WEBVTT 00:00.000 --> 00:11.760 Well, welcome everybody. Thank you all for coming. Today we're going to be giving a presentation 00:11.760 --> 00:19.400 about open leader RS about demand response and about open area, like we just said. So open 00:19.400 --> 00:23.760 area is a protocol and open source protocol. It stands for open automated demand response. 00:23.760 --> 00:27.400 We're going to get into what that is exactly in the next 50 minutes. We're going to be giving 00:27.400 --> 00:33.000 you some background about the open leader RS project, a primer on demand response. We're 00:33.000 --> 00:38.160 going to be looking at a Dutch adoption of open area, a Dutch use case and we'll be looking 00:38.160 --> 00:43.000 at future developments. We don't have too much time. So if you have questions, we'd love 00:43.000 --> 00:48.400 to answer them, but preferably at the end. But first let us introduce ourselves. 00:49.400 --> 00:54.880 Hi, I'm Max. I'm working for the Dutch software consultancy 00:54.880 --> 01:02.400 critical, specialized in rust and the tech lead of the open leader RS implementation. 01:02.400 --> 01:06.800 And I'm staying on howling and I am a specification writer and software developer at 01:06.800 --> 01:12.000 ALADNL. We're going to be talking a bit about ALADL later on and I would call myself 01:12.000 --> 01:17.000 a power user of open leader RS. 01:17.000 --> 01:21.840 Right. Before we start to talk about demand response, we need to get into grid congestion 01:21.840 --> 01:27.200 a bit. Because this has a lot to do with demand response, we're looking at the grid congestion 01:27.200 --> 01:31.840 map of the Netherlands as you can see it's quite red and red is not good. It means that 01:31.840 --> 01:36.560 there's a capacity shortage. So essentially almost no new connections can be made, existing 01:36.560 --> 01:42.920 runs can't be expanded. And is this a big impact on society? Housing projects are being 01:42.920 --> 01:47.520 put and hold. Businesses can't expand. So the government is looking for ways to solve this 01:47.520 --> 01:52.120 problem. We could just build more transformers, but as you can imagine, this takes time, 01:52.120 --> 01:56.840 it costs money. So we're looking at other solutions to fix this. And it might seem 01:56.840 --> 02:01.880 there's no capacity left at all, but this map doesn't really tell the whole story. 02:01.880 --> 02:05.960 Because if you look at a day-to-day basis, we see something like this. We see a waveform 02:06.000 --> 02:12.040 with two peaks and values all besides them. And this kind of means that there's two peaks 02:12.040 --> 02:15.960 in which lots of people use electricity and in between we've got a lot of space on the grid 02:15.960 --> 02:21.200 left. So if you can make better use of them, we still have some capacity. And we can 02:21.200 --> 02:26.400 do that with demand response programs. demand response is the idea that demands, so customer 02:26.400 --> 02:32.760 appliances like washing machines, heat pumps, what have you can respond to the status of the 02:32.760 --> 02:38.200 grid? So how busy is it on the grid? And demand response program is a grid operator 02:38.200 --> 02:44.200 communicating the sending signals and communicating the status of the grid to consumers. 02:44.200 --> 02:50.640 So you at home or companies, so they can change the way to use electricity. This can 02:50.640 --> 02:55.560 come into form of price signals or capacity limits, but the goal is to limit use during 02:55.560 --> 03:01.720 peak times either way. This benefits the year so because there's more capacity, so they 03:01.720 --> 03:06.680 can get people on the network faster or they have to build less transformers. But as a 03:06.680 --> 03:10.800 customer, it's also nice because electricity in those values is most of the time cheaper 03:10.800 --> 03:13.360 and a lot cleaner. 03:13.360 --> 03:17.680 How does work exist at DSO predicts a peak? They send a signal to the customer and then 03:17.680 --> 03:22.400 a customer appliance preferably automatically shuts down or it shifts its use to another 03:22.400 --> 03:28.320 time of the day. And that works exactly with open areas going to be explained by max. 03:28.320 --> 03:35.280 Yeah, so let's have a quick look and at what open ADR is now, open ADR is a standard to 03:35.280 --> 03:41.160 solve the problems that's time just laid out and it stands for open automated demand response. 