[BUG] Sparse Matrix Error in ill-conditioned state estimation for meshed grids

[TonyXiang8787]

After #853 is fixed by #862 , the source of the SparseMatrixError problem has been fixed using pivot perturbation. However, pivot perturbation is not enabled (yet) for meshed grids, which means the ill-conditioned state estimation for meshed grids could still throw SparseMatrixError.

The reason that we cannot enable pivot perturbation in meshed grids is the lack of the sufficient observability condition check for meshed grids. If we are not sure whether the system is sufficiently observable, enable pivot perturbation could create fake observability.

This issue is automatically solved once the full sufficient observability check for all grids has been implemented as described in #84.

We can create additional ill-conditioned meshed grids for validation cases with xfail.

[TonyXiang8787]

@petersalemink95 this is needed after we fix the LU solver. When the observability check is fully implemented, this issue is automatically solved.