Transformer Protection Basics

Hello, everyone. Welcome to the EasyPower refresher webinar series. Today's topic is Transform Protection basics. Happy Y'all could join us. And as we are in the habit of doing, we'd like to start the session with a couple poll questions. And very much would appreciate your response. So, the first poll question, How long since coordination was checked or updated at facilities that you're familiar with? If you would respond, no obligation or liability. Strictly just getting some feedback from the audience. You might suspect that to coordination is a topic today. It turns out it's not, but it does intersect with the discussions of Transformer protection. So, it looks like we're close to a quorum. Let's leave this open another ten seconds. Here's how folks have weighed in. Excellent. And then the second question, as your facility experience transformer or other equipment failures due to improper coordination? Again, this is more to the point of the cost of repair. Coordination is primarily and improves the reliability of the facility by allowing the downstream protective devices to trip or react if a remote fault is detected. Part of that scheme is to increase the delay upstream of protective devices. In some cases, those are meant to protect more expensive elements So, you can end up in a quandary between coordination and detecting fairly expensive pieces of the system. Alright? And so folks have weighed in on this. Okay. Fairly wide distribution of responses. Excellent. Thank you. So, we're talking primarily about electrical transformers, and in their failure modes, can include both internal faults and through faults, or what's regarded as through faults. So, we're going to be talking about the types of faults, the protection options, get into a little bit of a discussion of the transformer damage curves and how to utilize them as far as making sure your transformer is being protected as much as possible. And then look at the alternatives. Much of what we're covering today is included in the the buff book also known as I triply standard two forty two, and this is not These are not regulations. Or requirements. There are more guidelines. Hopefully, it's offered as suggestions on how to work with your particular system. So very often, the transformer can be one of the most expensive elements in the system. And since but they are a major system cost, it's important to consider what replacement or delays you will find if that transformer fails. And any way you look at it, loss of service can be costly. Sometimes even an installed spare or specking out a transformer that has more capacity than required, it's warranted, but in any case protection protection scheme is justified and basically you should be looking at the best protection scheme you can afford and make sure you're keeping up with maintenance and monitoring of the transformer itself. Transformers are susceptible to overloads, internal faults through faults. We'll be talking about the transformer damage damage curves But frankly, they're only there to compare against through false. If you have internal false, you've already damaged your transformer. And so the damage curves are really of no value at that point. It's important to understand the level of damage is important, and while it's virtually impossible, to prevent wear and tear, if you will, or damage partial damage to a transformer. That level of damage can make a big difference on what the alternatives are. For instance, winding failure is easier to repair than core damage. The windings can be removed from the tank and repaired physically rewired. As opposed to a core damage, which could be a catastrophic or involving fire. So proper protection can limit the damage but not always prevent the damage. And again, the bottom line and this will come up repeatedly preventive maintenance is important in this whole discussion. And this includes things like oil samples, testing the chemistry, testing the turns ratio, maybe power factor tests, and and and oil temperature tracking. The overloads basically have an effect on the winding over temperature. And the bottom line is hotter the temperature, the shorter the life, because there's some potential for degrading the insulation of the transformer windings themselves. Internal faults can include phase to ground, winding to winding, turn to turn, and loss of oil, in addition to the through faults that we'll be discussing in a bit. Thai temperature shorten the life of winding insulation. The best protection for long term overloading is monitoring or based upon tracking that winding temperature. Now, oil temperature and the bath is relatively easy to measure, as well as dry dry transfer temperature. And so, as we look at the name plate, specifies the maximum allowable winding temperature rise, for instance, a typical value is sixty five degrees c, And this is based upon a presumption that the twenty four hour average temperature range, temperature is thirty degrees Celsius. And so, the way this is utilized is sixty five degree rise on top of the thirty degree ambient gives us ninety five degrees average temperature for the windings, and then allow a ten degree band for the hot spot, and that Determines that hundred and five degree limit is the hottest you want your hot spot ever to be. Now, it won't immediately fail if we Even if we exceed this temperature, but it will have an impact on the transformer life. Now, it is true that the lower the ambient temperature, the the more capacity transformer is capable of. And so frequently, people will have Take advantage of that. Take advantage of the fact