03:41.160 --> 03:48.240 It originated back in 2009 from the Lawrence Berkeley National Laboratory and has quickly 03:48.240 --> 03:55.600 been taken over by the newly founded Open ADR Alliance that still governs the spec maintenance 03:55.600 --> 04:01.600 and development. And for us, most interesting are actually the versions 3.0 in 3.1 that 04:01.600 --> 04:11.680 have been released 24, 25, respectively. Because what mainly changed from the 2.0 versions 04:11.680 --> 04:18.760 there is that they switched from XML and XMPP to JSON and HTTP, which makes development a lot 04:18.840 --> 04:26.480 easier as many can imagine, I guess. In general, open ADR is a very high level messaging 04:26.480 --> 04:31.240 protocol, a very generic one that can be used for many use cases and supports many use 04:31.240 --> 04:37.480 cases. And because of that, to have a very specific use case, you typically need to have 04:37.480 --> 04:42.720 an additional so-called profile, which further specifies how exactly to use these messages 04:42.720 --> 04:47.960 and how exactly to behave. And one such profile is the grid where charging that a lot 04:47.960 --> 04:54.200 now is developing and maintaining. And I guess time will talk about that later a bit more 04:54.200 --> 05:01.480 in detail. So how does that work on a high level? We do have two parties involved here, 05:01.480 --> 05:08.440 generally, that is the distribution system operator, basically the grid operator. And in 05:08.440 --> 05:15.480 the case of the grid where charging, we have the charge point operators. In the Open ADR 05:15.560 --> 05:22.120 speech, they run a business logic and a customer logic. And they have so-called virtual 05:22.120 --> 05:27.800 top note, I would just call it a server and a virtual end note, basically the clients of that 05:27.800 --> 05:35.160 whole thing. And communication, at least for the basic use case, happens over two message types 05:35.160 --> 05:44.280 that are so-called events and reports. So the virtual top note, the VTN will send out events 05:45.240 --> 05:52.360 to the customer logic, the customer logic will somehow react to that and it may or may not send 05:52.360 --> 06:00.680 back reports that, for example, tell the DSO on how much energy they have used in the last hour 06:00.680 --> 06:07.080 or what they possibly predict to use in the future, maybe. All these are possible use cases, 06:07.080 --> 06:17.400 how you could use that. So our implementation, open leader RS, is then an implementation of 06:17.400 --> 06:24.600 that standard Open ADR. And that's joined effort by Elatinel and Tfeirhov. And it was born out of 06:24.600 --> 06:30.440 the desire from Tfeirhov to add or to use our expertise that we already gained in other internet 06:30.440 --> 06:36.200 protocols like the network time protocol to create a positive impact into the energy domain as well. 06:36.200 --> 06:43.240 And that way, our ways we're crossing with Elat, that we're also planning to work with Open 06:43.240 --> 06:52.680 ADR. And so we joined our forces to get open leader up and running. And then we also 06:52.680 --> 07:02.280 sponsor this to the Elatinel G organization that now is housing this whole project. So what does 07:02.280 --> 07:09.240 that actually, what what can open leader do? We do have a well-tested Open ADR 3.0 implementation. 07:09.240 --> 07:15.080 All right, that's used in production. It's time we'll talk about that in a second. We just 07:15.080 --> 07:21.960 released a better version of the Open ADR 3.1 implementation. And we are currently 07:22.600 --> 07:30.200 implementing a so-called subscription feature which basically lets you send real-time signals for the 07:31.160 --> 07:38.520 grid. We are working on a CLI that makes debugging and prototyping easier. And good news is 07:38.520 --> 07:43.320 we are already funded for 2026. So maintenance and development is going on. 07:44.760 --> 07:51.800 Short notes on why we're using Rust. Rins are fairly simple. Reliability is the biggest reason 07:51.800 --> 07:57.960 here because for the energy grid we do need a reliable and preferably memory safe and efficient 07:58.840 --> 08:04.600 core code here. Deployment is pretty easy because you just get a single binary that you can just run. 08:04.600 --> 08:08.120 You don't need a job of virtual machine to be set up. You don't need a Python into a return 08:08.120 --> 08:13.560 to the right version. Something like that. And growing Rust adoption across the infrastructure 08:14.360 --> 08:17.800 and embedded systems communities also shows us that we are on the right path here. 08:19.160 --> 08:24.600 Don't get worried. Interoperability is easy. Communication just happens over HTTP as I mentioned. 08:24.600 --> 08:29.000 So virtually any program language can interact with them. Doesn't matter if you're implementing 08:29.000 --> 08:35.240 the business logic, the client logic, whatever you are, you can just interact with your HTTP. 08:35.240 --> 08:38.040 You don't have to mess with the Rust code if you're not familiar with that. 08:40.040 --> 08:46.200 Having said that, Stein will talk about how this actually is in practice where we are 08:46.200 --> 08:52.840 and what's going on in the future. Thank you, Max. So we just talked about open leader 08:52.920 --> 08:58.120 about open ADR, but who's actually using these projects, these protocols. 