Training D12: Voltage Droop Control

over the next 10 years I've learned that I've learned to keep my expectations reasonable it happens slowly I'm afraid that they sounds like they're related to each other and as I present them they're similar but they really are two different things inside the powerful so I'm going to start with what's called voltage drift control in the power and there's a link on here where you can go to the health documentation to read about this I hover over this take me there here's a whole thing that talks about it in the health documentation you'll be able to tell I made these slides while I was working on the help documentation cuz some of the slides look the same where it started was starting in about 2013 this is what your renewable plants look like inside the power flow case I started getting questions from a guy at Idaho named Eric Bakey who many of you probably know hopefully regulate a bus in the power flow which you guys are used to doing our world has line drop compensation you can put on our generator where it looks out a certain impedance away which kind of worked for this but not exactly but what they really had is they wanted to regulate this point of interconnection against a curve and what they wanted to do is they'll have a group of generators up here these might be wind turbine or solar plants whoever they're all regulating this controlled bus voltage not to a specific voltage but against a curve and they also have a dead man so they'd say if the voltage is within this dead band I want the reactive power reaching the control bus to be some usually at zero basically I don't want to provide any bars but within the dead band and then outside of that as the voltage drops the reactive power reaching the controlled bus keeps going talking until eventually it hits some maximum output and the same thing happens on the other ok Isis and at first when I talked to Eric about it I was like the dead man seems kind of stupid to me to be honest with you but he was like he pulled up we run a WebEx he pulled up screens in his EMS and was like no look this is exactly what we put in our EMS for these wind plants and they have a dead band I feel like it's contractual it doesn't really cost them anything to provide it to provide bars but there is a dead man during which they don't provide any bars so then the question was can you do this in the power flow and the answer was no time you can you cannot model that in the powerful so before we talk about how it is modeled in the power flow now now let's step back and talk about what what is a normal generator a normal generator is kind of doing the same thing except its voltage is at a fixed set point until it hits a limit and then it the cube becomes fixed so this is the curve that it goes against and back in school you guys learned about PV and from Q buses and hitting and backing off limits and that's all built around this so the next thing to do is why don't we just kind of stretch that out and instead of having a constant voltage go ahead and have a slope across it solar farms do stuff like this where they had their regulating bus against the voltage droop control this is not new I was talking about the re PCA PCA model earlier today this is exactly what this path is doing this is voltage Dru control with dead band here's the dead band and here's the drew you either have complex power drew or just reactive characters one of the two so transient stability models have always included this as species do the same thing as well absolutely yeah so we can do this some transient stability but you could not do it in power flow could you take the parameters from stability or translate to the to their steady state equivalent to come up with the settings definitely yeah these this is the usually it's going this way because if you look at the curve I'm plotting this isn't complex power it's just key by itself the slope of this curve is the K value that's in in the stability except the stability models all in per unit normalized to the MBA basement machines you have to scale things based on so what we want to do or what I was being asked to do by a couple customers was provide the ability in the power flow to have this bus this bus right here map so that the queue coming in here and the voltage at this bus obeyed this curve so that is what we call voltage droop control and simulator now so inside simulator what we're gonna do is I do is what are we what do we already have what have we had for 25 years all right generators regulate a bus we already have that so you already go into power flow we go look at one of your hydro plants there might be 15 units all regulating a bus together so we already have to happen so we have the grouping of the generators but we didn't really have before is we need some way to say which reactive power am I talking about so what we do as software vendors in the transient stability realm is you have to tell us which branch for measuring reactive power on I've always hated that because frankly as a user community you guys suck at telling us that we've done it you leave it off if you go look at the RT PCA models actin power on I've always hated that because frankly as a user community you guys suck at telling us that you leave it off if you go look at the RT PCA models and in wet cases very very frequently the measured branch is just a minute it's not included so when up until about a year and a half ago in simulator when we would read a case like that we would treat that as an error and you would not be allowed to run the simulation until you told us the branch okay after about 15 support calls asking why do you keep giving us this air I don't get this air in pslf I was ie male 1 I was like what are you guys doing is like well if they don't tell us the branch we just assume the measurements always zero which means the device is doing nothing which means the device is doing nothing it's the same as disabling but people don't realize that so finally in simulator we do the same thing we just don't do anything but anyway the point is on this there's no reason you should have to tell us what this branch is we can figure it out we have all these generators you told us what bus it's controlling we can just figure out what the branches are and so that's what simulator does in the powerful so I got to read it first so I'm like don't tell us because you you'll forget so it will just search the topology so for instance I'll go through this in the software in a second but it it isn't