FAQ - 1 Hz Testing: A New Dawn in PF Testing

FAQ's from this webinar session

 

1 Hz Testing: A New Dawn in PF Testing Q&A

 

This is the transcript of the Q&A session from our webinar 1 Hz Testing: A New Dawn in PF Testing held on 9/22/2023. If you would like to watch the presentation leading up to the session below please click here.

 

With LF Power Factor testing, we have known 1% or 0.5% maximums due to historical data. Would 1Hz still be based on assumptions or is there a defined maximum limit? Or, can it only be compared to LF similar to tip-up testing?

"The question asked is absolutely correct. The 1% and 0.5% limits have been used for many, many years. And what Megger has done, and through some of our partners, is worked to develop the limits at 1Hz. In particular on the transformer side, there was a study done in a publication in Transformer magazines where, in particular, they overloaded transformers and they looked at how the power factor changed across many different frequencies. Based on what they saw, they were able to say that a 2% limit at 1Hz is what they would look for to indicate that insulation has started to degrade inside of transformers. On the other side, when you consider bushings, we had access to over 1500 DFR results through the IDAX equipment. And based on those 1500 results, we looked at bushings that had been deemed failures, bushings that were marginal, and bushings that were acceptable. And that is how we derived that 1.2% limit proportions at 1Hz. So it's not just a shot in the dark. It is something that we have investigated and put a lot of effort into developing limits that are applicable and informative for the user. I will say that most of the work that Megger has done has been around oil impregnated paper and there is still quite a bit of investigation to do with respect to dry type transformers, resin impregnated bushings, things of that nature.  So the work continues for us to drive limits there." - Ken Petroff

What does it mean if you have lower PF at 1Hz?

"Coming to this question, when we talk about power factor testing, I want to kind of give some background. When you talk about power factor testing, you are talking about the losses taking place in the insulation. And in one of the slides, Yash mentioned why we do the testing at 1Hz; because you have better visibility of the losses taking place in the insulation. So typically, when you compare the results that you would get here at your line frequency, which is 60Hz, and you compare the results that come at 1Hz, typically the numbers would be higher at 1Hz because you are just using a magnifying glass to look at the losses taking place in the insulation. And 1Hz would give you a better picture of the sense of the losses in the insulation. So the majority of the time, the 1Hz power factor will be higher than the 60Hz power factor. In very rare occurrences, like we have not seen that many instances where the 1Hz power factor is lower than the 60Hz power factor- In case you run into that, I have two recommendations. One is that I want to make sure that 1Hz data is also correct to 20°C, because when you are in the in the lower temperature range, temperature can play a big role. So I want to make sure that when I'm analyzing 1Hz data, it is correct to 20°C. Okay, so that's one thing that I would look at. And the second thing is, let’s say I corrected 20°C, and 1Hz data is still less than that, then I want to see how much difference there is from 60Hz to 1Hz. If it is within range, for example, let's say the 60Hz power factor is, let's say 0.5, and 1Hz power factor is let's say 0.4, then I wouldn't be much concerned about it, because what you are really looking at is that is it staying flat or is it going exponentially high as you go from 60Hz to 1Hz? Because within our DELTA4000, not only do you get 1Hz results, but you also get a narrowband DFR graph where you can see the follow up of the graph as you go from 500Hz all the way to 1Hz, you can see the trending of the graph. And I don't want to see the graph going exponentially high as I go lower in the frequency. So as I said, as long as the curve is staying flat, I have nothing to worry about. I can see that these are valid results. These are good transformers because the losses are not increasing as you look at the 1Hz data. So those two things I would look at, I would correct 20°C and then I would, rather than focusing on 1Hz value by itself when the number is low,  I would look at the trend of the narrowband DFR graph." - Dinesh Chhajer

How does IDAX measure moisture in oil and paper based on frequency?

