This should not be happening. In fact, even when using a much smaller tolerance
the power flow should solve without iterating.
>> mpc = loadcase('case14');
>> mpopt = mpoption('verbose', 2, 'out.all', 0, 'pf.tol', 1e-5);
>> r = runpf(runopf(mpc, mpopt), mpopt);
MATPOWER Version 6.0b1, 01-Jun-2016 -- AC Optimal Power Flow
MATLAB Interior Point Solver -- MIPS, Version 1.2.1, 01-Jun-2016
(using built-in linear solver)
it objective step size feascond gradcond compcond costcond
---- ------------ --------- ------------ ------------ ------------ ------------
0 13212.573 0.624343 0.275 18.5289 0
1 8077.4637 0.77403 0.0237554 0.0669394 6.14623 0.221221
2 8332.0151 0.14646 5.5179e-05 0.0423581 1.61662 0.0140811
3 8272.2352 0.095342 6.05388e-05 0.058572 0.193778 0.00326095
4 8132.6028 0.38603 0.000981254 0.00654534 0.0253149 0.00764178
5 8108.5602 0.1963 0.000743622 0.00106529 0.00469505 0.00132593
6 8086.8978 0.13748 0.000934896 0.00124858 0.00124516 0.00119625
7 8083.3486 0.085386 0.000153596 0.00183126 0.000400753 0.000196232
8 8081.8608 0.052248 7.45399e-05 0.000847285 7.24952e-05 8.22733e-05
9 8081.5752 0.022569 2.13288e-05 0.000271899 1.1834e-05 1.57921e-05
10 8081.5296 0.0036972 1.37272e-07 5.77172e-08 1.26644e-06 2.52494e-06
11 8081.5251 0.00016701 7.2761e-10 1.53437e-10 1.26893e-07 2.46571e-07
Converged!
MATPOWER Version 6.0b1, 01-Jun-2016 -- AC Power Flow (Newton)
it max P & Q mismatch (p.u.)
---- ---------------------------
0 2.142e-09
Converged!
What version of MATPOWER are you using? I think older versions (not sure which
ones without checking) did not update the generator voltage set points in the
OPF solution, so feeding it directly to the power flow used the original,
rather than the solved, voltage magnitudes for generator buses. Maybe that’s
the problem?
Ray
> On Jul 1, 2016, at 1:17 AM, Georgiadis Dionysios
> <dionysios.georgia...@frs.ethz.ch> wrote:
>
> Dear all,
>
> I am running a OPF on a the IEEE14, storing the results.
> I then pass the same results to a PF solver, with an accuracy of 1e-2. To my
> surprise, the solver iterates, returning a solution different from that of
> the OPF.
>
> My question is, should the above be happening? Since the constraints of the
> OPF ensures that the returned case already adhere to the conservation of
> flow and is thus an acceptable solution.
>
> Thank you for your time and help.
>
> Kind regards,
> Dionysios Georgiadis