quite as easy as just only basing it on who regulates the bus and grouping them together because as I talk to folks several of our customers as like customers I was like do you ever have a situation where you have a bunch of generators here and in a wind farm that are regulating this bus and then another group that are regulating the same bus with different group characteristics kind of behind and I was hoping the answer was no yeah we actually do have one like that now which means over time they'll have more and more so that means we need some way to group these generators into I call voltage control groups so one simulator in a second I'll show you how you group them together so you do have to tell us that but that's okay these can regulate these buses as long as there is some true we just discussed that earlier today so inside simulator actually you guys can open this I do have a little case set up like this for use for you so if you have simulator open you can go to file open case and it's in the cases and there's one called top 12 voltage drink control and it's there's a one in there called droopy PowerPoint so it's the example that goes with my PowerPoint slides and so I have this set up so I've got the generators shown here in orange are assigned to one voltage droop droop control object and the ones in yellow are in another group so if I come up here to the model Explorer and then I'm already there but if you come over here on the left under network down near the bottom there's something cuz something called voltage droop control and I have one group called Group A and another group called Group B and then if I click on the tab here for all generators you can see that there's a column that assigns the generators to which groups they're going so I've got some of them in a and some so that's the only extra intestine sorry I need to go through the droop control but for the generator itself that's the only extra data the regulated bus number also matters in here in order to be in a group together they have to be regulating the same bus so then we need to define what is a voltage droop object so if I'm on the voltage droop control and I'm on the voltage control trolls here let's go look at drew a if I right click on this and choose show dialog this is my little input data for my droop control so I my input data is what is the name of my control and then I'm just specifying these data points here so there's the low of the DB low the DB high and the V high and then to 3 Q values we don't let there be a slope on this one that one's just fixed so this is defining the curve that this group of generators needs to dispatch against hit close here and then there's a there's a second drink control or VDD for drew beep so actually let's go ahead and go over here to the all regulated buses and this is listing all the buses in the system that have this kind of droop control with them and you can see that there are bus bus one has two different drift controls at it that's what's shown right here let me just go ahead and make this image a little bit bigger what this is showing you is the to droop controls a and B so a is in green and B is in blue and they're regulating the same bus so they're both de jizz going to be the same and right now they're solved they're right there on this part of the curve and they're just kind of show you what happens in the power flow now has its all's is where I need more than where I need more than one monitor we can't afford that you guys have them on your desk right - - that's all I have on my desk caroline has like three or four but I get distracted two's enough so what I'm gonna do here is I'm gonna go ahead and go up to tools and hit the little green arrow to have it animate and I'm gonna start messing with the voltage set point at a different generator in the case this one down here so the droop controls over there so as I mess with this generators voltage setpoint then these guys are going to respond to that and we will look at the little red dots and see how they move around and traverse these two curves play here that's right yeah as I reduce the voltage here the voltage is dropping here and they're following those two curves one just to make clear maybe we got weird well eventually to go too far well one of the no one went off that's because the ad limits perfect thanks James okay keep going Oh point nine you get two point nine you here you go yeah so they went off because this guy if we go look at it this is B I think well let's just look at the diagram so what happened there is these generators up here we're trying to pull the voltage up but in this group both generators hit their max bar limit so they can't they can't meet and then if I start going back down do the start back up you'll see them there now they they back back off there all right so it's just a different kind of voltage control so it's instead of instead of trying to regulate the voltage up here for we are replacing this equation so we're not going to talk a ton about the powerful if you haven't done it but in this training inside the normal power flow when you when you have a remotely regulated bus there's an equation somewhere in that giant list of equations that says V reg is equal to the set point all right we just through that equation away and we replaced it with voltage is on this curve somewhere and so it actually was not a huge change in the power flow at first it felt really like it was gonna be but it's no longer TB fuss it's obeying this equation bus instead so you have an option to populate it from the re-themed eventually we could yeah alright so going back to the droop curve again there are some other ways you can do it too you could say that so again this is what it's like a needed customer before I do stuff because when I started to code this I almost didn't include cute men and q-max because to me I was like why wouldn't why wouldn't the generators just keep going until they hit limits because the generators themselves have their own internal limits and Eric Becky's response to me because that's not what the contract says so those he's like no I needed to follow this so in the example we just went through as it was going up there it wasn't able to get there because the individual units hit hit the limit first the generators down there it's not exactly correct because there are bar losses in the feeder network up to there so it's not like the maximum Q I can get here is the sum of those probably lower just probably not a lot of cat banks in