I assume you have modified the calculation of the constraint Jacobian (in opf_consfcn) and Hessian (in opf_hessfcn) as well as the constraints themselves. Have you verified that these calculations are correct? I suspect that an error in the Jacobian or Hessian is the most likely cause of convergence problems.
-- Ray Zimmerman Senior Research Associate 419A Warren Hall, Cornell University, Ithaca, NY 14853 phone: (607) 255-9645 On Dec 14, 2011, at 11:38 AM, 筱雅 谭 wrote: > Der Prof. Zimmerman, > > Thanks very much for your reply! > > I have downloaded the new version, but I'm still wondering why the > calculation sometimes won't converge after adding some constraints to > opf_consfcn.m. I've tried with small systems, as case 9, it worked very well, > but with case 30, it won't converge as far as I tried with different > parameters. > > Could u please give me some tips about how to find the problem by your > experience. > > Thanks very much in advance! > > Best regards, > Xiaoya > > From: [email protected] > Subject: Re: extension of nonlinear constrains > Date: Wed, 14 Dec 2011 11:15:58 -0500 > To: [email protected] > > Dear Xiaoya, > > I don't know if you saw that one of the features of the newly released > MATPOWER 4.1 is a very basic modeling of DC lines (see section 6.5.3 in the > manual for the details). This may not be a sufficiently detailed model for > you. Unfortunately, MATPOWER does not have an easy way to add non-linear > constraints to the problem. So adding them to opf_consfcn.m (and opf_hessfcn) > is currently the only way to do it. These changes will only apply to solvers > that use these functions (MIPS, fmincon, Knitro, Ipopt). > > -- > Ray Zimmerman > Senior Research Associate > 419A Warren Hall, Cornell University, Ithaca, NY 14853 > phone: (607) 255-9645 > > > > > On Dec 14, 2011, at 10:53 AM, 筱雅 谭 wrote: > > Dear Prof. Zimmerman, > > I'm working on adding DC lines to AC systems, and calculate the optimal power > flow. > > Is there any possibility to add nonlinear constraint to Matpower? As I > haven't found extension methods of nonlinear constrain. I just add the > constrain in opf_consfcn.m. The problem is that sometimes with different > scale of AC systems, the runopf can't converge. > > Looking forward to your advice. > > Best regards, > Xiaoya > > > > -
