The A matrix of (6.38) in the MATPOWER 7.0b1 User’s Manual will have to include the coefficients for all variables, both Pg and z.
Ray > On Apr 22, 2019, at 10:06 PM, Jubeyer Rahman <[email protected]> wrote: > > > Hi, > > Thanks a lot for your reply. > > Are you referring to Ar matrix ( The A matrix multiplied with Pg here) from > the reserve formulation (since this is a subtraction with the variable z)? > > So, If I want to implement anything like, Pg-z*factor= Po; where Po is the > dispatch in pre-contingency case how should I do it? > > The way I am thinking to do is, I will run the pre-contingency case, preserve > the 'Pg' data, and pass it as the 'Pmin' for the redistpatch period. Then I > will add the z variable with the OPF structure. > > I apologize I could not come up with a proper formulation as I am not quite > sure what I am thinking is correct or not. > > Regards, > Jubeyer > > > > > > > > > > > > On Tue, Apr 16, 2019 at 9:01 AM Ray Zimmerman <[email protected] > <mailto:[email protected]>> wrote: > Add a single constraint of the form … > > A * Pg - z = 0 > > … where A is a row vector, Pg is the vector of generation and z is a new > scalar variable representing the total amount of the resdispatch. > > Ray > > >> On Apr 9, 2019, at 5:56 PM, Jubeyer Rahman <[email protected] >> <mailto:[email protected]>> wrote: >> >> And for redispatch case as you have mentioned , if I want the generators >> redispatch in a certain way (make them produce power like Pg+some >> factor*some variable), how should I do that? >> >> On Tue, Apr 9, 2019 at 4:56 PM Jubeyer Rahman <[email protected] >> <mailto:[email protected]>> wrote: >> I am really sorry that I haven't been able to make it more clear. However, I >> will come up with a clear formulation later on. >> >> Actually, I have already formualted that and am getting some value for R. >> But I am not quite sure whether I have been able to get the cost formulation >> correct. >> >> So, for now , can you only tell me, that if I do get some value for R, which >> is supposed to come from some certain generator's, how can I make that >> calculated as a total power generation Pg+R from the gencost information and >> not from the reserve cost. >> >> On Tue, Apr 9, 2019 at 4:50 PM Ray Zimmerman <[email protected] >> <mailto:[email protected]>> wrote: >> I’m afraid I still don’t understand the problem you intend to solve. The >> formulation you provided is completely equivalent to the standard OPF with >> some additional variables and constraints that have no effect ultimately >> (since r_i = 0 is feasible). >> >> It sounds like you want to apply some kind of constraint to Pg to restrict >> redispatches from some initial dispatch or something, but the formulation >> you provided does not accomplish that. >> >> So, it seems like the first step would be to get the problem formulation >> clear and correct, then we can help if you have questions about the >> implementation. >> >> Ray >> >> >>> On Apr 9, 2019, at 4:33 PM, Jubeyer Rahman <[email protected] >>> <mailto:[email protected]>> wrote: >>> >>> Yes, I don't want the objective function to be affected in that way. The >>> reason of formulating that way is if I don't have all the generators active >>> in my system (in case I loose one generator for example, I want the rest >>> of the generator's to respond in a certain way). >>> >>> On Tue, Apr 9, 2019 at 4:30 PM Ray Zimmerman <[email protected] >>> <mailto:[email protected]>> wrote: >>> I think there must be something missing. Because the addition of these >>> variables and constraints will not affect the OPF solution at all. It will >>> be the same as the standard OPF formulation with no reserves. That is, the >>> original solution will still be both optimal (since there is no change to >>> the objective function in (6.34)) and feasible (since r_i = 0 is feasible >>> and imposes no additional restrictions on the problem). >>> >>> Your formulation as stated includes no reason for r_i to be non-zero. >>> >>> Ray >>> >>> >>> >>>> On Apr 9, 2019, at 3:33 PM, Jubeyer Rahman <[email protected] >>>> <mailto:[email protected]>> wrote: >>>> >>>> Ok. Here they go: >>>> >>>> similar to 7.2 >>>> >>>> 0<=r_i<=Pmax >>>> >>>> for 7.3: Since I don't want the cost to be calculated separately, I don't >>>> need anything here (reserves from generators should be calculated as the >>>> total power generation cost, no separate cost for generator) >>>> >>>> for 7.4 >>>> pg^i+x*r_i<=pg^i,max >>>> >>>> x is given as parameters here. >>>> >>>> for 7.5; since I don't have zonal requirement I don't have anything for >>>> that. >>>> >>>> Let me know if you need anything for more clarification. >>>> >>>> >>>> >>>> On Tue, Apr 9, 2019 at 3:13 PM Ray Zimmerman <[email protected] >>>> <mailto:[email protected]>> wrote: >>>> Ok, so you are attempting to modify the existing fixed reserves >>>> implementation to something with a similar, but not identical structure. I >>>> think I need to fully understand the formulation. Can you provide the >>>> equivalent of equations (7.2)–(7.5) for your problem so I can see exactly >>>> where the differences are? >>>> >>>> Ray >>>> >>>> >>>>> On Apr 9, 2019, at 11:43 AM, Jubeyer Rahman <[email protected] >>>>> <mailto:[email protected]>> wrote: >>>>> >>>>> Ok. Thanks a lot for your reply. >>>>> >>>>> What I am trying to implement is something like this 'Pg+x*R', where >>>>> 'Pg' is real power generation, x is a collection of factors (parameters) >>>>> usually fraction number ranging between 0 and 1, and R will be a variable >>>>> for reserves. Usually the minimum value for R is 0 and maximum value is >>>>> equal to the 'Pmax' for each generator asked to provide reserves. What I >>>>> also want is that the reserve cost to be ignored ,rather the cost of >>>>> total power generation 'Pg+x*R' should be calculated from the generator >>>>> cost information and not from the reserve costs ( I have tried that by >>>>> making all the reserve costs zero). In addition to these I have no zonal >>>>> reserve requirement ( I have made the constraint deactivated and deleted >>>>> the second row of the mpc.reserves.zones, deactivated mpc.reserves.req >>>>> and also deactivated where 'req' has been implemented). >>>>> >>>>> >>>>> Can you suggest how can how I do it? or do you have any comments on the >>>>> process I am already following? >>>>> >>>>> Just to illustrate more on the reserve cost modification: >>>>> >>>>> For example, I have 'Pg' from a particular generator (generator 1) 5 >>>>> MW, now after implementing the reserve , it is supplying another 1 MW >>>>> from its capacity (it's Pmax is 10 MW). Now what I want is that this >>>>> (5+1)=6 MW generation cost to be calculated by using the polynomial cost >>>>> information from the mpc.gencost section. >>>>> >>>>> >>>>> >>>>> On Tue, Apr 9, 2019 at 10:38 AM Ray Zimmerman <[email protected] >>>>> <mailto:[email protected]>> wrote: >>>>> I think it might help me to have a high-level view of what you are trying >>>>> to accomplish. If you are simply trying to *use* the already implemented >>>>> fixed reserve capability, you shouldn’t need to even concern yourself at >>>>> all with the implementation (i.e. the Ar matrix and the various callback >>>>> functions, etc.). In that case, all you need is to understand the inputs >>>>> in Table 7-5. If, on the other hand, you are modifying the implementation >>>>> to do something other than what is currently implemented, then I need to >>>>> understand what that is. >>>>> >>>>> In what is already implemented, the generation cost is simply the cost of >>>>> Pg. There is a separate cost of R that is added as a user cost. See >>>>> (7.3). So the cost coefficients of R are provided in mpc.reserves.cost >>>>> (see Table 7-5). >>>>> >>>>> Ray >>>>> >>>>> >>>>> >>>>>> On Apr 8, 2019, at 2:39 PM, Jubeyer Rahman <[email protected] >>>>>> <mailto:[email protected]>> wrote: >>>>>> >>>>>> Hi, >>>>>> >>>>>> I have another few questions regarding the addition of the fixed