here but could be there's also another option where instead of the voltage values being you know absolute voltages you can set and choose to make them Delta's away from the setpoint so I could make V low minus 0.05 I could make the DB low minus 0.01 and then as you change the set points on the generators the curve would just kind of move so that would be an option it's on there as well all right so coming back so the way you do this droop control define its curve and then each generator will refer to what droop control it belongs to and then to emphasize something I already said you do not have to tell us which branches to pick you come back here simulator just goes and figures it out it does a graph search starting at the generators where you're at figures out okay I'm here okay it's this so the actual cue that's being tracked is the summation of the cue coming into the regulated bus from those branches added to any local generator that happens to be at that regulated bus also so the little generators aggregating up there voltages regulate that well it's what's going on is the equation is enforcing the Q that arrives here love question is how do you split it up amongst the generators no how do they know that it's that box o every generator has a regulated bus so it's using the common Rivlin correct yeah the generators have a regulated bus and they have what yeah may have to they have to regulate that bus but not exactly that so like if you had and this kind of stuff I added this from the beginning because someday when people use full topology models sometime before the end of my career it's happened it's happening that's the people he's doing it you guys do it here in your you know data planning alright so if it was you know if your topology was like this if this generator is regulating here the generator is regulating here and they're connected by a low impedance branch simulator is gonna notice that and group those together and say oh I that's that's not what they meant so it'll figure that out as well so what simulator has to figure out is what generators are at the regulated point and what branches are arriving it's like in the code literally I call them the arriving branches what branches are connected to that point and the paths coming from the generators that are doing the regulation so we kind of walk through this the next thing that you have to ask so this does this override the regulation percentage No okay perfect questions that's literally where I was going next oh I'm sorry that still has to occur so in this example I have four generators acting together to regulate that bus that was true even back when I was doing voltage control that it was like how do I split up the VARs amongst all the generators doing the whole coordinated voltage control so to do that you guys have about you have this remote reg percentage that's part of the input data that's in what cases now so that feature of splitting it up is the same as it used to be that hasn't changed the only thing that's different is we're no longer enforcing that voltage equation we're forcing the curve instead so in simulator if you go look at the mismatched table you guys probably don't spend spend a ton of time in here unless things are going badly but this type is telling you what kind of equation are we solving in the power flow and the number of mismatched tables you guys probably don't spend spend a ton of time in here unless things are going badly but this type is telling you what kind of equation are we solving in the power flow and the number of equation types about doubles or triples in version 21 because of this feature and another one I'll talk about next but you'll see here that bus 1 says that it is a PQ and it's a voltage droop rag bust all right so that's telling you that this is a PQ equation but it's kind of special because it's following that curve that's associated with this bus and then the generators over here that are participating in this will say we're on voltage drift control as a remote bus so those are at busses 9 10 and 11 so that's 4 so we get so that's 4 so we get plus 11 and 10 we've had generators there and then bus 9 is over load buses is there a further limit you could put on the terminal voltages what do you mean by like if there was a voltage limit on the terminal that you don't want to exceed the terminals in the generator no there's the second person this training class that Nick was asking about that that's the same job on this gear I need to go think about that some more that what that would require is an outer loop mm-hmm as part of limit checking where you're like all these generators are regulating out here there's an outer loop and then if one of these generators went too high you would basically kick it out of the group and it would become its own little PV bus back locally inside the feeder Network and just set its voltage regulation to whatever limit it was trying to yes and then solve again then you need to come up with additions when it should back off the limit if there was only one I'm not too worried about it the thing I'm really worried about is that your hydro plants you got like 15 units here so one will hit the limit go to moral to the limit will go one will back off we'll go to moral and it was just an oscillating I'm worried about that that you could just fix them at voltage yeah whatever V limited said I'm worried you'll oscillate around but that's a different question so we're not fixing that problem with regulating buses both there's too much but the equation that's being enforced is basically that whatever the droop curve says it should be it equals the mega var mega VARs arriving at that bus added to the generators that are regulating that bus that are going so this replaces what used to be a voltage equation in the power flow so this kind of goes through it and kind of see maybe this is just talking through the numbers I think the picture kind of showed you how how it navigates and keeps it on that curve all right I mentioned this the limits are the same if you sum up the generator limits there's usually as the voltage increases I don't know if that the slants right one of it goes up or goes down but sort of superimposed on this there's some generator limits that we talked about again they they hit those before and the example we just did as we were going up here started going off this way because of the tail limit but the backing off and hitting limits is very similar in the power flow code as it used to be that didn't change in