08:59.000 --> 09:03.720 One project is by the Dutch Research Institute ALAT, which I'm a part of. 09:03.720 --> 09:08.840 It's a Dutch Research Institute funded by the district service operators or the grid operators in 09:08.840 --> 09:14.520 an Netherlands. And the goal of the project or of this use case was making EV charges 09:14.520 --> 09:18.760 aware of the grid specifically. So the idea was that EV charges could be limited during those 09:18.760 --> 09:24.120 feet times to make sure the grid doesn't get overloaded. It's a part of the Dutch National 09:24.120 --> 09:29.960 Charging Infrastructure agenda, which is a government program to oversee the rollouts of EV charges 09:29.960 --> 09:36.760 in an Netherlands. How it works is a CPU pulse for data. It's a simple rest API so they do a 09:36.760 --> 09:43.480 get requests, they get events and they see when their charges are limited and then they apply those 09:43.480 --> 09:49.720 limits via a protocol like OCPP on their charges. As part of the project we also made custom 09:49.720 --> 09:54.360 tooling, made for the DSO so we can check that the events they put on the servers are actually 09:54.360 --> 10:02.760 correct. So currently the use case of mainly public EV charges, charging stations of what we call 10:02.760 --> 10:08.760 slow charges, stations or 11 kilowatts charges. We're also looking to broaden the scope to target 10:08.760 --> 10:13.640 homes via home energy management systems and because OpenADR is very high level, this is not 10:13.640 --> 10:20.840 actually that big of a problem. We're also looking to define a model for grid congestion or multiple 10:20.840 --> 10:29.000 levels so we can make more efficient use of grid resources. A-Latters also created a few open 10:29.000 --> 10:34.040 source projects. We have a Python SDK for the EV inside so if you want to connect to OpenADR 10:34.040 --> 10:39.800 you can do so in Python as well with the grid or charging compliance plugin and we're also making 10:39.800 --> 10:45.800 a grid graphical user interface mainly also for our own pilots and test use case so we can more 10:45.800 --> 10:54.280 easily manage those. There will open source so you can look them up on GitHub. Looking ahead, 10:54.280 --> 10:58.280 A-Latters later off are committed to this project like Mark said the funding is secured for 10:58.440 --> 11:05.880 2026. Please go and talk to us after this talk or reach out to us via Slack. There's a one 11:05.880 --> 11:12.040 monthly technical steering committee meeting of OpenADR if you want to get involved. You can also 11:12.040 --> 11:17.560 create a PR with good first issues, Max has made sure that there's a few of those and if you want 11:17.560 --> 11:22.600 to learn more about OpenADR you can join us at A-Latters for the European Flexibility Initiative 11:22.680 --> 11:30.440 in March which will include a vehicle to rent demo using OpenADR in Anim. That was a presentation 11:30.440 --> 11:32.920 if you have any questions we would be happy to answer them. Thank you. 11:40.360 --> 11:40.920 Right, yes. 11:41.880 --> 11:54.680 So your question was if OpenADR implements the client side of the server side? Do you think 11:54.680 --> 12:01.880 Max can tell us something about that? Yes, so as of now we're mostly focused on the server side, 12:01.880 --> 12:09.720 the VTN. We do also have a VEN client though the thing is the VEN is mostly a very abstract 12:11.160 --> 12:18.920 idea in OpenADR and so it's a very thin layer which mainly is abstracting away the HDP 12:18.920 --> 12:27.320 requests. So we do have sort of rust SDK if you want to say so but most of the work is the server side, 12:27.320 --> 12:38.120 the VTN. Yes, so would I invite? It already is some home appliances or something like that or 12:38.120 --> 12:46.440 is it just some of the future right now and you're only working with New Administration so 12:47.640 --> 12:53.000 so the question was whether it's already in home appliances or at the moment still limited to 12:53.000 --> 12:59.880 public charging and I guess that is something for you. So OpenADR is not really targeting specific 12:59.880 --> 13:05.320 devices so the idea is that it's very high level it will send a signal or kind of a description of 13:06.280 --> 13:11.640 a group of appliances need to do so what you could see in a future is a home energy management system 13:11.640 --> 13:18.760 for instance being able to ingest OpenADR signals and then going to your charger or your heat pump 13:18.760 --> 13:25.960 and then applying limits or the reading price is something like that. Any other questions? 13:25.960 --> 13:47.240 Oh yes. So the question was if this is mainly for Ethernet or Internet user also for over 13:47.240 --> 13:52.040 power lines if it will communicate over power lines so OpenADR is very specifically a rest API so it's 13:52.120 --> 13:58.760 very specifically an Internet protocol because it's so high level. The idea is that if you want to apply 13:58.760 --> 14:04.040 OpenADR to a device you can translate those and then you can use something like ZB may be your matter 14:04.040 --> 14:10.040 or whatever else S2 of course if you want to. Right, yes. 14:10.280 --> 14:18.760 So many times I'm building like a big people. So you play it like that or as a server, 14:19.160 --> 14:23.480 so the question is of what a virtual power plant would use a server or a client in that case that 14:23.480 --> 14:27.400 would be a v-end so it would be a client so you would can be communicating with the grid operator 14:27.400 --> 14:33.880 and then getting signals from their server and then using those, yes, that's up. Ah, too bad. 14:34.920 --> 14:36.280 Right thank you very much. 14:40.040 --> 14:42.040 Thank you. 15:10.040 --> 15:12.040 Thank you. 15:40.040 --> 15:42.040 Thank you. 16:10.040 --> 16:12.040 Thank you. 16:40.040 --> 16:42.040 Thank you. 17:05.040 --> 17:07.040 Just hold it like this. 17:08.040 --> 17:11.040 No, not just talking too. 17:12.040 --> 17:22.040 Don't go too close to the speaker because I'm back and let's stay here to test obviously on this one. 17:27.040 --> 17:30.040 I must have two minutes. 