that a lower ambient temperature allows more capacity for the transformer itself. That's because the k v a rating is based on maintaining that winding temperature in a safe max mode, based upon the specified average temperature. If the ambient temperature is higher loading loading must be reduced, Whereas, if it's lowered, you may be able to increase it. And so this gives us the ability, in fact, some cases to have a seasonal rating that we can apply to the use of the loading on the transformer. The best overload protection monitors are actually the winding temperatures. Now, that means that we need to have a temperature sensing device. A thermocouple or an RTD and the vicinity of the hot spot. The next best location or ability to test the temperature is the oil temperature. And then transformer protection relays provide utilizing the thermal model of the transformer and this can be based on loading, ambient, and measured temperatures in real time. Internal faults include short circuit and the windings themselves. These can be turn to turn. And ground faults, but those can produce a very low magnitude fault current that may not be picked up by relays. And, whereas, we describe differential relays as the gold standard for detecting false, they're really only viable and used for detecting major faults. So, we're left, especially for faults that occur near to the neutral or in a low magnitude range with the inability to use a differential relay capabilities. So, that gets to surge protection. Transformer windings are the weak link. This is because surges are the cause of surges like lightning or switching transients, can produce spikes that could impact winding to winding voltage differential. So, properly sized and installed surge arrestors can reduce the winding failures. And you want to make sure, and these can be on both the primary and secondary control or connections on the transformer itself. And you want to make sure you locate the surge arresters as close to transformers as possible. Talk a little bit more about transformer differential protection in a bit. But the way the relay works, is that the current n, the sum total of the current n equals the sum total of the current out. And if it doesn't, in that current, it's going some place and it's probably a bad route. As I just mentioned, This won't be sensitive to very low magnitude false, but it does require as far as comparing protection scheme to alternatives, which are primarily fuse, primary fusing, It is required that we have a breaker for the related trip upstream. So, the combination of the breaker and the relay can be fairly significant, but at the higher capacity transformers, it can be a modest investment compared to the cost of losing the transformer itself. At the same time you need high quality c t's that won't saturate for external faults, and the biggest advantage is that they are very fast. The relays themselves are very fast acting. And the placement of the c t's describe the zone of protection. And, at this, in this case, we're including both the main breaker downstream and the transformer primary breaker upstream as part of that zone. And as we discussed in other webinars or discussions about coordination, this zone can be extended by placing the CTs lower in the distribution system, and that will that can determine better coverage on the higher current elements. So Deferential protection is considered the gold standard, but you need CTs in all three phases both in and out. And again, you need to have a relatively costly main voltage medium voltage or high voltage breaker to be able to. Switch to then put current. An alternative then for her very low level currents, detecting low liver failures is measuring the pressure relay or the pressure of the oil itself, and so that element's called a sudden pressure relay that can detect the pressure wave that occurs when oil pressure in the tank is changed. It's more sensitive scheme is more sensitive than differential relays. It can detect turn to turn false, and it can be of a very fast response. Unfortunately, or it can be prone to nuisance stripping, so calibration and proper installation and maintenance is critical. And for conservator type transformers, so those that have intermediate oil source. There's a similar type of relay that responds to the input and output flow to that reservoir. It's called the book hall relay. So, these can be very sensitive and aid in the detection, especially if you're tracking the pressure the oil level and temperature in combination with the relay status. Again, my recommendation would be to make sure you have a copy of the buff book. This is chapter eleven that focuses on relay transformers. But, yes, I'd be happy to share a copy of the slides when we post the video. Through faults are the result of an external fault on the secondary of the transformer. The issue here is the high current that's involved, and we do see a fault in the high, low voltage system. For instance, with a five percent impedance transformer, the effect of fall current can be twenty times typical values. And where we're considering heating and magnetic forces at that high squared effect, then that twenty times factor is multiplied by itself. Which potentially be four hundred times normal current, and the heating and magnetic effects would be that much greater. And the primary fault damage type is overheating of the windings doing due to this long duration current. Mechanical damage on the other hand is due to the magnetic forces which are caused when spikes are high current dumps are placed on a transformer, and the the windings themselves have a tendency