[BUG] Sparse Matrix Error in some state estimation cases due to lack of numerical pivoting in the LU sparse solver

[mgovers]

Describe the bug

The in-house LU Sparse solver does not do numerical pivoting for performance optimization. It assumes that the matrix is always diagonal major.

However, this assumption does not hold for certain real-world state estimation cases. This results in Sparse Matrix Error where the system is actually perfectly observable and solvable.

To Reproduce

Run the test cases introduced in #852 and #854 . See that they run as expected fail (xfail) instead of pass.

Expected behavior

The test cases introduced in #852 and #854 pass