Dear Dr Ray, I was able to implement what i wanted when I said I was trying to implement an OPF where the real power dispatch is pre-determined (as in the case of an energy market where the Q schedules are managed separately by the transmission operator), by fixing the Pmin=Pmax equal to the energy schedules, and allowing only one "slack" bus to move, since the real power variation was only minimal.
Now I am trying to determine an optimum "reactive power schedule" by trying to simulate the given the current ancillary service payment arrangement here in our country. The reactive power outputs of generators beyond 0.9 lead and 0.85 lag are paid with fixed 4$/kVAr cost. All reactive power generated within these limits are not compensated. So i tried to model this scenario by using a piecewise linear cost function for generators. The cost function below is for one specific generator in my data. 1 0 0 5 -78 4000 -51.82 -0.000001 0 0 66.3 0.000001 102 4000; Please see attached figure for the cost function I wanted to achieve. Basically I wanted the solution of the OPF to schedule as much MVArs in region A. However it seems that because of the (-) negative term for the reactive power absorption, it tends to schedule more MVArs in region B. Did I use the piecewise linear function correctly? How can I improve my cost function model to achieve the least amount of VArs shceduled outside region A? Thank you very much Sir. Vids On Fri, Oct 17, 2014 at 2:52 AM, Ray Zimmerman <[email protected]> wrote: > If you used the A, l and u fields of the MATPOWER case struct (mpc) to > specify the constraints, then the multipliers are in … > > results.lin.mu.l.usr > results.lin.mu.u.usr > > Note that these will be relative to power in p.u. not in MW, so you’ll have > to divide by baseMVA to get prices in $/MW. > > -- > Ray Zimmerman > Senior Research Associate > B30 Warren Hall, Cornell University, Ithaca, NY 14853 USA > phone: (607) 255-9645 > > > On Oct 16, 2014, at 12:52 PM, vids <[email protected]> wrote: > > Hi Dr Zimmerman, > How do i access the multipliers for the additional constraint that I > added? thank you > > On Tue, Sep 2, 2014 at 10:26 PM, Ray Zimmerman <[email protected]> wrote: > > Can you accomplish what you want simply by including the base dispatch in > the PD column of the bus matrix (where injections would appears as negative > loads)? > > -- > Ray Zimmerman > Senior Research Associate > B30 Warren Hall, Cornell University, Ithaca, NY 14853 USA > phone: (607) 255-9645 > > On Aug 28, 2014, at 10:10 PM, vids <[email protected]> wrote: > > Thank you very much, Dr. Shri. Yes i think your suggestion about > making the delta Pg as the control variables is a great idea. However > I am new to Matpower. Can you help me/ direct me to some examples on > how i can accomplish this? Thank you very much... > > On Thu, Aug 28, 2014 at 12:23 AM, Abhyankar, Shrirang G. > <[email protected]> wrote: > > Vids, > Implementing your reformulated OPF equations, written in complementarity > form, is non-trivial in MATPOWER as it will require modifying the > variable/equation sizes and muddling with the OPF data structures. Note > that you'll also need additional equations, perhaps expressed in > semi-smooth form, relating your upward/downward balancing service to > generator power deviation. You will have to spend some time to understand > the OPF data structures and how they are used in the various OPF routines. > > One other possible way (that I think will work) is by using the real power > generator deviation \Delta{Pg} as the control variable instead of Pg (see > the attached equations). This will keep the sizes of the > variables/equations for the reformulated OPF same as the original one. > However, you will have to modify the cost function, gradient, Hessian, and > the generator real power limits accordingly. > > Shri > > -----Original Message----- > From: vids <[email protected]> > Reply-To: MATPOWER discussion forum <[email protected]> > Date: Wed, 27 Aug 2014 16:18:14 +0800 > To: <[email protected]> > Subject: Modifying the Power Balance Equations > > Hi Dr Zimmerman and Matpower Community, > > I am trying implement an OPF where the real power dispatch is > pre-determined (as in the case of an energy market where the Q > schedules are managed separately by the transmission operator). > > Is it possible to implement it in Matpower? My idea is to add "slack" > variables in the nodal energy equations > > Pgi - ΔPgi + Pb1i - Pb2i - Pdi = ΣViVjYij(cos(θij + δi -δj) > > Pgi - fixed/predetermined real power generated at node i > ΔPgi - real power 're-scheduling' due to the reactive power dispatch > Pb1i - upward balancing service at node i > Pb2i - downward balancing service at node i > > Pb1i and Pb2 will have non-zero values when ΔPgi is nonzero, prompting > other generators to compensate the real power 're-scheduling' when > needed. > > This is the formulation in the dissertation of Dr. El-Samahy, and I > am wondering if this can be implemented in matpower. > > Any ideas would greatly be appreciated. Thank you very much. > > -- > 2 Cor 12:9 > Each time he said, "My grace is all you need. My power works best in > weakness." So now I am glad to boast about my weakness, so that the > power of Christ can work through me. > > > > > > > -- > 2 Cor 12:9 > Each time he said, "My grace is all you need. My power works best in > weakness." So now I am glad to boast about my weakness, so that the > power of Christ can work through me. > > > > > > > -- > 2 Cor 12:9 > Each time he said, "My grace is all you need. My power works best in > weakness." So now I am glad to boast about my weakness, so that the > power of Christ can work through me. > > > -- 2 Cor 12:9 Each time he said, "My grace is all you need. My power works best in weakness." So now I am glad to boast about my weakness, so that the power of Christ can work through me.