to repel each other, other from the core or even from windings to windings. So, in that case, mechanical damage is cumulative. And and can be added up as multiple faults have happened during the life of the transformer. NCC fifty seven has transformer damage curves that define the magnitude and duration. Of through false and describe what the transformer should be able to withstand. And this is in c fifty seven dot one zero nine for liquid transformers and c fifty seven dot twelve dot fifty nine for dry. And what it describes and what you will see when we get to easy power is what I describe as the key points of the milestones in the TCC plot. So, And we'll show this in easy power here in a couple minutes. If I describe time current characteristic curves of a transformer, the tool will show me with a tick mark here at the top the full load current at the voltage we have specified as a reference. Based upon what I've selected in the dialog box for the transformer, it will display the hundred percent damage curve. So this this is a result of the capacity and the voltage an impedance ratio, so it's a fixed value. A hundred percent damage curve is therefore described by the, as I say, the standard c fifty seven, the upper limit, which is the longer time exposure, although it's lower current. Now, those of you that aren't familiar with the t c c plots, We have time on the vertical axis and current on the horizontal axis. So at the lower current values, we can see that the damage curve will impact our transformer health. On longer exposures, whereas mechanical damage will be the result of high current spikes even at very low value of time. So, this is less than a second, and we can see potential exposure to damage if we plot in this case, thirty thousand amps for even two seconds as far as mechanical damage. And then there's this dot at the bottom of the the plot. That's described as n rush, and we will get into more description on this a little bit later in the slides. But the overall goal then is to make sure my my fault current and my load current will remain to the left side of the hundred percent damage curve and any protection that I have for the transformer will not respond to the inrush plot. So, my typical or my best case protective device will have a curve that comes where plots somewhere to the left of the hundred percent damage curve and to the right of the in risk curve. And this is only for primary protection. And, on secondary protection, we don't have to worry about the the inrush. Jonathan asks, of course, sorry. Jonathan says, is there any way to accommodate the cumulative damage on the transformer damage curve? You're reading my mail, Jonathan. Thank you for the question. It's my next slide. Now, as we So we describe c fifty seven dot one zero nine defined in what's called a frequent fall curve. And so, it gives us four different categories of both single and three phase transformers. And then an adjustment that's referred to as frequent fault. And what this amounts to is the ability of the transformer to survive that number of faults in its lifetime that have exceeded the frequent fall limit, but have not exceeded the hundred percent damage limit. So, this is also included in easy power automatically, and It shows up as this plot down at the bottom. So this jagged plot which comes up to be less current It's regarded as a frequent fault curve. So the damage curve is, again, for through false, and we have the ability to select frequent versus infrequent. So, frequent fault is this jag down at the bottom, then frequent is the extension of the hundred percent damage curve. So, if I design my system or my protection scheme, to avoid entering this area of the damage curve, I've got a more conservative protection scheme and if I disregard that, and I just potentially allow the system to routinely be affected by false, only with reference to the hundred percent damage curve. So let's see if I can show what this means. Zero one They call these z curves because they look like kind of a z excused. Or stretched out. And so, again, this is handled in easy power as I dialogue, the transformer parameters. So so this is the transformer for this bus two, and we're showing under our t c c plots, We're plotting a hundred percent withstand curve, the unbalanced, a derating curve that we'll talk about in a little bit. And then the frequent fault curves. So, I wanna see what those look like. If we go to coordination, pick out a single transformer and plot it. The t c c curves, we see the result of the two plots, I'm disregarding in rush for a minute. The dotted line, hopefully you can see it. The dotted line on the right is a hundred percent damage curve. And if I open the dialog box to the TCC plot, and I take off the hundred percent damage curve, that line disappears. I'm gonna put it back on, and then take off the frequent fault curve. Now, the second line that we're we're talking about is the unbalanced derating curve, and we'll talk about that shortly. So I'm gonna take it off We start with a hundred percent damage curve, and then use the frequent fault curve to do our coordination. I want us to compare all three of these. Just real briefly to show you what what to expect. I've saved a t c c, I'm comparing the three plots for these three different transformers. Cat one, as we noticed from our chart from from c fifty seven, does not have a frequent fault category. So, we design protection scheme against a hundred percent damage curve. For category two, you'll see that shows a seventy five percent of maximum value as a frequent fall curve. For categories three and four, which are