17:37.040 --> 17:40.040 Just kidding. 18:07.040 --> 18:30.040 You're not going to have to have control about the current we have to have control of the temperature so that's what we're doing. 18:30.040 --> 18:43.040 So what we're doing is we're modeling the temperature with that we can say hey we have this much thermal capacity left in our grid so we can increase the current which is allowed to use in the grid. 18:43.040 --> 19:03.040 Then we can use the model as well on this part in the monitoring so we can measure the current used in the grid we model again the temperature so we know the temperatures it had in real time and with that we can do again. 19:03.040 --> 19:17.040 Then again we can see like how warm can still get more or maybe a bit less so this way by having the cycle we are optimizing the use of our grid. 19:17.040 --> 19:26.040 And the better the calculation is the better we are optimizing use of our grid and we need thermal models. 19:26.040 --> 19:36.040 So we said critical assets which assets are we talking about we talk a lot about about transformers because that's one of our largest main models we have at the moment. 19:36.040 --> 19:49.040 But other importance models are as well the bear contactors the switch gear and cables and cables is done actually by one of our sibling teams in the company and Derek is here as well. 19:49.040 --> 19:53.040 So if you want to know a bit more about cables instead of transformers go to Derek. 19:56.040 --> 20:06.040 And this was a bit of theory but how to apply this models to add more clients to our grid we do chain analysis studies. 20:06.040 --> 20:15.040 What you see on this picture is a substation and what you see on this picture is a schematic overview of a substation. 20:15.040 --> 20:29.040 Don't worry you don't have to read it in the standard image but every of assets you can find here is put in a list and we put there behind it it's no one of value given by the manufacturer. 20:29.040 --> 20:37.040 And then you see that one at the assets is limiting all the other assets so the other assets have more capacity than one of the specific assets. 20:37.040 --> 20:50.040 And in this case it's a transformer so for this transformer we are running as well the our thermal model and we see that with current on the substation the thermal model. 20:50.040 --> 21:03.040 Yeah, we have some thermal capacity left so it can be overloaded it can be above 100% and still be good enough it won't decrease so fast. 21:03.040 --> 21:09.040 So we do that and then the whole substation can have more clients. 21:09.040 --> 21:22.040 This way we found already 600 MVA as potential space in our grid and that's more or less a whole city 100,000 households. 21:22.040 --> 21:27.040 But then spread over the whole area of course but to imagine. 21:27.040 --> 21:31.040 Sorry part we've built an LV dashboard low voltage. 21:31.040 --> 21:41.040 And here you see a bit more practice what the thermal model does in blue you see the current so currently we are always measuring the current. 21:41.040 --> 21:49.040 But why are we saying look at the temperature you can see the day around 23rd of May. 21:49.040 --> 21:55.040 There's a very high current but the tent in blue but in pink or red. 21:55.040 --> 22:03.040 It's quite a low value so maybe it was a cool day or something happens why the transformer didn't warm up as fast. 22:03.040 --> 22:14.040 So that day nobody would have had to be good off or whatever because we have thermal capacity left and on other days more on the beginning of this graph. 22:14.040 --> 22:22.040 Around May 4th we have a lower current but unexpectedly a higher temperature. 22:22.040 --> 22:32.040 So that day our transformer decrades a little bit faster and if you add all those degradations up on those days where it goes a little bit faster you expected. 22:32.040 --> 22:41.040 You can have this overview and every green dot is that the transformer decrades as fast as we expected to be. 22:41.040 --> 22:44.040 Red ones are going faster than we expected. 22:44.040 --> 22:50.040 So you can imagine for an management department if this whole area gets red we panic. 22:50.040 --> 22:56.040 You don't want it because then we are this year with preparing everything that we're building so it's counterproductive. 22:56.040 --> 23:04.040 So we have to have it mostly green but a few red are okay because we can manage we have people outside working on it. 23:04.040 --> 23:09.040 So this way we can balance if you are using the grid optimal. 