No, as I said below "If the OPF converges successfully all constraints will be satisfied." In other words, if success = 1 there should be no violated constraints. The extra printout is for binding constraints. These are not violated, they are simply operating at their specified limit.
-- Ray Zimmerman Senior Research Associate 419A Warren Hall, Cornell University, Ithaca, NY 14853 phone: (607) 255-9645 On Oct 1, 2012, at 12:19 PM, Evangelos Galinas <[email protected]> wrote: > Hello Ray, > > thanks a lot for the reply. So, if I get it right the suc=1 does not mean > necessarily at all that everything went right. From what I understood, > whenever there is a constrained violated there are outputs that tell you > which got violated (extra prints for gen and branch constraints). So in order > to make my code more eefficient and avoid all these checks is there a way to > get to know if at least one constraint got violated, for example somehow > accessing these extra prints? > > Thanks once more, > > Vagelis > > From: [email protected] > Subject: Re: OPF questions > Date: Mon, 1 Oct 2012 10:20:50 -0400 > To: [email protected] > > On Sep 30, 2012, at 3:06 PM, Evangelos Galinas <[email protected]> wrote: > > 1st: Is there a way to tell to the opf function to disregard the power limits > of Generators (min/max) or the branch flow limits? Or I have to change the > limits myself accrdingly in the system struct before I run a new OPF? > > You have to change the limits in the case data before running. Btw, setting > RATE_A for a line to zero removes the branch flow limit. > > 2nd: I realised that the suc value get zero only if the algortihm does not > converge. What I need to know for my implementation though, is at any point > to know if even one contraint (generator or branch) was violated. Is there a > way to quickly access a flag or something for that, or I'll have to do checks > myself between the 2 colums (e.g. Pg and Pmax) after every run? > Alternatively, maybe you have a more efficient way to do that? > > If the OPF converges successfully all constraints will be satisfied. If it > does not converge successfully, the values returned are not necessarily > meaningful at all. Specifically, I don't know of any of the solvers that will > return a solution with all constraints satisfied except a few that are > conflicting. > > In general, determining the minimal number of changes required to make an > infeasible problem become feasible is a difficult problem. > > > -- > Ray Zimmerman > Senior Research Associate > 419A Warren Hall, Cornell University, Ithaca, NY 14853 > phone: (607) 255-9645 > >