the very large power transformers. We To be conservative, we try to stay under the sixty percent frequent fall curve. So, this is all handled by AZ Power automatically. Alright. Oh, I'm just gonna show the reference full detail. So Right now, if we look at the plots, the scale, we're showing this as current times a hundred, at the four eighty volt range. Easy power selected that because that's the lowest voltage in our system. And usually that'll be the higher current. And consequently, we see in a category one, The current range goes from Looks like Four thousand amps down to three hundred amps, ish, Oh, as I'm moving the cursor around, you can see at the bottom of the screen, the coordinates of the cursor location. So, if I place this on the upper end of the damage curve for the category one, I see my range is 4zero amps, at four hundred and twenty six seconds. Whereas, it tops out at three thousand six hundred amps at two, that's just over two seconds. If I wanna know what this is in primary voltage, our primary voltage scale, I can right click on the bus above set it as a reference. Now, we're talking lower current. Here, we're talking a hundred thirty six amps at two seconds. And fourteen amps at four hundred and forty seven seconds. And then I can look at the full load current, the cursor is at five point two six amps, so that's the reference point. That's used to calculate the loading and the protection scheme between the full load current and the hundred percent damage curve. So, the first slides I had as far as the damages, what we can see if I exceed my fault current exceeds this upper edge, then I potentially have insulation damage if I let that condition run for an extended period of time, whereas if I have spikes that exceed this limit for a short period of time, I can get mechanical damage in the transformer itself. And as we just described, those are hard to to pick up. Those those types of damages are hard to pick up with my current sensing relays and fuses. And so, that's why other schemes as far as the protection have to be considered for transformers. Now, there's a This is interesting and it's it's nontrivial and it's also non intuitive. But, basically, there's a difference based upon the configuration of the transformers, typically in a delta y configuration. The per unit seen by the primary protection. Others, if we have face protection, face current protection on the primary, it doesn't see something necessarily proportional to the secondary side. So, if I have fault current on the secondary side and I'm expecting the primary protection deal with it, I have to be aware of this non conventional consideration. And again, winding damage on through false is a function of through current on any one winding. So, we must adjust the damage curve to take into account that's difference in per unit current seeing between the primary and the secondary. If you look at the way delta system is connected. There's no ground connection on the primary side. And on the y side, very often, the neutral connection can be consider or is grounded. The point that I'm trying to make is if we measure the current through the transformer on the phase downstream on the secondary, we can measure one per unit current, but it corresponds to the phase on the upstream. But in the case of a primary, I mean, a secondary fault on the secondary side, that one per unit current is shared by two phases on the primary. So, again, Currents for single line to ground faults on the y side, where we have our protection on the primary side. So, no, it's not nonlinear characteristics of the iron. This is strictly phase, winding shift. So, there's a thirty degree shift just automatically in current versus voltage realm, but in the magnitude of the current For a line to ground fault downstream, we need to adjust the fact that on the primary side, we're only looking at fifty eight percent of what that current is. So the point is we need to have a more sensitive protective device if we're using primary protection. And that's taken into account by offsetting that damage curve when seen by a primary device, or primary protection. Now, again, this applies only for ground faults on the secondary and it applies only for primary protection. And again, EC Power does this automatically and that's why I'm kind of hammering on this. Let me figure out which one I'm looking for. Ground current t c c. And and it's important to understand this is only for primary protection. So let's kinda not save this. Okay. And so Let's just look at this. If I plot this curve for this transformer, I see on the t c c plot, two curves. As I mentioned before, That second curve is the unbalanced derating curve. If I take it off, left with a hundred percent damage curve. So, the the message is, if I have primary protection, There's if my protection is a fuse and a primary side, then I want to make sure it responds before it gets to this shifted curve. Now, let's see. Let me go back and compare these three, because it it'll change based upon how I've configured my ground Notice here, I have high resistance ground, hard ground, and no ground. So let's see how I've done that. Let's look at the high resistance, excuse me, the hard ground first. That's just what we were seeing. And so, if I have primary protection, I need to make sure it stays clear. It protects me from currents that potentially exceed this shifted curve. Now, if I have secondary protection, I could disregard that curve. And Make sure that I'm not exceeding the hundred percent damage curve or the frequent fall curve. If we look at the hard ground, we just did that. Let's look at