 I'm Jonas van Bogat, Digital Spread Elite at Alliander and also the open source office lead for Alliander's open source program office and I'm here today with Nico. I'm open source ambassador at the open source program office of Alliander and also acting as a CICD specialist. For those who are not familiar with Alliander, let me give you a short overview of who Alliander is. Alliander is the largest distribution system operated in the Netherlands and we develop and manage energy networks. We are nearly a sponsor for 6 million customer connections. We manage over nearly 90,000 km of electricity grid and over 40,000 km of gas grid. And as Alliander, it's our key task to make sure Dutch households and companies stay warm and have access to energy at all times. And today, in specific, we want to tell you more about the challenges that energy transition poses on companies like Alliander. And how open source projects like OpenStaff, the power grid model and a shape shifter help Alliander to tackle these challenges. Also we will provide you a short overview of the Lean Foundation energy, the open source foundation for the power system sector and hopefully that will also have some time at the end for some questions and discussions. So what challenges are distribution system operators like Alliander facing today? In 10 years time, we see that the way we live, work and travel won't be the same and it's very much necessary since fossil fuels are becoming more scarce and have a growing impact on our environment. However, luckily, we see that new initiatives are emerging around the world to conserve energy and make it more sustainable. And especially in the Netherlands, we see that the way we generate, distribute and use energy is changing very rapidly. And we see that fossil fuel power stations are closing down in the Netherlands in order to reduce CO2 emissions and we see a really strong increase in the use of solar panels, wind energy and also the emerge of electric cars and e-mobility. And this creates major challenges for distribution system operators like Alliander to support this transition. The Netherlands for example has nowadays one of the most reliable electricity grids in the world. However, this grid back in the day is not designed for the rapid developments that are happening today. We see especially in the Netherlands that the demand for electricity is growing so fast that it's driving the grid to its limits. There are already places in our grid that we unfortunately have to say no to new business consumers or business consumers that will expand their businesses and our need in electric power. And that's for a huge problem for us as a company and also for these customers. Even though we as Alliander have significantly increased our investment in expanding the grid in the last few years. And we see that it requires new investments and new smart solutions, especially in digitalisation to better use the capacity that's currently available in the grid. And one of these smart solutions we as Alliander are investing in and I would like to talk about today is the steering of supply in the manned when there's a risk that the maximum grid capacity is reached. And this can be due to for example a sunny day in summer where there's a lot of sun and solar energy and a lot of wind and wind energy but maybe only a little demand. Or in a period in winter where there's too much demand, think of a day where it's very cold but there's nearly no sun or wind production at that time. And this steering of supply and demand is also often referred to as congestion management. And congestion management can be in the form of direct control such as the curtailment of the excess of renewable energy or in the form of a market-based mechanism where incentives are used to enter supplies, grid parties to adjust their supply or demand. And when we as Alliander are able to active manage congestion we will create an electricity grid that has more room for new businesses, consumers and business consumers that want to expand their business and are in need of extra electric power. And to perform active congestion management new digital capabilities and solutions are required and open source plays a very important role here. We see that open source software and companies like Alliander can no longer be ignored. If you take studies from Gardner you see that open source software is used in mission critical environments by more than 95% of the organization worldwide. And honestly I speak to a lot of companies in this situation, I still have to find those 5% who don't. And many innovative technologies for building cloud native of digital applications are also open source. And I personally really believe that digitalization in today's world without open source software is really impossible. For example it's really hard to imagine in modern IT stack today without open source software. I would like to challenge you to raise your hands if not at least one of these applications is used in your organization as well. So what makes open source software particularly interesting for companies like Alliander? One we see that open source software tend to lead to more stable software. Two we see that open source software also allows for more incremental innovation with a limited budget you have. Three, and maybe even most importantly, we see open source software enables the power of