Thankyou for the answers. For the moment I just excluded the pf calculation from the algorithm since I don't really need it, I just need the se. I also excluded the loads from the calculation, since i have data from a real network and the loads are mesured (or estimated but it's the same concept) with an uncertanty. I was thinking to just consider them as negative generators and that should do. It might be a problem if I have to add multiple loads or generators to the same node, since the se algorithm can only evaluate the total power injection in the node, and can not distinguish between them. It would be interesting to think of a solution to be able to include multiple power mesurements to the same node, I might have to do it in the next days. I also excluded the reference voltage magnitude, leaving just the reference angle, since I need it to be free to change according to the available mesurements. If I find the time i will try the new algorithm you send me and give you some feedback, thankyou for the help. If I have more questions I will write them here. My final goal is to think of some way to include network parameters uncertanties in the se calculations, if I manage to do something meaningful I'll send it to you. Thankyou again, Niccolò Citroni
2016-12-06 15:55 GMT+01:00 Ray Zimmerman <[email protected]>: > Thanks Rui. > > Quick question … is there a reason to do the updating of the Pg and Qg > values after, rather than before calling pfsoln()? I’ve attached the > updated version I propose to include in v6.0. > > And what about Niccolò’s second question regarding the results being > affected by system load rather than simply the measurements? > > Thanks, > Ray > > > > > > On Dec 6, 2016, at 2:35 AM, Ray Bo <[email protected]> wrote: > > > > Hi Ray, > > Sure I will take care of it. > > > > Niccolò, > > As I don't have 6.02b2 right now, I did a quick check on v5.1, and did > see the issue you brought up. The state estimation results are actually > correctly presented in the last section of the screen output, where you can > see the Pg2=0.3034 pu, Pg3=0.1336 pu. So the gen, load and losses are > balanced. The Pg2 and Pg3 values you see in the 'Generator Data' output > section are actually the default values from the 3-bus case, and appear to > be inbalanced with load. The reason Pg2 and Pg3 values do not get updated > in the 'Generator Data' output section is that, to output the power flow > solution in a nice format, I simply took advantage of the MATPOWER function > 'pfsoln' to update bus, gen, branch data structures to match power flow > solution. This function however only updates Pg for the slack bus generator > and not for the rest of the generators (because those are PQ and PV buses > and there is no need to do it). > > > > To clean up the output to avoid such confusion, I have implemented a > quick fix for the issue by updating the Pg and Qg using state estimation > results. The code with the quick fix is attached. I haven't tested it > extensively. Please use it and let me know if you find it useful or if you > have additional questions. I can be reached at [email protected]. > > > > Thanks, > > Rui > > > > -----Original Message----- From: Ray Zimmerman > > Sent: Monday, December 5, 2016 9:27 AM > > To: Rui Bo > > Cc: MATPOWER discussion forum > > Subject: Re: state estimation: active power not balanced in test cases > > > > Hi Rui and Niccolò, > > > > The state estimation code was contributed by Rui Bo, so I’m not that > familiar with it. Rui, I was wondering if you might be able to address > Niccolò's questions. > > > > Thanks, > > > > Ray > > > > > > > >> On Dec 1, 2016, at 6:20 PM, Niccolò Citroni <[email protected]> > wrote: > >> > >> Hallo, I'm new to matpower, using it for my master thesis. > >> I'm trying to figure out how the se program works, and running the > included test cases I noticed the following: > >> there is a big unbalance in active power in the network, for example in > the 3 bus case there are more than 380 MW of active power in excess, > considering generation, load, and losses. How is that possible? > >> Another thing: modifing the load i get different results runnning the > se. Why is that? shouldn't the state estimation be based only on the input > mesurements and the topology of the network? How has the load and generator > power anything to do with the se, when not included in the mesurements? > From what I know the power balance at each node should be a result of the > se, not part of the input data (when not as mesurements of course). > >> I hope I've been clear enough. > >> I'm using version 6.0b2 > >> Thankyou for the help and the program. > >> Niccolò Citroni > > > > <run_se.m> > > >
