufficient convergence to pass the derivatives inside the summation, this gives

    vmn(t)+kλmnvmn(t)=0for t>0 and m,nN{0}.

    This ordinary differential equation has general solution

    vmn(t)=vmn(0)exp(kλmnt)for t0.

    Hence we get

    u(x,y,t)=m,n=0vmn(0)exp(kλmnt)cosmπxacosnπya.

    By the initial condition, we get vmn(0)=Gmn for m,nN{0}. Therefore, we get

    u(x,y,t)=m,n=0Gmnexp(kλmnt)cosmπxacosnπya.


  11. Suppose U=Ω×(0,T) and uC2;1(U)C(ˉU) satisfies utΔu+cu in U, where c0 is a constant. If u0, show that (11) holds for u. Give a counterexample when c<0 and u<0 in ˉU.
  12. SolutionIf the condition "u0" means u0 in ˉU, then it is easy to find that cu0 in U, and hence vtΔu+cuΔu. Clearly, (11) follows from the weak maximum principle.

    Now we consider the case that u(x0,t0)0 for some (x0,t0)ˉU. Without loss generality, we may assume that u(x0,t0)=maxˉUu0. Let us firstly assume that ut<Δu+cu in U. If (x0,t0)U(Ω×{T}), then u(x0,t0)=0 and vt(x0,t0)0 and Δv(x0,t0)0, which implies

    0ut(x0,t0)<Δu(x0,t0)+cu(x0,t0)0.

    This leads a contradiction, and hence (x0,t0)Γ, i.e., (11) holds true. For the case that utΔu+cu in U, we define uϵ(x,t)=u(x,t)ϵt for (x,t)ˉU and ϵ>0. Then uϵ satisfies

    uϵt=utϵΔu+cuϵ=Δuϵ+cuϵ+ϵ(ct1)<Δuϵ+cuϵ.

    Thus, by the previous case, we have

    max(x,t)ˉU[u(x,t)ϵt]=max(x,t)ˉUuϵ(x,t)=max(x,t)Γuϵ(x,t)=max(x,t)Γ[u(x,t)ϵt].

    Note that

    max(x,t)ˉUu(x,t)ϵTmax(x,y)ˉU[u(x,t)ϵt]max(x,t)ˉUu(x,t),max(x,t)Γu(x,t)ϵTmax(x,y)Γ[u(x,t)ϵt]max(x,t)Γu(x,t)for all ϵ>0.

    Letting ϵ0+, we obtain (11). The proof is complete.

    Suppose that c<0 and u<0 in ˉU. I thinkt the simplest example is that v(x,t)=ect for (x,t)ˉU. Then it is easy to verify that

    vt(x,t)=cect=cv=Δv+cuin U.

    Since c<0, function v is increasing in t and the maximum value of v occurs at (x,t)ˉΩ×{T}. Clearly, (11) does not hold.

    Remark 1. It seems useless that consider v(x,t)=ectu(x,t) for (x,t)ˉU. Surely, one may obtainvt(x,t)=cectu(x,t)+ectut(x,t)cv(x,t)+ectΔu(x,t)+cectu(x,t)=Δv(x,t)in U.Note that ect1 in ˉU when c0. Thus, by the weak maximum principle for v, we getmax(x,t)ˉUu(x,t)max(x,t)ˉU[ectu(x,t)]=max(x,t)ˉUv(x,t)=max(x,t)Γv(x,t)=max(x,t)Γectu(x,t),where Γ=(Ω×[0,T])(ˉΩ×{0}). This seems impossbile to derive (11).

    Remark 2. The original question is that "If u0, show that (11) holds for u. Give a counterexample without the condition u0." The condition u0 is ambiguous, which is already dealt within the solution. However, when c=0, the weak maximum principle still holds true so we need to require c<0. In addition, we need to make the assumption that u<0 in ˉU.

    Remark 3. When c0, we have max(x,t)ˉUu(x,t)max(x,t)Γu+(x,t), where u+=max{u,0}.


  13. If u satisfies (1), define its heat energy by E(t)=Ωu2(x,t)dx.
    1. If U=Ω×(0,) and uC2;1(ˉU) satisfies (1) and either (i) u=0 on Ω, or (ii) u/ν=0 on Ω, then prove that E(t) is nonincreasing in t.
    2. Use (a) to conclude the uniqueness of a solution uC2;1(ˉU) for either the nonhomogeneous Dirichlet problem (9) or the corresponding nonhomogeneous Neumann problem.
  14. Solution
    1. Since uC2,1(ˉU) satisfies ut=kΔu in U, we have

      E(t)=2Ωu(x,t)ut(x,t)dx=2kΩu(x,t)Δu(x,t)dx=2kΩu(x,t)uν(x,t)dSx2kΩ|u(x,t)|2dx=2kΩ|u(x,t)|2dx0(k>0).

      Here we have used the fact that uuν0 on Ω, which follows from either (i) u0 on Ω or (ii) uν0 on Ω. Threfore, E is nonincreasing in t. The proof is complete.
      • Suppose that there are two solutions ˜u and u of the nonhomogeneous Dirichlet problem (9), i.e.,

        {˜ut(x,t)=Δ˜u(x,t)+f(x,t)for xΩ and t>0,˜u(x,0)=g(x)for xˉΩ,˜u(x,t)=h(x,t)for xΩ and t>0,{ut(x,t)=Δu(x,t)+f(x,t)for xΩ and t>0,u(x,0)=g(x)for xˉΩ,u(x,t)=h(x,t)for xΩ and t>0,

        Let w(x,t)=˜u(x,t)u(x,t) for (x,t)ˉU. Then w satisfies

        {wt(x,t)=Δw(x,t)for xΩ and t>0,w(x,0)=0for xˉΩ,w(x,t)=0for xΩ and t>0

        and its heat energy is given by Ew(t)=Ωw2(x,t)dx. By (a), it is clear that

        Ωw2(x,t)dx=Ew(t)Ew(0)=Ωw2(x,0)dx=0,

        which implies w0 in ˉU and hence ˜uu in ˉU. Therefore, we complete the proof of uniquenees of solution of the nonhomogeneous Dirichlet problem.
      • Suppose that there are two solutions ˜u and u of the nonhomogeneous Neumann problem, i.e.,

        {˜ut(x,t)=Δ˜u(x,t)+f(x,t)for xΩ and t>0,˜u(x,0)=g(x)for xˉΩ,˜uν(x,t)=h(x,t)for xΩ and t>0,{ut(x,t)=Δu(x,t)+f(x,t)for xΩ and t>0,u(x,0)=g(x)for xˉΩ,uν(x,t)=h(x,t)for xΩ and t>0,

        Let w(x,t)=˜u(x,t)u(x,t) for (x,t)ˉU. Then w satisfies

        {wt(x,t)=Δw(x,t)for xΩ and t>0,w(x,0)=0for xˉΩ,wν(x,t)=0for xΩ and t>0

        and its heat energy is given by Ew(t)=Ωw2(x,t)dx. By (a), it is clear that

        Ωw2(x,t)dx=Ew(t)Ew(0)=Ωw2(x,0)dx=0,

        which implies w0 in ˉU and hence ˜uu in ˉU. Therefore, we complete the proof of uniquenees of solution of the nonhomogeneous Neumann problem.

沒有留言:

張貼留言