the crowd. Innovation tends to come more often from outside your organization than inside your organization. And this is also true what we see at Alliander. We see that Alliander we are using more and more open source software. We are participating in more and more open source projects and we also continue to open source internal projects which is also of interest for broader scope and broader parties. And today we would like to focus in particular on the LF energy projects open staff, the power grid model and shape shifter and share with you how these three projects help building new digital capabilities for congestion management and help Alliander to integrate more renewables in the electricity grid. As I mentioned before the goal of active congestion management is to steer supply and demand in periods when there is a risk that the maximum grid capacity is reached. It is to better use the capacity that's available on the grid and create room for the new businesses and businesses that want to expand their business and our need of electric power. So let's get started with our open staff. For anticipating local congestion, it's essential that we as Alliander are able to forecast a load on the electricity grid for the next hours to days. And we are specifically interested to forecasting the load on the grid for the next 24 hours or four different eight hours. Because we need to know what the load on the grid will be so we can determine if there's a risk that the maximum grid capacity is reached. And it's also important for us that we know this in advance so there's still sufficient time to take action if there is a potential risk that that maximum grid capacity will be reached. And here the open staff project comes to play. Open staff provides the software for forecasting the load and generation of electricity on the electricity grid for upcoming hours and days. And open staff uses smart algorithms and specific machine learning models to produce forecasts based on measurements, weather forecasts, pricing on the energy market and many other important features. And this makes it a really important building block for active congestion management at Alliander. And besides active congestion management, open staff is also important and enabler for us, for grid safety analyzers and determine grid loss. So open staff itself is a Python package which you can use to make short term forecasts for the energy sector. So how does open staff work? Open staff provides a fully automated machine learning pipeline which creates and codify, creates energy forecasts for the energy sector. And I'm not familiar with machine learning pipelines. Let me give you a short overview. A machine learning pipeline is a way to codify and automate the workflow it takes to produce and deploy a machine learning model. And machine learning pipelines consist of multiple steps that do everything from data extraction of pre-processing to model training and to deployment. And open staff in specific encapsulates all the learned best practices of producing machine learning models for load forecast on electricity grids. And also, importantly, allows users to use this pipeline at skill. As input, open staff can use data sets for multiple data sources. And it is very easy to add new data sets or new sources depending on your needs. Key input data for Alliander, for example, are historical load and generation measurements, weather data, and data from energy markets. Next, open staff performs validation on this input data and combines the historical load and generation measurements with external predictors like weather data and the market prices. This data is then used to train a model for each location, where we do forecast. And here, in principle, any skyskid compatible machine learning can be used. The current model that we use at Alliander is the extreme gradient boosting algorithm. So, especially at Alliander, we use to make forecasts at substation level. And every substation has its own unique model, trade model. This because, in some substation, the influence of wind speed has a huge influence on the load of that substation. Due to that, there's a lot of wind protection behind that substation. For example, in other substations, this can be different. There's, for example, a lot of solar production in that area. And then you see that factors such as radiation has a way larger input. Our forecasts at Alliander are renewed every 50 minutes. But it also depends on your use case, what's an optimal time. Also, the forecasts we make are automatically consciously evaluated. And when we see that the forecast at specific location or envelope performs, then the underlying model will be re-trained or re-optimized. And of course, in the post-pressing step, we make also sure that the trained models and the forecasts are made available through APIs for other systems to be used. To give this a little bit more imagination, I took this image on this slide. And here you see one of the forecasts that's performed for one of Alliander substations. And what you see in this substation particular is that we make forecasts for different time horizons. For example, one hour ahead, four hours ahead, 24 hours ahead, and more. Also, as she looked closely, you can see that there's a lot of times when the energy production behind this substation is negative, which means that there's more energy produced