Fix problem in state estimation

[aldalahmeh]

In test_se.m file solves problem 6.6 in the book "Computational Methods for Electric Power Systems"
by Mariesa Crow, which seems to be the first edition of the book. In the 2nd edition the problem's number
changed to 9. Since the solution is not available, I tried using the code on example 6.17,
which is a 3-bus system similar to problem 9 previously mentioned. However, the code is hardwired
for problem 9 and needs to be changed in several places. For example, real and reactive power demands
at bus 3 (P3, Q3) are not defiend as measurements, so they need to be added. Consequently,
the files:
test_se.m
run_se.m
doSE.m
outputsesoln.m

should be modified to include the new parameters, which is simple to do.

Unfortunately, just accommodating the new parameters in the code does not yield the same intermediate
calus in the solution steps of example 6.17. This is because the derivative matrix (H) has
some wrong values. This is traced back to computing the complex power matrix S.

As such, I created a new function to compute it (cmptSmat.m). So, the code is expanded to accommodate any sets of measurements. A test file "test_se_Ex_6_17.m" is provided that solves example 6.17 in the above 2nd edit of the book. The file shows the difference in computing H between the original, which is edited to allow more measurements, and the modified code. The modified code produced results that coincide with the solution of example 6.17 in the book. I can provide snapshots of example 6.17 solution upon request, since I cannot upload them here.

[rdzman]

Since the se directory is a subrepo here in matpower-extras, I've moved this to the master repo. Let's take the discussion to MATPOWER/mx-se#1.

Thanks!