23:09.040 --> 23:12.040 And now harm is going to tell you a bit more about the model. 23:12.040 --> 23:13.040 Yeah. 23:13.040 --> 23:15.040 Thank you. 23:15.040 --> 23:20.040 So on to the transformer model itself so how does it work? 23:20.040 --> 23:27.040 Well for our model we use the loading guide described by the IEC. 23:27.040 --> 23:36.040 And in the IEC they give a wide variety of international standards and one of the standards they give is about the thermal behavior of transformers. 23:37.040 --> 23:46.040 And we really took this standard and digitalized it into Python code and used that as the basis of our model. 23:46.040 --> 23:50.040 And the goal really is to predict the whole spot temperature. 23:50.040 --> 23:57.040 Because the whole spot temperature will be the temperature at the as the name suggests at the whole spot inside the transformer. 23:57.040 --> 24:04.040 And this will really, yeah, limit the transformer and prevent it from processing high loads. 24:05.040 --> 24:11.040 But mostly this will be located right above the windings deep inside a transformer. 24:11.040 --> 24:14.040 And we can't really access easily access it. 24:14.040 --> 24:18.040 So we can put a thermal meter there to verify how hard it gets. 24:18.040 --> 24:24.040 And to solve this the model also predicts the top oil temperature. 24:24.040 --> 24:31.040 And again at the name suggests this is the temperature at the top of the oil inside a transformer. 24:31.040 --> 24:41.040 This actually is really accessible. You can put a thermal meter there and you can measure it and you can verify where the model predicts the reality. 24:41.040 --> 24:47.040 And these temperatures are linked with each other and both predicted by the model. 24:47.040 --> 24:52.040 And to predict them the user needs to specify some inputs. 24:52.040 --> 24:55.040 You need to provide a load profile. 24:55.040 --> 24:59.040 So how much current is flowing through the transformer. 25:00.040 --> 25:03.040 We also need the ambient temperature profile. 25:03.040 --> 25:11.040 You can imagine that in a summer transformer is generally warmer compared to the wind when it's cold outside. 25:11.040 --> 25:16.040 And we need to have some transformers specific specifications. 25:16.040 --> 25:22.040 So we know that there's a wide variety of transformers and they all have their own thermal behavior. 25:22.040 --> 25:28.040 And if you want to make specific calculations to those transformers, we need some input from the user. 25:28.040 --> 25:38.040 And to make this easier we created general classes inside our model for general transformers that we know are out there. 25:38.040 --> 25:46.040 So for example the large power transformers or the smaller distribution transformers. 25:46.040 --> 25:53.040 And in those classes we incorporate that most of the logic specific to those type of transformers. 25:53.040 --> 25:57.040 So the user needs to specify as little as possible. 25:58.040 --> 26:01.040 So this is the idea of the model. 26:01.040 --> 26:04.040 But of course we have to verify where that is actually correct. 26:04.040 --> 26:07.040 Where do we predict the reality? 26:07.040 --> 26:11.040 And we can do that by comparing our model with measurements. 26:11.040 --> 26:14.040 And we actually see that it models it quite well. 26:14.040 --> 26:16.040 It most ranges. 26:16.040 --> 26:19.040 It follows the reality quite good. 26:19.040 --> 26:21.040 But we have some issues. 26:21.040 --> 26:26.040 And maybe the biggest one is that we lack data during high loads. 26:26.040 --> 26:29.040 So as Imka already mentioned the greatest fool. 26:29.040 --> 26:32.040 But we don't really see that at a temperature. 26:32.040 --> 26:37.040 And that's a problem because if you want to push the transformers. 26:37.040 --> 26:43.040 We need to know what actually happens when we push them and when we go over the limits. 26:43.040 --> 26:46.040 So this is it is an issue. 26:46.040 --> 26:47.040 And still is an issue. 26:47.040 --> 26:49.040 And we thought how could we solve this? 26:49.040 --> 26:54.040 How could we contact other people and tackle this problem? 26:55.040 --> 26:58.040 Well, we knew that we were not the only ones. 26:58.040 --> 27:00.040 The whole of the Netherlands was full. 27:00.040 --> 27:05.040 So we started collaborating with the DSOs and the DSOs in the Netherlands. 27:05.040 --> 27:08.040 And we shared data with each other. 27:08.040 --> 27:11.040 But there were more people out there who had the problem. 27:11.040 --> 27:14.040 Other countries or companies who had their own transformers. 27:14.040 --> 27:17.040 Or maybe people we didn't even think about. 27:17.040 --> 27:21.040 So we thought why not share it with the world. 27:21.040 --> 27:26.040 And make transformers a model and open source package. 27:26.040 --> 27:33.040 And we really wanted to create a community where we could tackle this together. 27:33.040 --> 27:36.040 And that's what we did. 27:36.040 --> 27:44.040 We hosted a release party where we invited all sorts of people who had an opinion about the model. 