in that area than consumed. And that's why that's also really true for this substation because this is an area where there's a lot of wind production behind this substation. In addition, OpenSTEP also provides forecasts for the amount of solar and wind energy generated. And since solar energy and particular wind energy have the greatest uncertainty, it really helps to have insight in the contribution of these components to the total load. And as already mentioned, OpenSTEP by itself is the supply-ton package. And there are multiple ways to implement OpenSTEP in your organization or in your application. In this slide, which I will go into detail, I'll give you an example of a reference implementation for OpenSTEP, where open source software, technology, and standards are used in a microservice-organized architecture that is optimized for cloud deployment. Although Alienda is still the largest contributor of OpenSTEP project, the community has come more diverse in the last few years. For example, RG, the French transmission system operator, and also a member of Linux Foundation Energy, start also contributing to OpenSTEP and are currently replacing their legacy-based forecasting solution based on the solution on OpenSTEP. Another contribution by OpenSTEP is made by the Agi University, who implemented the support for the ProLove model. And ProLove is an other holistic load forecasting model that can be used to forecast the load and electricity grid. And with ProLove included, OpenSTEP supports multiple models, and it's up to the user to choose which model best fits your needs. And we would love to see the OpenSTEP community grow further. So if anyone of you is interested in this project, I recommend you to check out the OpenSTEP website or dive into the OpenSTEP GitHub community, where you can get started with OpenSTEP. And this brings us to the end of the OpenSTEP introduction. And now it's time to introduce our second project, the PowerGrid model. And here I would like to give over to Nica. Yeah, thank you, Jonas. Yeah, that brings us to PowerGrid model and how that ties onto the data that we get from OpenSTEP. And PowerGrid model is a library, also a Python library, you can use for power system analysis. Basically doing calculations on the grid. Talking electricity grid here. And it uses the OpenSTEP data as input, together with characteristics of the grid, to come to a conclusion about what actually is the impact of the predictions. So if we go back to the use case we have at hand, active congestion management, we have the forecast of OpenSTEP, right? We see what's coming, and we want to determine if this is good or bad, or if we need to do something in any way. And then we use PowerGrid model for power system analysis to evaluate that. We take the data and see if it's an issue, or what could be an improvement. And mainly a PowerGrid model is used in this case for like a steady state calculation, but it also is used far widely using different calculation methods that have become part of the software. Jonas already introduced you about the scenarios, like maybe having wind or solar in a day. One of the ways you can deal with, say, curtailment is to maybe prevent wind farm from producing too much, right? They just turn the wind turbines out of the wind. That could be a way to mitigate congestion. And first you need to determine if there is an issue and in what way. Now, PowerGrid model was started at Alieander. We were needing a calculation engine that was fast, that was easy to integrate. And in the meantime, it has become a fundamental building block of everything we do as a distribution system operator, because it's not just this scenario, but oftentimes you find yourself in any analysis needing a bit of calculation on the grid, because there's a lot of things that affect the network. So we have it in a lot of ideas like here, grid planning, automatic network design, like we're trying to come up with new parts of the network, monitoring of where the asset allocation is happening, and now here, active condition management. And some of these include a human in the loop, but a lot of them don't, so it's really focused on automation. Now, a traditional workflow of power system analysis, say in the past, is a more manual step. Ever since we've been putting cables in the ground or power poles, there have been people running calculations, first by hand, then with calculators perhaps, and then with desktop software applications. So typically a grid architect would check out a grid representation and do the calculations and come up to some conclusions and that will end up somewhere and the action will be taken. But this is not an option if you do active congestion management. Like Jonah said, every five minutes we have to come up with an analysis, oftentimes involving multiple calculations, so the human needs to get out of the loop. And so we're moving to a world where we take something from a database out to a database or a queue or whatever other system, right? Data in, data out, and the integration is really important. And so we need models that are fit for that. And so why should you use PowerGrid model as you're moving to this more integrated use case? Well, let me tell you, these are some of the highlights of PowerGrid model where it strengths lie. So it has multiple power system calculations, functionalities, probably the most typical