Kernel generalized divergence convergence

97 Views Asked by At

Let $\Omega \subset \mathbb{R}^n$ be bounded, open, Lipschitz. Let $A_n,A^{-1}_n \in L^{\infty}(\Omega,\mathbb{R}^n \times \mathbb{R}^n )$ such that:

$$A_n,A^{-1}_n \to I_d \quad \text{uniformly}$$

We can then define a "divergence type operator" $div_{An} :H^1(\Omega)^n\to L^2(\Omega)$ as:

$$div_{An}(u) = Tr(A_n\nabla u)$$

Cleary $div_{An}\to div$ in the operator norm since:

$$\|div_{An}(u)-div(u)\|_{L^{2}(\Omega)}\leq\|A_n-I\|_{L^\infty}\|u\|_{H^{1}(\Omega)}$$

I want to show that if $v \in Kernel(div)$ then there is sequence $v_n\in Kernel(div_{A_n})$ such that:

$$v_n \to v $$

In other words, that the corresponding orthogonal projections converges. I have try to use the Leray projection, but this approach involves a Neumann problem that it's not unique up to a constant.

Any ideas or thought are very welcome. Thanks!