Simulation of thermal disturbances with finite wave speeds using a high order method (Q2271953)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Simulation of thermal disturbances with finite wave speeds using a high order method
scientific article

    Statements

    Simulation of thermal disturbances with finite wave speeds using a high order method (English)
    0 references
    5 August 2009
    0 references
    The paper presents a numerical resolution of a hyperbolic 1D (in space) heat-flux equation using a least square method. The heat-flux equation is written as \(\gamma \partial T/\partial t+\partial q/\partial x=0\); \(\tau \partial q/\partial t+q=-k\partial T/\partial x\), posed in \(\Omega =(0,L)\times (0,T)\), where \(T\) is the temperature and \(q\) the heat flux. \( \gamma \), \(\tau \) and \(k\) are the usual physical parameters. When \(\tau =0\), the second equation called Cattaneo's equation reduces to Fourier's heat conduction law. The problem can be written as a system \(\mathcal{L}u=g\), where \(\mathcal{L}\) is a linear operator easily deduced from the above equations. Boundary conditions \(\mathcal{B}u=u_{\Gamma }\) are imposed on \( \Gamma =\{0,L\}\times (0,T)\). The least square method first consists to minimize the energy functional \(1/2\left\| \mathcal{L}u-g\right\| _{Y(\Omega )}^{2}+1/2\left\| \mathcal{B}u-u_{\Gamma }\right\| _{Y(\Gamma )}^{2}\) where the spaces \(Y(\Omega )\) and \(Y(\Gamma )\) are such that \((\mathcal{L},\mathcal{B})\) is a continuous mapping from a space \( X(\Omega )\) onto the space \(Y(\Omega )\times Y(\Gamma )\). This leads to a variational formulation: find \(u\in X(\Omega )\) such that \(\mathcal{A}(u,v)= \mathcal{F}(v)\) for every \(v\in X(\Omega )\). The domain \(\Omega \) is then decomposed in \(N_{e}\) non-overlapping subdomains \(\Omega _{e}\) of diameter \( h_{e}\) each of which being mapped to the unit cube \([-1,1]^{d}\), where \(d\) is the dimension of \(\Omega \), by an invertible mapping. In each \(\Omega _{e} \), the unknown solution \(u_{h}^{e}\) is then approximated by a polynomial of degree less or equal than \(Q_{e}\). The approximate solution is obtained when gluing these approximate solutions \(u_{h}^{e}\). Finally a \(C^{11}\) approximation is built using interpolation basis functions. The main part of the paper presents two examples of situations where this method can be used. The author indeed considers a parallel sided slab and a pulse heating. In each case, he compares the numerical simulations when considering Cattaneo's and Fourier's laws.
    0 references
    0 references
    0 references
    0 references
    0 references
    heat transfer
    0 references
    transport properties
    0 references
    hyperbolic heat equation
    0 references
    Cattaneo's law
    0 references
    Fourier's heat conduction law
    0 references
    least square method
    0 references
    space-time formulation
    0 references
    0 references
    0 references