a high resistance ground, and this kind of give you a clue or suggestion. It doesn't show that second curve, but it does show the current through the fuse, This is another point I didn't emphasize. When I'm plotting a faulted bus, The tool shows me a tick mark through any protective device that's sensing that current. So, in this case, we have a fuse here in the primary the fault on the secondary because of the resistor, I have a hundred amps limitation on the amount of ground current I can produce. But if I look at the current in the primary, that's only two amps. And that's way down here on my scale. And so, there's no chance that that's gonna trip a primary protection anytime. So that's just a fact. It says, I'll I'll kind of reemphasize that here in a second. And then if we look at a no ground situation where we haven't. Yeah, ungrounded delta to an ungrounded why. Again, there's no difference between ground current and and I'm looking at a line to ground fault, and there's no ground current as a result of the system having no ground at that point. Okay. Again, easy part takes this into account, and and handles it pretty nicely. So, the takeaways here is if I'm if I'm focused on primary protection, need to take into account the situation on my grounding transformer, specifically for a line of ground fault. Now, for a three phase fault, it doesn't matter because now we have full current going through the fuse. And we need to make sure we stay clear of that hundred percent curve. Okay. Now, and so so that's gives us this fifty eight percent adjustment for delta y. There's also an eighty five percent shift on a delta delta. And, again, this is handled automatically by EasyPower. Now, in the Nashville electrical code, table four fifty dash three, it shows the max settings for transformer overcurrent protection as a percentage of full load current. So, based upon the location, whether it's any location or supervised. We have systems primarily that are monitored and operated by qualified personnel that's considered a supervised installation. And we have the ability to look at primary voltages over a thousand volts and secondary voltages over and under a thousand. Then whether we're not using breaker on the primary fuse on the primary and breakers fuses on the secondary. What we see is potential for going to six hundred percent of the full load current, six times the full load current, out with a breaker and the primary. That doesn't mean we have to go to six hundred percent because we're cutting it pretty close as far as the damage curve. So, let's look at my three hundred percent fuse. This is this is where the rubber meets the road as far as. As far as what we choose to do for generator for transformer protection. Skip rid of this. Go to the fuse. So according to the chart, if we have a thousand volts, a thousand k v a, and thirteen point eight volts on the primary and four eighty volts, we could have three hundred percent value of fuse for protection. And so if we go to coordination, Let's plot the fuse and transformer. Okay. So we're looking at three phase fault. The straight lines here on the right and the cable damage occurs. So, again, this, we don't want our current time plot to exceed this rating for the cable, And yet, if we look at our fuse protection in the primary, this is rated Fifty amps, and our full load current, Four eighty volt range is seven twenty five. So, let's I'm gonna reference this bus I wanna look down here at my coordinates, and set this over the full load plot or tick mark, I see the full load current and the primary is forty one amps. So three hundred percent times that would be a hundred and twenty amps, So fifty e looks like it's about a hundred and twenty five e for as far as full current. The point I wanna make is, you can see how inadequate the fuse is for protection because there's some area here that I don't have protection from my single line to ground. If we look at a single line to ground fall, I guess I can fall down here. So I'm seeing two hundred and seventy six amps in the primary, and and that current here at app cycle, and here at five cycle, thirty cycles, This shows protection for my frequent fault, potentially, but it doesn't have protection for this lower level long term Correct. So that's one of the reasons it fuses are or not that acceptable or at least to have some considerations that you need to make sure your circuit covers this part. Again, the fuse is in the primary, and so it's doing a not quite adequate job at protecting the transformer. From both the single line to ground faults and and the frequent fault curves. Okay. Now, trying to cap on fuses, I'm just showing how to determine that effective coverage. Both for a three phase fault and a single phase fault. Now, again, we're not violating the The recommendation table will still stand under the three hundred percent, but it's not adequate enough for the purposes of this particular design. Okay. Question says, using the table Is it possible that we oversize protection as a result to get high arc flash values? And that's another great point is that The problem another problem with fuses is that it's not gonna be an adequate protection for art flash. So if we look at default on this particular bus, Because we have I mean, this is not bad, but what's also not necessarily good protection. So this is just a short cable. In fact, my paradox example, This is limiting to sixteen thousand amps, so that's pretty high. And, it's potentially gonna exceed the damage on the the cable itself. So, this table should not really be considered for arc flash protection. This is strictly the