"This presentation was mostly on 1Hz testing, but there was one slide that Yash mentioned or showed, and it showed the full DFR trades. And what it is, is that when you do the full spectrum of dielectric frequency response, which is what IDAX does. It goes all the way from 1000Hz, all the way to like 1mHz to 2mHz, depending upon the insulation temperature. And in the trace that you get, which is power factor versus frequency, there are the higher frequencies and lower frequencies which kind of gives you the response of the paper, and the middle frequency gives you the response of oil. Okay, now I have got the trace of a given transformer. How do I find out the moisture content present in the paper insulation? So just like how you have done many things in your in your studies where you have an unknown quantity, and you want to compare with a set of known quantities, and whichever matches with the known quantity, that's the value of that curve. So, what we have is that in the IDAX we have a library of multiple different curves which represent different moisture concentrations in the paper insulation, and different conductivity of the oil. So we have a library of curves, and with the curve that we got from this transformer, we can do a curve fitting technique where we are trying to measure, or trying to compare, the unknown transformer curve with the library of known curves, with the known percentage of moisture concentration and oil conductivity, all those kind of things. And then we use the curve fitting technique to see which curve would fit the best with the unknown curve, once there is a curve fitting technique done, then we know exactly what the moisture concentration is for that given transformer because now we have the known variable that we are comparing against. So thats how IDAX does the does the estimation of the moisture concentration of paper insulation. So, apart from this scope fitting technique, you can go in and read- there are tons of people on the IEEE website where it talks about the frequency dependence of power factor. So we know that as you vary the frequency your power factor would vary, right? And that's kind of the whole basis of this DFR concept where I'm wearing the frequency such that some frequency would highlight the response of paper insulation and some other frequency would highlight the response of the oil insulation. And once I can separate those out, I use the curve fitting technique to find out the moisture present in the paper insulation." - Dinesh Chhajer

What's the difference between TCT and ITC? Additionally, what does ITC stand for?

"TCT is temperature correction tables that use an average of all the readings collected in years and years of data, and then you get, based on those old readings, an average correction factor as to what should be the multiplication factor if im testing at, for example, 18°C, 19°C, you have your corresponding temperature correction factor from that table. Whereas ITC, that is individual temperature correction that is based on frequency sweep. You analyze the response from a certain material, you’ve tested different frequencies and you get a correction factor using Arrhenius equation. That's how you get the correction factor from evaluating that equation. So that's the difference between TCT and ITC. ITC specifically evaluates one particular dielectric and its response is based on that specific insulation material, whereas TCT takes an average of all of the assets or different dielectric insulation materials and then gives you a correction factor." - Yash Godhwani

Why use temperature correction tables when you can use ITC?

"When we talk about temperature correction factor tables, Yash gave an example of comparing your test results or your bloodwork results with the last 15 people who came to the doctor’s office, and taking the average of those 15 people and using your blood work and comparing your blood work level with those last 15 people; that may not be the best way to assess your results, because the profile of those 15 people who came to the doctor's office could be drastically different. 

We don't know. They're they're eating habits, their exercise, and all those kinds of things. So what temperature correction factor tables have been generated, as Ken pointed out, it is a combination of the values we got from OEM’s data submitted by users on the transformers that they have in the fleet and those kind of things. And if you go and look at the temperature correction factor tables, you would be so surprised that it treats the transformer that was built in 1960 and the transformer that is built in 2023, it treats both them the same way. We all know that a transformer, that 60 year old and a transformer that's brand new would have a completely different insulation condition right then. The other thing is that it treats a transformer that's of the size of half MVA and a transformer that the size of 500 MVA- It treats both the transformers the same. Not only that, if a transformer that's in the middle East, right where the temperatures are extremely high, the ambient temperature- and the transformer that's sitting in Canada, it again treats both the transformers the same way. It uses the same correction factors to bring the value from 30°C to 20°C or from 10°C to 20°C. With these kind of obvious this glaring kind of problems, it is very obvious that when you're going to use those correction factor tables, your compensation values may not be accurate. And that's the reason why IEEE standard c57.12.90, they used to have correction factored tables in the previous standard and in the in the standard that came out, I think somewhere in 2015 or 2013 or somewhere in that range, they got rid the correction factor tables and now there is a section there saying that the reason why they took the correction factor tables out is that no single correction factor the curve can fit all the cases. It's from IEEE standard, and that's the reason why we don't recommend using the temperature correction factor tables. And ITC obviously. Yes, Yash explained very well what ITC is then. The one other important thing I want to mention is that in our software we still offer the correction factor tables and ITC. And the reason being that a lot of customers, even though they are convinced that they are wrong, they want to maintain the trending capability. So some people, what they do is, they would compare their last reading with today's reading and they would use temperature collection factor table to make sure that, “how off was I?  I'm consistently off even in today's reading”.  And from this point onwards they start using ITC. So that's something, the reason why we offer the correction factor tables and ITC so that people can trend with their previous reading. But moving forward, they would use ITC to get the accurate temperature compensation." - Dinesh Chhajer

Can we do this test with the TRAX220?

"Yes, you can do Power Factor test with the TRAX220, but you would need a TDX box to go with it. That is basically the high voltage box, and then you can do it. It has the narrowband DFR from 1Hz all the way to 505Hz. So you have that much of a frequency band to test with. Yeah, you can do this test with that." - Yash Godhwani

Can I use 1Hz on dry type transformers?