time it used to be you just could kind of check which side of the curve am I on was easy because it was rectangles now I need to like look am I on the blower above that curve so yes we can see that here still got it solving as I keep increasing this this voltage here you'll see these generators keep ramping down and at some point right there just kind of zoom in here you'll see there's a little indication the generators are at there on a br I go up one more and they hit a limit so they're now stuck it away so they're hitting the limit backing off is the same kind of behaviour all right a couple since this is a more technical presentation that what meetings I don't go into the next step because I think it confuses people but I'm here to answer but let's say you implemented this deadband at all the generators what do you think the net result with with so they all have that dance now let's say you give them all this well your your scute you're encouraging them to operate at their whatever q0 is and then as you move away from that to fall off that should work fine there are two you'd be using more to keep the voltage up new people oh I see what you're saying if you gave a big dead band not even big just a dead man yeah how do you think it would change if I mean if they were all I mean they still have a set point or they're kind of moved all around and offset I don't see it being any different than that they had too big of a dead band they'd all be kind of letting the taps Johnstone's you guys feel like the power flow with generators to their limits that's the next presentation that's something different like that might be more stable correct I'm actually the very next presentations this is where we're trying to regulate the queue coming up here there's another the next presentation I'm going to do is talking about changing that voltage to a tolerance where you give it a tolerance right so and that's it's a little bit different it's looking at the cues coming out of the generators instead but you could do this that you can do this on an individual generator that's the setpoint tolerance which is the next presentation how does this work with transformer trying to regulate that class - I mean transformers that are trying to regulate a bus regulated by a generator transformer trying to regulate that bus - I mean transformers that are trying to regulate a bus regulated by generators would never move because the generators have priority in the power Eric faculty in excelsis here set the generators on the crew [Music] correct this is this is all built right into the powerful salvation of the background no it's not what I mean I'm literally taking that equation in the Jacobean that's this and replacing it with the equation of that curve so if there's no there's this is not an outer loop iteration this is in the power flow equations itself so it's much more steeper so if there's no there's this is not an outer loop iteration this is in the power flow equations itself so it's much more stable no you'd have been oscillating oh yeah so but in order to put this in the power flow there are some issues here having having spent the last decade fighting with the network boundary equations and transient stability I've learned that anytime I have a point that looks like this where the derivatives are different at either side of it that to me what danger bells go off and I go I'm going to spend a month and a half because it was powerful if you all saw okay so we don't do that but you don't really notice it but you have to zoom in real close to support so if i zoom in on this that curve there curved it actually it actually cheats and it doesn't actually do what you're asking us to and there's like a little bit of a curve around that point and so if I solve this so if i zoom in on that I'm not actually following the input day that you gave and following this little dotted line so it's like hard-coded stuff in there to kind of calculate an appropriate way to curve around these points so the power flow doesn't beiong and it also means that depending on what kind of curve you give us if you were to give us a curve if you were to give us a curve that says I want to go like this our world would choke on that because we'd be like you just bounce across it so we like there's like a hard-coded limit that these the points have to be no more than point zero zero one per unit apart but thank you nobody has set points that are that accurate the next time your voltage set points and generators aren't nearly as precise as we think they are so there's a minimum so if I come in here and I set my BDD to 1.05 and my BBB hi to 1.06 and DVD lows you may have to go even closer than that there's some hard-coded stuff in here that ignores your input data if it's too tough if you go look in here okay it's gonna curve around yeah well the input data head was those are so close to each other we're just gonna take the average and ignore it you know there's some there's some tweaks in there I mean it's you have to put in stupid in predicative to make this happen but we have to prevent that from getting a vertical line because then over in the Jacobian we have to calculate the derivative of this and it's infinite and that's no good so there's some stuff to prevent that from happening but the idea of this is this is what your wind turbines are doing when they're grouped together like this and regulating a remote point so this was added to to handle them alright anybody have questions do we have this data you have this data on a wind farm of what that curve looks like somebody does I mean it's in the stability day that right there so it's it's kind of here because the dead band and the stability is telling you and then the K is giving you the slope just read the re P see a model yeah read it although you know you might find out that it's those parameters aren't set right yeah we want them to match otherwise you might get Wiggles in your no disturbance exactly yeah but you're gonna be well I'll start with your power value that it want to be different as you would know is the stability you just initialize around the present offer and stability you don't give it the set point you just give it a dead band and we assume you're initialized in the middle of the set point that's how that's how it always works all right mr. Reese they a drew so they just go to max so they are either providing zero bars or max or min now I can't well that's what I mean though they're gonna follow some curve up if you look at this well I'm