zonal >>>>>> reserves. So, far I understand, after adding the reserves, the real >>>>>> power output of the generator will be added with reserve amount, so in >>>>>> the part of the objective function where real power cost is being >>>>>> calculated, which power is fed into as for calculation is it the 'Pg' >>>>>> part of 'Pg+R' or is it the total 'Pg'? >>>>>> >>>>>> If I want to implement a reation like 'Pg+x*R' , where x is a collection >>>>>> of parameters (n-by-1) , which place can I feed into these parameters? I >>>>>> am assuming, this should be the second column of the Ar matrix. Is that >>>>>> correct? >>>>>> >>>>>> Regards >>>>>> >>>>>> >>>>>> >>>>>> On Fri, Apr 5, 2019 at 10:53 AM Jubeyer Rahman <[email protected] >>>>>> <mailto:[email protected]>> wrote: >>>>>> Please ignore the last email, I have figured this out. Every column in >>>>>> the first row corresponds the generators supposed to participate in the >>>>>> reserve provision , that's why they are made one. >>>>>> >>>>>> On Wed, Apr 3, 2019 at 5:27 PM Jubeyer Rahman <[email protected] >>>>>> <mailto:[email protected]>> wrote: >>>>>> Are you talking about the columns in the second row? >>>>>> >>>>>> On Mon, Apr 1, 2019 at 5:21 PM Ray Zimmerman <[email protected] >>>>>> <mailto:[email protected]>> wrote: >>>>>> The only thing you need to do is make sure the corresponding column in >>>>>> mpc.reserves.zones is all zeros. >>>>>> >>>>>> Ray >>>>>> >>>>>>> On Apr 1, 2019, at 10:31 AM, Jubeyer Rahman <[email protected] >>>>>>> <mailto:[email protected]>> wrote: >>>>>>> >>>>>>> Ok, I got your point and realized my mistake in understanding the zone >>>>>>> handling section. So, if I want some of the generator's choosing not to >>>>>>> provide ramp, should just setting the element of Identity matrix's >>>>>>> corresponding rows of first column of Ar be Ok? or I may need to change >>>>>>> something else as well? >>>>>>> >>>>>>> On Mon, Apr 1, 2019 at 9:45 AM Ray Zimmerman <[email protected] >>>>>>> <mailto:[email protected]>> wrote: >>>>>>> Regarding your first question, as described by (7.2) in the User’s >>>>>>> Manual, the reserve for a given generator is bounded above by both any >>>>>>> limit provided in mpc.reserves.qty (r_i^{max}) and by any physical ramp >>>>>>> rate (∆_i) given in mpc.gen(:, RAMP_10). It just so happens that the >>>>>>> example in t_case30_userfcn does not specify any physical ramp rates, >>>>>>> but the code still needs to handle cases which *do* provide physical >>>>>>> ramp limits. >>>>>>> >>>>>>> I’m not sure why you say only two generators are supposed to take part >>>>>>> in the reserve provision. In t_case30_userfcn there are two reserve >>>>>>> zones defined, but all 6 generators are able to participate in >>>>>>> providing the required reserves. >>>>>>> >>>>>>> You may want to review carefully the formulation in (7.2)–(7.5) and >>>>>>> Table 7-2. >>>>>>> >>>>>>> Ray >>>>>>> >>>>>>> >>>>>>> >>>>>>>> On Mar 29, 2019, at 4:06 PM, Jubeyer Rahman <[email protected] >>>>>>>> <mailto:[email protected]>> wrote: >>>>>>>> >>>>>>>> Referring to the 'userfcn_reserves_formulation', there is a line which >>>>>>>> is finding the value of k, which seems to be zero since none of the >>>>>>>> data in 'Ramp_10' column in t_case_30_userfcn is all zeros. so I >>>>>>>> don't see any point of using the line >>>>>>>> >>>>>>>> Rmax(k)=mpc.gen(k,Ramp_10), can you explain why the code is written >>>>>>>> that way. >>>>>>>> >>>>>>>> From my understanding only two generators are supposed to take part in >>>>>>>> the reserve provision, but the while putting the value for Rmax and >>>>>>>> Rmin, the code is considering all of them, which looks kind of >>>>>>>> unreasonable to me. Can you please explain this section as well? >>>>>>>> >>>>>>>> Regards, >>>>>>>> Jubeyer >>>>>>>> >>>>>>>> On Fri, Mar 29, 2019 at 12:43 PM Ray Zimmerman <[email protected] >>>>>>>> <mailto:[email protected]>> wrote: >>>>>>>> That is correct. All of the callbacks are technically optional. >>>>>>>> Typically