"So as I mentioned earlier, the ITC direction algorithm that has been developed, we primarily based that on OIP insulation. I do believe that you can do this ITC for dry type insulation as well. However, it is some research and investigation that Megger is continuing to seek out. If you go in to PowerDB today and you select dry type transformer, you will not have the option to perform ITC.  But if you were to leave that as oil, the software doesn't know any better, then it would attempt and apply ITC. And I have seen cases where ITC, you know, testing a transformer that just came out of service, and then waiting for it to cool down to around 20°C, I have seen ITC provide reliable correction back down to 20°C. But I would say that Megger as a whole, the industry has not done enough work to validate and make sure that the same activation energy that's used for OIP, applies also to the dry type transformers. I will say that you can use 1Hz, and 1Hz is an excellent tool for looking at sister assets, see if inflation is starting to change. The thing about dry type transformer, since it doesn't have any oil inside with the paper, you will see elevated values at 1Hz compared to what you would see with OIP. It doesn’t necessarily mean that it's bad. 1Hz ITC drytype transformers, Megger continues to investigate, and we look forward to providing some more concrete answers to these questions as we move forward." - Ken Petroff

How big of a role does temperature play on insulation condition assessments?

"That's a very interesting topic because temperature is a big, big component which would affect the health of the insulation. You know, there was a presentation that I believe I attended, this was some time back where they talked about four factors that would take the the life out of the transformer. And those four factors were temperature, second one was moisture, third one was force- whenever you have a short circuit, obviously there's a tremendous force acting on the windings, and the fourth was oxygen. And that presentation kind of sums up this question quite a bit. That temperature is big, big player when it comes to the deterioration of the insulation condition. That's the reason why we try to talk about hot spot temperature, temperature rise, and those kinds of things, because we want to operate a transformer within the acceptable range based upon how it was designed. So when you're doing the power factor testing, it is equally important for you to know what effect does the temperature play in the insulation condition assessment. If you perform the test at 5°C compared to when the transformer, let's say is at 40°C, the power factor readings would be different. And the reason why those would be different is because, and again, I'm going to go a little bit deeper into this topic, that 95 to 99% of the moisture resides in the paper insulation and only the remaining handful of maybe 1 to 5% reside in the insulating liquid. And what happens is that as the temperature changes, as you go higher in temperature, that moisture would come from the paper insulation migrating into oil insulation. And obviously when you have a higher concentration of moisture in the oil insulation, your power factor values would jump up as opposed to the moisture being absorbed in the paper insulation. So when you run the tests at 5°C or 40°C, your power factor values would come out of be different. That's one part. And then the second part is that obviously a deteriorated insulation- an aged insulation, would react differently at different temperature levels. A simple example that I can give is that a person who is a 20 year old and a person who's a 70 year old may not be able to run at the same speed. The same thing happens with the insulation as well. So a transformer that's age insulation is deteriorated, the effect of temperature on that insulation would be drastically higher than on a transformer that's brand new. And for those reasons, it's extremely important for us to take into account the temperature whenever we are evaluating the power factor results. And that's where the temperature connection factor table and its discussion with the ITC come into the picture, and you can have the best instrument in the world and you can get the most accurate reading, but if you're not correcting it accurately to 20°C, you are always going to be in doubt, whether I'm comparing the right numbers on or not. It's extremely, extremely important to correct the reading to 20°C in the most accurate way possible. So that's what the role of temperature is in the power factor assessment." - Dinesh Chhajer

Do I need to assess both 1Hz data and 60Hz data together?

"So yes, just real quick to Dinesh’s point, since temperature is so important in the power factor correction, if you want to get the most accurate temperature measurement, you know, obviously having an optical sensor is very good. But if you just look off the oil gauges, sometimes everyone can go up and tap on them, they can change. So if you have time to do a winding resistance measurement, then compare that to factory results, you can then correct that winding resistance to a temperature to get a very accurate temperature of the winding. 

Now, to the question about 60Hz and 1Hz together: what we are recommending is that you continue to do your 10kV, 60Hz power factor tests. It's what's been the standard for decades. We do not want you to go away from that. We do recommend you using ITC, though, to get that temperature correction that is accurate. Now, using the same connections and the same software, you would then do the 1Hz measurement at the same time, and that's going to help improve your assessment of the insulation. As Yash mentioned in the presentation, if your power factor is good and your 1Hz is good, you can be confident that your insulation is in good health. If you get a poor power factor measurement at 60Hz, 1Hz will confirm that degradation. And then lastly, 1Hz can be used to identify problems that will show up before it shows up at 60Hz. So that is sort of the strategy that you look at 60 Hz and 1Hz. We're not trying to change what you're doing today, we're just trying to add to that to improve your maintenance prediction on your assets." - Ken Petroff

Interested? Click here to watch the full webinar.