restrictions for transformer protection and it may not be adequate to do coordination downstream. Okay. So much for that. Just a real quick rehash of the inrush. When we're considering primary protection, whatever we have, whether it's a breaker or a fuse, we need to stay to the right of the in rush plot. So here we have roughly eight to ten times full load current for a tenth of a second. And this will be based on the self cooled rating and we see that either of these devices presumably in the primary side will be affected by the inrush. Most the problem is it's the inrush current is very very much harmonic included for harmonic currents and it's difficult to know their response on a protective device. Okay. We noticed the t c c plots we can control what we're showing on the t c c plot itself, and you can take these off the plot especially the d rating curve if you're not looking at single line to ground primary protection. So, questions, how to improve transformer protection. Transformers, again, can be very expensive and catastrophic failures if we haven't designed us a procedure to take care of. Potential damage. So, for better protection, use breakers, not fuses, breakers include relays, and breaker combinations for transformer protection against over current, differential protection is the gold standard. And then, for incipient, false, very small magnitude pressure detection and or tracking gas chemistry can be a more effective procedure. Now, we describe the delta y impact on primary protection. But there's others and there's other schemes that are used in this industry. These e z power takes care of automatically, but there's something to take into account based upon what your particular system is doing. Now, I haven't talked about resistance grounding other than that one example. The most common fault in the industry is aligned to ground fault. And the issue is, if it's not properly dealt with, it can be high enough current and potentially have an arc that can migrate to a full blown three face fault And so, it behoves system designers to pay attention to the possibility of limiting the amount of current in a single line to ground fault. One of the ways to do that is with high resistance ground. And so, as I showed in that one example, resistance grounding can be accommodated in multiple ways. But again, to determine or or remove the potential for grounding to be a problem, you should have all your sources in the same mode. Of limiting maximum ground current. Degree cap, it's difficult to provide full, through fall protection with only primary fuses. If you compare expulsive fuses with current limiting fuses, the current limiting fuses has more of a vertical slope to it, this gold band over here on the right. Is a current limiting fuse. And trying to use it in a primary protection mode, we can get frequent fault coverage, but we don't have coverage at the higher time delay at lower currents, whereas expulsion fuse can be a little bit more accommodating as far as the slope. And this one's on the secondary side. And can be worked in conjunction with a upstream breaker. Primary breaker is required for more advanced protection, it also helps with downstream arc protection. Multifunction, transformer relays provide good transformer protection, but you really need a backup because you don't wanna have all your eggs in one basket. When it comes to protecting the transformer that has such high value. So, here's a typical protection scheme that's offered a multifunction relay that can include differential, ground, phase over current, and then a backup relay, again, maybe lower capabilities looking only at overcurrent. But the ability to cover multiple loc multiple types of faults within the zone and and then there's as far as comparing this against a fuse, it could be considerably more investment but better coverage. More investment because the relay itself can be expensive. The CTs have to be selected, and then meeting voltage breaker as an investment. And just one more quick rehash on the n rush because there was a question there. Transformer protection must ride through the end rush. In other words, if I have primary protection for my transformer, it can't be allowed to be sensitive. You don't want it to be sensitive to the inrush plot, which generally is at the tenth of the second range. Here's a good question from Jesus. Using the four fifty-three chart, Is it possible that we oversize protection as a result of a high flash high arc flash values? So, I believe what he's referring to is the ability to oversize or include additional capacity in the transformer, so I can fall within the damage curves with my primary protection. And that's a good way to In other words, the lower the draw on the transformer, the better we've extended the life. So, temperature and time both affect the capacity utilization. And so the answer is, yeah, potentially, that's one way to get there. But you do end up with higher current potential on a fault because of the extra capacity of a transformer. So it's it's a Something you have to pay for at some point in the design, but overall, the the ability to extend the life of the transformer and reduce the service aspect can be substantial. Okay. I think that covers what I meant to cover today. Will include a copy of the slides when we post the video. Thank you everyone for joining us. Be sure to check out the new webinars we have scheduled, and we're doing live virtual training and act actually, this year we'll have a live in person trading in July. So check out the website to to get updates on that. Thank you for joining us. We'll talk to you later.