Navier-Stokes equations

Table of Contents

1. Mathematical formulation

We denote \(\Omega \subseteq \mathbb{R}^d\), \(d = 2\) or \(d=3\), the fluid domain and \(\partial \Omega = \Gamma_D \cup \Gamma_N\) its boundary. On \(\Gamma_D\) the Dirichlet boundary conditions are imposed and on \(\Gamma_N\) the Neumann boundary conditions are imposed. The fluid density is given by \(\rho_f\) and its viscosity by \(\mu\). The fluid velocity \(u(x,t) : \Omega \times\) ]\(0,T\)] \(\mapsto \mathbb{R}^{d}\) and pressure \(u(x,t) : \Omega \times \)]\(0,T\)] \(\mapsto \mathbb{R}\) satisfy the Navier-Stokes equations:

\[\begin{equation*} \left\{\begin{array}{rcl} \rho_f \frac{\partial u}{\partial t} + \rho_f (u \cdot \nabla)u - \mu \Delta u + \nabla p &=& f \mbox{,} \quad \mbox{in } \Omega \texttt{,} \\ \nabla \cdot u &=& 0 \mbox{,} \quad \mbox{in } \Omega \texttt{,}\\ u(x,t) &=& h(x,t) \;, \quad \mbox{ on } \Gamma_D \mbox{,} \\ \sigma(x)\vec{n} &=& (-p I_d + 2 \mu D(u))\vec{n} = g(x,t) \;, \quad \mbox{ on } \Gamma_N \mbox{.} \end{array}\right.\; \end{equation*}\]

where \(f(x,t): \Omega \times\) ]\(0,T\)] \(\mapsto \mathbb{R}^{d}\) represents the external forces.

Variational formulation

We use the following functional spaces as test and trial spaces :

\[X = H^1(\Omega)^{d} \mbox{,} \quad M = L^2_0(\Omega) \mbox{.}\]

The variational formulation of the Navier-Stokes problem is given by :

Find \((u,p) \in X \times M\) such that:

\[\int_{\Omega} \rho_f \frac{\partial u}{\partial t} \cdot v + \int_{\Omega} \rho_f (u \cdot \nabla)u \cdot v + 2 \mu \int_{\Omega} D(u):D(v) - \int_{\Omega} p \nabla \cdot v = \int_{\Gamma_N} g \cdot v + \int_{\Omega} f \cdot v \;, \quad \forall v \in X \mbox{,}\]
\[\int_{\Omega} q \nabla \cdot u = 0 \;, \quad \forall q \in M \mbox{,}\]

where \(D(u) = \frac{1}{2}(\nabla u + \nabla u^T)\), the deformation tensor.

We denote:

\[\begin{align*} a(u,v) &= \int_{\Omega} \rho_f \frac{\partial u}{\partial t} \cdot v + \int_{\Omega} \rho_f (u \cdot \nabla)u \cdot v + 2 \mu \int_{\Omega} D(u):D(v) \mbox{,} \\ b(v,p) &= - \int_{\Omega} p \nabla \cdot v \mbox{,} \\ F(v) &= \int_{\Omega} f \cdot v \mbox{,} \\ G(v) &= \int_{\Gamma_N} g \cdot v \mbox{.} \\ \end{align*}\]

Thus, the variational formulation can be rewritten as:

Find \((u,p) \in X \times M \) such that:

\[\begin{align*} a(u,v) + b(v,p) &= G(v) + F(v) \;, \quad \forall v \in X \mbox{,} \\ b(u,q) &= 0 \;, \quad \forall q \in M \mbox{.} \end{align*}\]

The integration in time is solved by the Euler-Backward method.