27:44.040 --> 27:50.040 We had data to share with us or maybe you only had a problem that it could be solved by the model. 27:50.040 --> 28:00.040 But before we actually could go open source, we had to modify the code base a bit. 28:00.040 --> 28:03.040 We of course had to push or package to pipe. 28:03.040 --> 28:08.040 So developers could use it in their own project. 28:08.040 --> 28:12.040 But we also had to add a lot of documentation. 28:12.040 --> 28:14.040 And we did this with MK docs. 28:14.040 --> 28:17.040 So we wrote our documentation in Markdown language. 28:18.040 --> 28:22.040 And we hosted this documentation on GitHub pages. 28:22.040 --> 28:25.040 And now we have our very own website. 28:25.040 --> 28:30.040 Yeah, describing all things concerning the model. 28:30.040 --> 28:36.040 And as you can see, for example, the theoretical documentation is shown here. 28:36.040 --> 28:39.040 So here you can find a which formula that I used in the model. 28:39.040 --> 28:44.040 Which types of transformers can be calculated with the model. 28:44.040 --> 28:46.040 But we also have a lot of examples. 28:46.040 --> 28:49.040 So all the features we added to the model are shown there. 28:49.040 --> 28:52.040 And you can find why you should want to use them. 28:52.040 --> 28:55.040 And how we can use them. 28:55.040 --> 29:00.040 And this is all to make it the user as easy as possible to use the model. 29:00.040 --> 29:04.040 But we do not only want users. 29:04.040 --> 29:10.040 We would also like to get your ID and get contribute to the model. 29:11.040 --> 29:15.040 And we did this by also showing how the model is made. 29:15.040 --> 29:20.040 So which decisions made for the code and where did we decide to make classes. 29:20.040 --> 29:21.040 For example. 29:21.040 --> 29:23.040 And we automated all of tests. 29:23.040 --> 29:25.040 So we have run by test. 29:25.040 --> 29:29.040 When a new contributor opens a new feature request. 29:29.040 --> 29:34.040 The by test are run and we check the code quality is still matching. 29:34.040 --> 29:37.040 The one we want is it's still good enough for us. 29:37.040 --> 29:39.040 And it's the security still good. 29:40.040 --> 29:44.040 And this gave some success stories. 29:44.040 --> 29:48.040 So stating came to us and they wanted to model the 29:48.040 --> 29:51.040 Horsepal temperature based on measurements to top oil. 29:51.040 --> 29:54.040 Yeah, to measurements of the top oil temperature. 29:54.040 --> 29:57.040 So they incorporated it themselves. 29:57.040 --> 30:01.040 And next is exchange data with us to help us verify. 30:01.040 --> 30:06.040 Yeah, where does the model work and where might it need some improvements. 30:06.040 --> 30:07.040 That tenant. 30:07.040 --> 30:11.040 They had different types of transformers compared to us. 30:11.040 --> 30:14.040 They had their transformers which were actively cooled. 30:14.040 --> 30:17.040 So ventilators were blowing air through it. 30:17.040 --> 30:19.040 And this could turn on and off. 30:19.040 --> 30:23.040 And we saw that this resulted in different thermal behavior. 30:23.040 --> 30:26.040 And tenant better the feature to also be able to. 30:26.040 --> 30:31.040 To model their transformers with our model or with our shared model, actually. 30:31.040 --> 30:35.040 And others gave a lot of feedback and tips on how we could improve the model. 30:35.040 --> 30:40.040 Which directions we still could explore. 30:40.040 --> 30:44.040 So the going open source really brought us a lot. 30:44.040 --> 30:50.040 And maybe yeah or take away messages if you have a work at a company. 30:50.040 --> 30:53.040 Which as transformers or you know use guys. 30:53.040 --> 30:58.040 Give data and try and see how much potential awaits you. 30:58.040 --> 31:08.040 Yeah. 31:08.040 --> 31:10.040 Questions. 31:10.040 --> 31:12.040 Yeah. 31:12.040 --> 31:15.040 And this is our get update. 31:15.040 --> 31:16.040 Yeah. 31:16.040 --> 31:27.040 What happens if transformer gets to warm. 31:27.040 --> 31:30.040 Around the winding there's a paper layer for insulation. 31:30.040 --> 31:32.040 And the oil protects the paper. 31:32.040 --> 31:38.040 But when the oil gets to warm the the chemical of the paper will change. 31:38.040 --> 31:43.040 And the more it happens the faster the the transformer will heat up next time. 31:43.040 --> 31:49.040 So once it goes around the tipping point, the it will be worse when it hits up. 31:49.040 --> 31:50.040 Yeah. 31:50.040 --> 31:53.040 So we have to stay away from that tipping point. 31:54.040 --> 31:57.040 Yeah. 31:57.040 --> 31:59.040 That's well one of the regular. 31:59.040 --> 32:02.040 Oh, can you change the oil? 32:02.040 --> 32:03.040 Yeah. 32:03.040 --> 32:06.040 It's one of the things which is done with maintenance. 32:06.040 --> 32:09.040 And they are now experimenting as well. 32:09.040 --> 32:11.040 If you can add things to the oil. 32:11.040 --> 32:14.040 So it will be even less. 32:14.040 --> 32:17.040 How do you go it in English? 