you need the formulation callback to implement the actual >>>>>>>> problem modifications, and possibly ext2int and int2ext if you need to >>>>>>>> do some handling of input and output data, respectively. The printpf >>>>>>>> and savecase callbacks are only needed if you want to add things to >>>>>>>> the standard pretty-printed output or saved case data. >>>>>>>> >>>>>>>> Ray >>>>>>>> >>>>>>>> >>>>>>>>> On Mar 29, 2019, at 12:15 PM, Jubeyer Rahman <[email protected] >>>>>>>>> <mailto:[email protected]>> wrote: >>>>>>>>> >>>>>>>>> Just how important it is to include printpf and savecase callback >>>>>>>>> during the extension of OPF, if I don't really need anything printed >>>>>>>>> out right after I call the power flow? Will it be still possible to >>>>>>>>> extract information from the 'results' when I say >>>>>>>>> results=runopf(mycase)? >>>>>>>>> >>>>>>>>> To my understanding, after runopf being called, 'results' struct will >>>>>>>>> be returned and can be accessed by writing some command like >>>>>>>>> results.gen(:,2), etc. Let me know if I am thinking correctly or not? >>>>>>>>> >>>>>>>>> On Fri, Mar 29, 2019 at 9:53 AM Jubeyer Rahman <[email protected] >>>>>>>>> <mailto:[email protected]>> wrote: >>>>>>>>> Thank you very much. >>>>>>>>> >>>>>>>>> On Fri, Mar 29, 2019 at 8:43 AM Ray Zimmerman <[email protected] >>>>>>>>> <mailto:[email protected]>> wrote: >>>>>>>>> Are you attempting to use the provided extension for fixed reserves, >>>>>>>>> or are you attempting to write your own extension? >>>>>>>>> >>>>>>>>> If it’s the former, the full implementation is included in >>>>>>>>> toggle_reserves() >>>>>>>>> <http://www.pserc.cornell.edu/matpower/docs/ref/matpower6.0/toggle_reserves.html>. >>>>>>>>> Simply load your case file, use toggle_reserves() to enable the >>>>>>>>> callbacks, then run the OPF (or just call runopf_w_res() >>>>>>>>> <http://www.pserc.cornell.edu/matpower/docs/ref/matpower6.0/runopf_w_res.html>, >>>>>>>>> which does these 3 steps automatically for you). >>>>>>>>> >>>>>>>>> If you are attempting to write your own extension, I suggest making a >>>>>>>>> copy of toggle_reserves.m and rename it and all of the functions in >>>>>>>>> it and use it as a template for your own extension. >>>>>>>>> >>>>>>>>> Ray >>>>>>>>> >>>>>>>>> >>>>>>>>>> On Mar 28, 2019, at 12:40 PM, Jubeyer Rahman <[email protected] >>>>>>>>>> <mailto:[email protected]>> wrote: >>>>>>>>>> >>>>>>>>>> Hi, >>>>>>>>>> >>>>>>>>>> Recently I was digging through the extending OPF chapter of Matpower >>>>>>>>>> manual, but I don't quite catch the process. Regarding the example >>>>>>>>>> given there on 'Fixed zonal reserves' what I understand from my >>>>>>>>>> reading is, it is required to write down a call back function for >>>>>>>>>> formulation along with some call of callback functions. I have >>>>>>>>>> followed every steps mentioned there but could not make the code run >>>>>>>>>> (I am using version 6.0). I am adding my code snippet here for >>>>>>>>>> better conveying. >>>>>>>>>> >>>>>>>>>> %%% >>>>>>>>>> mpc=loadcase('case30.m'); >>>>>>>>>> mpopt = mpoption('out.all', 0, 'verbose', 0); >>>>>>>>>> mpc=add_usefcn(mpc,'formulation',@userfcn_reserves_formulation); >>>>>>>>>> mpc=ext2int(mpc,mpopt); >>>>>>>>>> results=runopf(mpc); >>>>>>>>>> results=int2ext; >>>>>>>>>> >>>>>>>>>> %%%% >>>>>>>>>> Error message: >>>>>>>>>> Access to an object's fields is only permitted within its methods. >>>>>>>>>> >>>>>>>>>> I have added the mpc.reserve data(cost, req, zones) posted in >>>>>>>>>> 't_case30_userfcns.m' file. >>>>>>>>>> I have written the userfcn_reserves_formulation in a different >>>>>>>>>> script , but it is not working. >>>>>>>>>> I didn't write the add_var and add_constraint explicitly since the >>>>>>>>>> add_userfcn callback function already contains those. >>>>>>>>>> >>>>>>>>>> Can you tell me what I am missing? >>>>>>>>>> >>>>>>>>>> Regards, >>>>>>>>>> Jubeyer >>>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>> >>>>>> >>>>> >>>> >>> >> >
