Observations on blow up and dead cores for nonlinear parabolic equations (Q1113012)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Observations on blow up and dead cores for nonlinear parabolic equations
scientific article

    Statements

    Observations on blow up and dead cores for nonlinear parabolic equations (English)
    0 references
    0 references
    1989
    0 references
    Consider the following problem \((P_ 1):\) \[ (1)\quad u_ t-\Delta u=u^ p-a| \nabla u|^ q\quad t>0,\quad x\in \Omega, \] \[ u=0,\quad t>0,\quad x\in \partial \Omega;\quad u(0,x)=\Phi_ 1(x),\quad x\in \Omega. \] It has been known for quite a while that solutions of a simpler problem (without the gradient term) can blow up in finite time. Recently Chipot and Weissler investigated whether the blow up effect persists when one adds a dissipative term to the equation. They discussed the case \(1<p\) and \(1<q\leq 2p/(p+1).\) In the present paper the case \(q=2\) is studied in some detail. For positive a and \(1<p<2\) the gradient term is of no lower power than the source term \(u^ p\). Therefore the damping term does control the source term in the sense that there is no blow up. For any \(p>2\), however, one may still have blow up phenomena, so that the restriction on q of Chipot and Weissler appears to be technical. Blow up phenomena for other functions f(u), instead of \(u^ p\), are studied as well. An important step in the proofs is a transformation to so-called dead problems, such as \((P_ 2):\) \[ (2)\quad v_ t-\Delta v=-h(v)\quad t>0,\quad x\in \Omega, \] \[ v=1,\quad t>0,\quad x\in \partial \Omega;\quad v(0,x)=\Phi_ 2(x)\quad x\in \Omega. \] The function u blows up if and only if v tends to zero in finite time. Problem \((P_ 2)\) is analyzed by methods which are essentially due to Bandle and Stakgold. Finally the case \(a<0\) in \((P_ 1)\) is treated. This causes an effect which enhances blow up in the sense that sometimes blow up will occur on all of \(\Omega\) and not in a single point only. iii) If \(p>2\), there is single point blow up. The proof of this Theorem can be reduced to some known blow up results, once one finds the right transformation. If we set \(v(t,x)=e^{u(t,x)}- 1\), we obtain the transformed equation \[ v_ t-\Delta v=(1+v)(\log (1+v))^ p. \] Now the result follows immediately from the paper by \textit{A. A. Lacey} [Proc. R. Soc. Edinb., Sect. A 104, 161-167 (1986; Zbl 0627.35047)]. Remark 7. After this paper was completed we were kindly informed by \textit{V. A. Galaktionov} of the paper by himself and \textit{S. A. Posashkov} [Exact solutions of parabolic equations with quadratic nonlinearities, preprint, Moscow, in Russian (1988)], which contains exact solutions of \(u_ t-u_{xx}=u^ 2+| u_ x|^ 2\) on \({\mathbb{R}}^+\times {\mathbb{R}}\).
    0 references
    0 references
    0 references
    0 references
    0 references
    blow up
    0 references
    dissipative term
    0 references
    dead problems
    0 references
    0 references
    0 references