On some linear parabolic PDEs on moving hypersurfaces (Q500880): Difference between revisions

This paper deals with the direct application of the existence theory to linear parabolic equations on abstract evolving Hilbert spaces to the justification of well-posedness of four problems on moving domains. To this aim, the authors recall the results presented in their previous paper. The first three problems are, namely, the Cauchy problem for the surface advection-diffusion equation: find a unique function \( v\) that verifies \[ b(v,\gamma)=0,\quad \text{ for any test function }\gamma; \] the Cauchy-Dirichlet problem for the bulk equation: find a unique function \(u\) that verifies \[ a(D; u,\omega)+\int_0^T\int_{\Omega(t)}(\mathbf{b}-\mathbf{w}) \cdot\nabla u \omega= \int_0^T\int_{\Omega(t)} f\omega, \] for any test function \(\omega\); the Cauchy-Robin problem for the coupled bulk-surface system: find a unique pair \((u,v)\) that verifies \[ \begin{aligned}\alpha a(1; u,\omega)+\beta b(v,\gamma) + \int_0^T\int_{\Gamma(t)}(\beta v-\alpha u)(\beta\gamma-\alpha \omega )=\alpha \int_0^T &\int_{\Omega(t)} f\omega+ \beta\int_0^T\int_{\Gamma(t)} g\gamma \\ &\text{ for any test function }(\omega,\gamma).\end{aligned} \] Here, \(D,\alpha,\beta>0\), the vector fields \(\mathbf{b},\mathbf{w}\) are sufficiently smooth functions defined in \(\mathbb{R}^{n}\times ]0,T[\), \[ a(D; u,\omega)=\int_0^T \langle \dot u(t),\omega(t)\rangle_{ \left( H^1(\Omega(t))\right)^*\times H^1(\Omega(t)) }+ \int_0^T\int_{\Omega(t)}\left( D \nabla u\cdot \nabla\omega+ u\omega\nabla\cdot\mathbf{b}\right), \] and \[ b(v,\gamma)=\int_0^T \langle \dot v(t),\gamma(t)\rangle_{ H^{-1}(\Gamma(t)) \times H^1(\Gamma(t)) }+ \int_0^T\int_{\Gamma(t)}\left( \nabla_\Gamma v\cdot \nabla_\Gamma\gamma+ v\gamma\nabla_\Gamma\cdot\mathbf{b}\right). \] The fourth problem is the dynamic boundary problem for the Laplace equation: find of a unique function \(u\), with \(v(t) := u(t)|_{\Gamma (t)}\) for each \(t\in ]0,T[\), that verifies in the weak sense, \(\Delta u(t)=0\) in \(\Omega(t)\), \(\dot v(t)+\frac{\partial u(t)}{\partial \nu(t)} +v(t)=f(t)\) on \(\Gamma(t)\), and \(v(0)=v_0\) on \(\Gamma(0)\).