32:17.040 --> 32:19.040 It will heat up less. 32:19.040 --> 32:22.040 So the oil is even more delayed. 32:23.040 --> 32:24.040 Yeah. 32:24.040 --> 32:25.040 Yeah. 32:31.040 --> 32:35.040 To be honest, I don't know that much of the transformers. 32:35.040 --> 32:40.040 But I think yeah, there's something done with maintenance where they look at the oil. 32:40.040 --> 32:44.040 So I don't know if it's really oil change or if it's an oil filtering. 32:44.040 --> 32:47.040 But they inspect the oil every six to seven years. 32:47.040 --> 32:51.040 And then they will do something about it if they find something. 32:52.040 --> 32:53.040 Yeah. 33:01.040 --> 33:07.040 If the oil is actively circulated through the system. 33:07.040 --> 33:08.040 I don't think so. 33:08.040 --> 33:10.040 I think it's one of the types you can have. 33:10.040 --> 33:16.040 But we have only onan and onan means it's just a big tank of oil. 33:16.040 --> 33:20.040 And onan is with air ventilation around it. 33:21.040 --> 33:22.040 The oil is natural. 33:22.040 --> 33:24.040 The onan stands for oil natural. 33:24.040 --> 33:26.040 So the oil is not pushed through it. 33:26.040 --> 33:28.040 It's just floating there. 33:30.040 --> 33:31.040 Yeah. 33:31.040 --> 33:37.040 If the outside number is so elemental and the H&J should just work on it. 33:37.040 --> 33:41.040 And if yes, then I would expect that the more extend they are, 33:41.040 --> 33:43.040 the more I open my zone. 33:43.040 --> 33:45.040 It's important for us. 33:45.040 --> 33:47.040 Yeah. 33:47.040 --> 33:48.040 Yeah. 33:48.040 --> 33:52.040 The question was whether the ambient temperature is really relevant. 33:52.040 --> 33:54.040 If I understand you correctly. 33:54.040 --> 33:55.040 Yeah. 33:55.040 --> 33:59.040 So what we see is that the effect of the ambient temperature is quite linear. 33:59.040 --> 34:10.040 So if the ambient temperature is 10 degrees higher, then the transformer will also get 10 degrees higher. 34:10.040 --> 34:12.040 So I would say it's quite relevant. 34:12.040 --> 34:15.040 More than the ambient temperature or maybe predicting it. 34:15.040 --> 34:20.040 If you want to know what happens to more out to a transformer is also quite relevant. 34:20.040 --> 34:24.040 Yeah. 34:24.040 --> 34:28.040 Yeah. 34:28.040 --> 34:29.040 Yeah. 34:29.040 --> 34:33.040 And you've got any feedback from the people doing the norms. 34:33.040 --> 34:37.040 Because basically you're enforcing the calculation based on the standards. 34:37.040 --> 34:39.040 So I would do the interface with that. 34:39.040 --> 34:43.040 Because at some point, does it start to become independent? 34:47.040 --> 34:50.040 We know that all the other repeated question. 34:50.040 --> 34:51.040 Yeah. 34:51.040 --> 34:55.040 So if you had a reaction of the IEC because we use the standard. 34:55.040 --> 34:56.040 No, we have not. 34:56.040 --> 35:02.040 But we know that all the TSO standards and all the other DSOs are using the same standard as well. 35:02.040 --> 35:06.040 So I think it's widely used in the market. 35:06.040 --> 35:07.040 Yeah. 35:07.040 --> 35:13.040 I see IEC happy that you are opening the calculation of the standard because at some point, 35:13.040 --> 35:16.040 if it's open-source and everyone can access it. 35:16.040 --> 35:21.040 And maybe you don't need to text anymore because you are like the calculation. 35:21.040 --> 35:22.040 Yeah. 35:22.040 --> 35:25.040 So if the IEC is happy, we don't know. 35:25.040 --> 35:27.040 Because we didn't have contact with them. 35:27.040 --> 35:33.040 But it's just a very, yeah, I didn't study science. 35:33.040 --> 35:39.040 But it's a, I have a few and the other team members that it's a very normal heat formula. 35:39.040 --> 35:43.040 So it's not rocket science. 35:43.040 --> 35:45.040 So it's a. 35:45.040 --> 35:50.040 So we did contact them before we went open-source to sort of. 35:52.040 --> 35:54.040 As far as I know we didn't. 35:54.040 --> 35:55.040 Oh, yeah. 36:00.040 --> 36:02.040 We have to check. 36:02.040 --> 36:04.040 But I think we didn't know. 36:04.040 --> 36:05.040 Yeah. 36:05.040 --> 36:06.040 I know. 36:06.040 --> 36:09.040 And when did you get your load profile from? 36:09.040 --> 36:14.040 I also think, yeah, we have to do things like this. 36:14.040 --> 36:15.040 Yeah. 36:15.040 --> 36:16.040 So temperature profiles. 36:16.040 --> 36:18.040 We are very afraid of data from. 36:18.040 --> 36:19.040 So temperature profiles. 36:19.040 --> 36:23.040 We have to have the economy, which is a Dutch weather. 36:23.040 --> 36:24.040 Service. 36:24.040 --> 36:28.040 So to say, so they have 20 around 20 stations in the Netherlands. 36:28.040 --> 36:30.040 So we take the temperature of them. 36:30.040 --> 36:33.040 And current is measured within our company. 36:33.040 --> 36:38.040 So on high voltage, everything is measured and on low voltage, 36:38.040 --> 36:43.040 I think, then to 20% I would say, is measured. 36:43.040 --> 36:45.040 So we can interpolate of that. 36:45.040 --> 36:47.040 Can we get your vehicle? 36:47.040 --> 36:52.040 There is a part of it. 36:52.040 --> 36:54.040 Open data on our website. 