This is good question. It turns out that, in general, increasing the voltage
tends to reduce losses, which in turn reduces total generation and generation
cost. So minimizing generation cost also tends to dispatch reactive power so as
to push voltages upward.
Fiddling with the OPF starting point or optimizer parameters will sometimes
allow it to converge to a slightly different locally optimal solution, often
with nearly identical active power dispatch, but slightly different reactive
power dispatch and voltage profile.
I too find it note-worthy that in practice these multiple local optima are not
encountered more frequently than they are.
Ray
> On Oct 2, 2022, at 5:59 AM, Mohsen Jorjani <[email protected]> wrote:
>
> Dear all,
>
> Hi. I have a question about the OPF problem modelled in the MAFPOWER package.
> I know that if we just define the generator active power cost curve data in
> the gencost matrix and leave the generator reactive power cost curve data
> related fields empty, the objective function of the OPF problem would be
> reducing the generators active power cost in the system. What I can't figure
> out is that what forces the generators reactive powers and bus voltages
> values to be derived as same values when we run the OPF problem again and
> again on the same system. As you know, when the objective function is to
> reduce the generators active power cost, controllers like generators reactive
> power or bus voltages have negligibe impact in the solution. So, is it
> possible to have one set of generator active powers and various sets of
> generators reactive power and bus voltages as the final solution of the OPF
> problem when we run it again and again on the same system? Your explanation
> regarding my question would be much appreciated.
>
> Kind regards
