Linear solver for c = theta * S

[RemDelaporteMathurin]

https://github.com/RemDelaporteMathurin/FESTIM/blob/5e710594217d05156e9fd947897b86eb0f9d0b76/festim/concentration/theta.py#L84-L101

The problem here should be a LinearProblem instead of nonlinear.

The following changes should be done

    def post_processing_solution_to_concentration(self):
        """Converts the post_processing_solution from theta to mobile
        concentration.
        c = theta * S.
        The attribute post_processing_solution is fenics.Product (if self.S is
        festim.ArheniusCoeff)
        """
        problem = f.LinearVariationalProblem(f.lhs(self.form_post_processing), f.rhs(self.form_post_processing), self.post_processing_solution, [])
        solver = f.LinearVariationalSolver(problem)
        solver.solve()

    def create_form_post_processing(self, V, materials, dx):
        F = 0
        v = f.TestFunction(V)
        self.post_processing_solution = f.TrialFunction(V)
        F += -self.post_processing_solution * v * dx
        for mat in materials.materials:
            if mat.solubility_law == "sievert":
                F += self.solution * self.S * v * dx(mat.id)
            elif mat.solubility_law == "henry":
                F += self.solution**2 * self.S * v * dx(mat.id)
        self.form_post_processing = F
        self.post_processing_solution = f.Function(V)

[RemDelaporteMathurin]

This is needed as sometimes the nonlinear solver for theta doesn't converge...