letting you put those points within ten to the minus one so my hunch is they're you know they're not 0.97 and 0.96 six nine it's like I'm literally letting you put them you could have a point nine six nine that would be okay and then it wouldn't ramp up real quick to go up there but but again it seems like as soon as you cross them all the way to the max immediately I met it just slowly goes down I mean if what you're saying is true and I looked at the output of your wind farms they would either be at zero max or men like 99% of the time correct at zero but when they're not zero they're not always at max Ehrmann okay well then then you give a really tight range there and and you put a big slope in would it be would it depend on the system conditions though I mean even even if it were straight up and down you're wondering what what reactive power is needed from this collection of generators to you know solve it a certain voltage and you may not need all of it so I mean it could be anywhere on that line the other few just need to give it a slope there has to be some so I mean we let you put it really small but it can't be exactly the same otherwise we'll have the America because otherwise you're just bouncing around trying to figure out which curve so we need something a little bit before assault yeah so wind farm ever going to voltage control there holds the point yeah so what yeah so what I'm saying in power world you would need easy on the curve you'd have to say point nine for nine point nine five zero those afeared you like the downside and the upside group is zero on this report that I'm looking at yeah so in power world I'd say are you guys okay being within one one thousandth of a per unit so correct yeah good do witcher's well well that's where I remember problems trying to write that myself 95 correct you go like up here so this is going to handle this and it's gonna kind of smooth it around that corner and you'll be close I mean you're gonna be within the measurement error of your EMS by a factor of 10 right so it's it's still close enough okay and even if it were straight-up down I'm sorry I'm still trying to think about this man so okay and even if it were straight-up down I'm sure I'm still trying to think about this man so if you were if you did a QB up at this bus and you need a certain cue to solve that at certain V and then you're just coming to yeah yeah you're you're some point on that line but it wouldn't have to be at the max I think the vertical lines yeah then you're just guessing which of these curves are which is what the power flow does the traditionally the power flow is am I in a limit or am I not and there's an outer loop where you keep guessing until you get it right or voltage here yeah which is the vertical one in here operating anywhere get ever it needed to stand yeah but at a time it doesn't mean much if it's just corrected and what simulator is gonna do is it's going to go along a line oh ok so yeah you're right that's somewhere between point nine five and yeah so I need an on infinite slope there he's good all right the Laurette the rest of the slides that cover great but actually I tell you there's a lot of error checking that has to go on here so like there are certain certain configurations the user can give us that don't make sense this one's fine I've got the green guys are all in a group together the blue guys are in a group together regulating this bus the green guys are gonna go figure out that my arriving branches are these two there can be more than one and they're in the generators up here that's all fine this is a problem if I've got a group of generators here regulating this bus and another group regulating this bus but there's a path that connects the two groups together alright that doesn't I don't know how any bars that leave here I don't know that they really come here some of them might go around and that the whole logic falls apart to get an error that says I can't do this if you set that up in simulator similarly if you had a generator hanging out in this system that was on voltage control regulating its own bus here that would violate it and it would kick it out saying I can't do it it's like everything that's in this group needs to be on the same page can't be doing something else it is okay if you look back at my example it is okay to have other generators hanging out in these as long as they are not on control so these generators are off ABR or you can have a cap bank you're going to other stuff as long as it's just not playing it's just fixed that's okay this one should be obvious if this generator is in this crew and there's no path that gets me back here literally that can't work so that that will come in as an error this is okay even though it looks weird this is a situation where I've had these generators regulating this bus down here and this generator regulating this bus up here but all these branches through here are essentially breakers simulators okay with this it's going to group everything in this purple region treat it as one blob and they're all regulating the same point together in a in your planning cases we're not going to do stuff like that in an EMS model and we'll see stuff like that all the time inside a switcher it's like who knows what bus they picked to regulate so we need to handle this is the mother of all examples I put together to test the algorithm where you've got all these generators all these busses up here which is all fine there's three you actually have three groups all regulating a bus so you can get kind of crazy topologies and that's just fine and this just goes through the fact that I can't have these corner points that I was talking about I kind of have to smooth around them this goes through all basically I have a white paper out on our website and I have this presentation so that other software vendors can do this too anyway this is one all right and then there's a white paper you want to read a whole bunch of math go read that that's on our website in our knowledgebase there's something in here on renewable generators voltage droop control with deadband so the taper walks through all the mathematics associated with this it was a good entertaining these are called cubic splines it's like how do I fit a cubic curve two to three points so they get a nice curve around the corner

Channel: PowerWorldCorp

Views: 5,280

Keywords: software training, PowerWorld Simulator, power flow

Id: BEiPpKCvGwE

Length: 47min 53sec (2873 seconds)

Published: Wed Mar 11 2020