ones you were expecting. Also, this linear method, we have proven scientifically that it's within the error bounds that we expect from traditional methods like Newsome and Repson. So it's good enough, but it's a giant speed booster. We also can do low voltage grid where we have different loads on the different feeders of three voltage systems. And as I said, it's a high performance implementation in C++ to really use the best performance of the CPU it's running on. It has a Python API that's well tested to all these use cases and cross-platform. It's not focused on a UI, but I'll get back to that afterwards with the integration. And so this is more an overview of a typical scenario where we take the predictions sort of below, right? We have assume load, we have maybe measurements coming in depending on the use case. We have the topology of the grid which also changes over time. So it's an added complexity, right? One line might be out for maintenance. The topology changes over time and we have asset information. What type of cables are there? What transformers? How does that change the impedance and the calculations? All this is part of PowerGrid model and all the data interaction is according to various open standards. And we can determine also if there's an issue. So in this case, there's an example calculation it might be an overload situation but also you can think of voltages like there's an overvoltage, the voltage with all the solar coming in, the voltage is rise and you get a voltage level that's out of code. That's also something you have to account for. Now, the PowerGrid community is also growing. I think like OpenStaff or the Mincretributor at Alieander, but we have a good buy-in now from other grid operators and universities that are using it and testing it out to see if it's adding value to them. And through that discussion and our meetups we get a lot of insight in where it needs to go next, how we can improve it and also how we can integrate it with other solutions. For example, you see the PandaPower which is another great solution in this space which covers a more different use case, right? A lot of more models, more modeling where we have a specific need for integration and speed. But we work together and we have integrated PowerGrid model in PandaPower and the same for GridCal, it's a UI where you can use PowerGrid model as a calculation engine. So we're really looking to work together with other organizations, other open source projects. If you wanna know more about PowerGrid model, as said, it's an active community with great meetups. This is the team. You can reach them out on GitHub as well. Mail them and we've done webinars as well for international attendees. There might be some in the future, so please just reach out and maybe we can even set up something special for that. Now we have done the analysis, but what do we do with that? Because knowing there's an issue is not enough. You actually need to act on it. And that brings us to another project, ShapeShifter. ShapeShifter has an origin by the universal smart energy framework, USAF, which is a great accomplishing document framework like how should the energy market incorporating flexibility look like? A lot of great minds came together to come up with standards, with suggestions. And one part of that was the universal flex trading protocol, U of T P, designed to enable the trading of flexibility separate from other energy trading. This got put under the Linux Foundation Energy and it's now called ShapeShifter. And it's really about enabling trading of flexibility to mitigate issues like congestion. So in the overview we had of active congestion management, we now have done an analysis. You wanna know, there is an issue, we know there was an issue and we wanna address it some way and we can do that through ShapeShifter. I explain why, I wanna how. This is one of the diagrams that's from ShapeShifter. ShapeShifter is pretty much a really standard, a protocol, a way of working with a couple of libraries in Java and Python to enable you to implement this in your own solution. So it's not like this encompassing application, there might be one in the future, but not at the moment. The idea is that for flexibility we can use the flexibility that is already there. On the left you see some examples. So you can see with electric cars, for example, they can charge or even feed back into the grid at various times. If you can control this, this is a great opportunity for flexibility. But also you can think of a cooling cell if you have like these huge industrial refrigerators for vegetables, keeping them cool, for example. Of course they have to remain in a certain temperature bound, but within that you have some wiggle room to adjust and adjust the amount of energy consumption. Now this wiggle room can be considered flexibility and you can put a market around it. Now here's multiple reasons why you wanna work with flexibility for various causes. And our use case as a distribution system operator is to focus on constraint management. And we could do so without touching the other market acts that are already in place. And so the real principle here is to consider flexibility separate from other market mechanisms. Now here's an example where you see we have a cable that is typically overloaded. There's oftentimes there's a congestion happening there in shift shift