28.4. Linear vector equation#
Differential equation.
Integral eqaution. On a control volume \(V\) at rest
Integral equation in an arbitrary domain \(v_t\) can be derived using Reynolds’ transport theorem.
Method of characteristics. Let \(\mathbf{A}\) be diagonalizable,
Using spectral decomposition and multiplying the PDE by \(\mathbf{L}\), the PDE is transformed in a set of scalar PDE equations,
or
having introduced the characteristic variables \(\mathbf{v}(x,t) := \mathbf{L} \mathbf{u}(x,t)\). If the source is not function of the unknown function \(\mathbf{u}(x,t)\), this is a system of decoupled PDEs. The solution of the vector PDE follows from the solution of \(n\) scalar PDE linear equations. Using methods of characteristics, the solution of the \(i^{th}\) scalar PDE equation can be recast as the solution of pairs of ODEs, after writing \(V_i(t) = v_i(X_i(t), t)\), being \(X_i(t)\) the equation of a curve belonging to the \(i^{th}\) familiy of characteristics,
As the spectral decomposition of a constant matrix is constant, the characteristic lines are straight lines \(X_i(t) = s_i t + c\). In general, the solution of the dynamical equations of the characteristic variables \(V_i\) is a full system of \(n\) ODEs.
Particular case: no source. As a particular case, if the source term is identically zero, \(\mathbf{r}(x,t) = \mathbf{0}\), characteristic variable \(V_i(t)\) is constant along each characteristic line \(X_i(t;\dots)\) of the \(i^{th}\) family of characteristics.
…