36:54.040 --> 36:59.040 So I think around 12 to 15 substations have been published online. 36:59.040 --> 37:02.040 So there is a group working on it. 37:02.040 --> 37:03.040 Yeah. 37:03.040 --> 37:04.040 Yeah. 37:04.040 --> 37:05.040 Yeah. 37:19.040 --> 37:21.040 What means MPA? 37:21.040 --> 37:26.040 That's how we express how much capacity a station has. 37:26.040 --> 37:30.040 I don't know. 37:30.040 --> 37:38.040 We tried to make it a bit more understandable by saying 600 MBA equals 100,000 households. 37:38.040 --> 37:45.040 So it's the energy provision for a very large, quite a large city. 37:45.040 --> 37:46.040 Yeah. 37:46.040 --> 37:49.040 Yeah. 37:49.040 --> 37:54.040 I'm sorry. 37:54.040 --> 37:59.040 I cannot hear it properly. 37:59.040 --> 38:14.040 If you want to open up more models. 38:14.040 --> 38:16.040 Yes. 38:16.040 --> 38:19.040 We are now currently working on the switch care model. 38:19.040 --> 38:21.040 And we very much have the ambition. 38:21.040 --> 38:24.040 Hopefully this year to have that open source as well. 38:24.040 --> 38:25.040 Yeah. 38:25.040 --> 38:26.040 Definitely. 38:26.040 --> 38:33.040 Because it really, the discussion we have mainly for now with all the DSOs is worth it. 38:33.040 --> 38:35.040 Because we are just a team of six people. 38:35.040 --> 38:39.040 And when we have all the 12 people within the company thinking about it. 38:39.040 --> 38:42.040 And then we triple four time it. 38:42.040 --> 38:44.040 It gives you a very valuable discussion. 38:44.040 --> 38:45.040 So yes. 38:45.040 --> 38:46.040 Definitely. 38:46.040 --> 38:49.040 Yeah. 38:49.040 --> 38:50.040 Yes. 38:50.040 --> 39:01.040 If we have a target capacity for our transformer. 39:01.040 --> 39:19.040 Yeah. 39:19.040 --> 39:20.040 Yeah. 39:20.040 --> 39:26.040 So the question was whether we looked at a specific transformer or at the complete range. 39:26.040 --> 39:27.040 Insighted. 39:27.040 --> 39:28.040 Yeah. 39:28.040 --> 39:30.040 The idea is to look at the complete range. 39:30.040 --> 39:35.040 So to look at the small ones and look at the really big ones. 39:35.040 --> 39:37.040 But we are a DSO. 39:37.040 --> 39:40.040 So the extremely large ones. 39:40.040 --> 39:43.040 I don't know if we also incorporated them. 39:43.040 --> 39:44.040 No. 39:44.040 --> 39:48.040 We worked last year for the first time on a three winding. 39:48.040 --> 39:50.040 So that's one of the larger ones. 39:50.040 --> 39:53.040 And our one of our biggest one is 100 MVA. 39:53.040 --> 39:56.040 And we did a test with that one. 39:56.040 --> 39:59.040 But that is by far our biggest. 39:59.040 --> 40:02.040 So I think tenet has even larger ones. 40:02.040 --> 40:03.040 Yeah. 40:03.040 --> 40:11.040 But you can model every type of transformer with our model. 40:11.040 --> 40:16.040 It depends on the transformer parameters you fill in. 40:16.040 --> 40:19.040 So now we added the three winding one. 40:19.040 --> 40:21.040 That had a bit of a model change. 40:21.040 --> 40:23.040 But after that you can model everything. 40:23.040 --> 40:30.040 If you know your parameters, which is 9 out of 10 times in the fed test, which is delivered 40:30.040 --> 40:31.040 which is a transformer. 40:31.040 --> 40:32.040 Yeah. 40:32.040 --> 40:33.040 Yeah. 40:33.040 --> 40:38.040 Because the vendor for a transformer I entrusted by what you are doing all of them 40:38.040 --> 40:43.040 or to the team that you better end to the end. 40:43.040 --> 40:44.040 Yeah. 40:44.040 --> 40:52.040 If the vendor is looking at this model as well, mainly related to the means 40:52.040 --> 40:55.040 I'm between failures. 40:55.040 --> 40:57.040 We they know about it. 40:57.040 --> 41:02.040 Especially I talked in person with a couple of people for switch gear. 41:02.040 --> 41:06.040 And they are interested. 41:06.040 --> 41:10.040 But as well, they say this is 100%. 41:10.040 --> 41:12.040 So that's what they can guarantee. 41:12.040 --> 41:14.040 And they have the standards as well of course. 41:14.040 --> 41:16.040 So I think for the thought testing, etc. 41:16.040 --> 41:19.040 They are using already this formula as well. 41:19.040 --> 41:24.040 So they are familiar with it. 41:24.040 --> 41:26.040 What did I want to say about it? 41:26.040 --> 41:27.040 Yeah. 41:27.040 --> 41:28.040 So they know. 41:28.040 --> 41:29.040 Yeah. 41:29.040 --> 41:34.040 And what we noticed is that the little bit all the ones, because we were over-dimensioning 41:34.040 --> 41:39.040 by design in the past a bit more than we are doing now. 41:39.040 --> 41:43.040 And technique has become a bit more precise. 41:43.040 --> 41:48.040 So we see that all their transformer can be overheated more or not overheated. 41:48.040 --> 41:51.040 Yeah, they can apply high percentage. 41:51.040 --> 41:55.040 And the newer ones have to stay a bit lower because we know more so we are more precise. 41:55.040 --> 41:57.040 So there's a difference in that. 41:57.040 --> 41:58.040 Yeah. 41:58.040 --> 41:59.040 Time's up. 41:59.040 --> 42:00.040 Thank you a lot. 42:00.040 --> 42:02.040 Thank you.