it is called a congestion point. And this is addressed, right? We had a calculation from a power grid model that says here we have an issue, we need to resolve it. What we can do is on both sides of this congestion point trying to create an incentive, a financial incentive through this flexibility market to on the one hand create demand on the other hand create supply. Then on both sides the local balance changes and there's less need to transport energy across that congestion management. If there is then a difference financially that is compensated for and that is sort of the price to actually resolve this congestion. The main thing here is that they balance out or at least they can balance out so that for example the transmission system operator who looks at a higher level to maintain balance in the grid is not affected. They see a net zero, oh there's nothing it's just the DSO shifting load a bit. But this is a great resolver for this active congestion. Now shift shift it goes across a couple of steps. Contract, plan, validate, operate and settle. And the main thing you think of with active congestion management is happening in the validate phase. This is where all the trading is happening and the predictions are updated and new trades are coming in. The rest before that the contract and the plan are really to set up and identify congestion points, identify possible partners that can help as aggregators and then operate this during the process within the time bound where the congestion is mitigated is live there might still be updates or a live need for additional changes that is done there and finally the settle to make sure that all the balances match out and the payment is secured. And all the while power grid model and open step can really help during that validate phase to make that possible. Here you see an example of all these messages of flex requests and flex orders coming in or continuous evaluation. This is like what's happening in that validate phase. And so you see that open step is crucial for the forecast and power grid model is crucial to evaluate if these offers come in, would they resolve it? Do we need maybe adjust something? So it's a continuous iterative process and we need the other tools to make this work. Now, shapeshifter asset is a standard with a couple of libraries and it's really neutral about how it's implemented. So on the one hand you can have every distribution system operator dealing with every aggregator that is there, say an operator of EV charging points and with all the balanced responsible parties for the electricity market. Or you could even centralize it which might be more efficient. But of course there has to be a legal framework in place that has to be some buy in to enable that. That doesn't really matter. And shapeshifter brings a mature specification because there have been, I'll show multiple implementations in the field that are continuously working on improving the specification to be more clear and more easy to implement. There's a community of practice with how do you put this in practice and recommendations and I said there's a library in Java and Python to get you started. These are the two projects that are doing this now in the wild and we get a lot of experience from that refining shapeshifter protocol in the libraries. The fusion project in the United Kingdom and GoPaks in the Netherlands. And so you see as a result of that that we have already quite a substantial community of grid operators working on this but also consultancy company, DMVGL, Grid Imp. And a funny thing is that SP Energy Networks from Scottish Power and working in the UK is actually owned by Iberdrola. Maybe you recognize the three leaves. They're also here with the giant building here in Bilbao and you see them all over the street. So there's also a Bilbao tint to that. And who knows, maybe here in Spain there will be another shapeshifter project. Now if you want to know more about shapeshifter and the active community that have been there already since quite some time, you see some of the people and as we got more people coming in, I just added a couple of them. There might be even more to mention. And you can reach them of course on GitHub on the website. And we also have an online documentation you can view about the protocol with all the diagrams there and that should be a good explainer. And then I'll give it back to Jonas about ElevEnergy. Thank you Nika. So I also want to mention that all the three projects that we just introduced, OpenSapp, BioGrid, Model Line, Shapeshifter are hosted under the Bella of Linux Foundation Energy. And to just recap, ElevEnergy is an open source foundation for the power system sector hosted within the Linux Foundation. And the mission of ElevEnergy is to provide a neutral collaborative community to build shared digital investments around energy. So how does ElevEnergy particularly help us but also the OpenSapp, BioGrid, Model and Shapeshifter projects in particular? And the four things that really stand out for us. First, ElevEnergy provides us with a neutral ecosystem and open governance to host open source projects and collaborate with organization within and outside the energy sector. Second, ElevEnergy provides us also with to know how to best leverage and adopt open source best practices and practices around community management. And third, ElevEnergy also helps us with IP and legal related questions. And fourth, maybe even most importantly, ElevEnergy is also great network organizations to get in touch with other organizations working in the energy space and also interested in the open source domain as well. And we see that ElevEnergy has grown a lot in the last few years. ElevEnergy currently hosts 20 projects with many more in the pipeline. And this includes projects in the application domain like OpenSapp, but also projects in the edge domain like FlashPower. And today, ElevEnergy has over 50 members and this includes distribution system operators and transmission system operators like Alliander, Archie, Tenet, Stutnet, EnergyNet, but also universities and large corporates like Microsoft, Google, GE, and Shell. And this creates a very diverse and valuable community. And I also want to highlight three reports that has recently published and created in collaboration with Linux Foundation Research, where you can learn more about the developments that are happening in the energy space and particular how open source plays a role in these developments. And I don't have the time to go into them in detail, but if you want to read them and download them, you can use the QR code on this slide. And this brings us to the end of our presentation. And today, we have talked about the challenges we as distribution system operator are facing and we talked about the importance open source plays in Alliander's digitalization ambitions. We introduced the open source projects, OpenSapp, BioGrid model and ShapeShifter, and talked about how these projects in particular help Alliander to integrate more renewable energy in the electricity grid. Also, we give you some insight in how Linux Foundation Energy is helping us to grow these projects further. And I also want to highlight today and take this moment to also address the importance of these innovations. Because I believe that energy and moving to more renewable energy is not only key for the energy sector itself, but for all sectors. And I hope that this story can inspire you as well. Because I really believe that if we bring the great minds and the power of the crowd together, we can build great innovative solutions which will help us and the world to move to a more decarbonized world. And if you want to get to know more about our open source efforts or the open source efforts within the Linux Foundation Energy, please reach out to us or one, check out one of the links present on this slide. And I hope that we still have some time for some questions and discussions before we have to hand it over for the next presentation. Thank you very much. Yes. There's a lot of fluctuation there and you can't control the production. So one way to mitigate those problems is I hope to do it. So there's one way to provide that extra boost. Do you see or what's your perspective on let's say consumer electronic devices? We say, for example, a household where you have an AC unit that consumes a lot of electricity but it's not like time critical. It does not have to work at a certain time. Do you see there something that's for a project or are there any initiatives within the Linux Foundation that basically work on controlling that part of the... Yeah, I can answer. Let me repeat the question for online viewers. Basically, what projects are there, especially for the consumer side, people in households, they also have flexibility, right? AC vehicle to grid with electric vehicles. What opportunities are there to leverage that flexibility rather than businesses which have more opportunity to integrate in these pilot projects? As far as I know, there's not really anything going on at the moment. Ideally, I would say we need open standards to get something off the ground. But I think talking about ShapeShifter, we talk about aggregators having a giant role here. So you see some equipment manufacturers, say boiler producers, experimenting with the idea of making them internet connected and controllable. So you get maybe also for your car, you get a car and the car manufacturer or the charging point manufacturer has the ability to control it for you and sort of deal with flexibility trade. What I personally would really like to see is to enable households to take that into their own hands through automation to protocols. You see a lot of things happening with home automation. You see a lot of energy division coming on with the home assistant, for example, getting that connected to whether or not to aggregators or directly to this flexibility trade. I think would be really beneficial and would also help us as a grid operator, especially in the neighborhoods because talking to industrial parks and managing it there is quite different than actual neighborhoods. Yes, Adam. I saw one at 4 p.m. and 7 p.m. when people were pooping and that certainly helps drive behavior. There's also some quite interesting projects like the Octopus Intelligent for that project for EVs where you'll plug your car in, it'll charge for a moment, then it'll have a thing that will be like, nope, we're gonna put in these time spots where it's gonna be a much reduced range to sort of incentivize that time shipping. Yeah, that's very interesting you mentioned that. You see also Netherlands dynamic contracts are coming up and we as a distribution system are really happy because then you create also the incentivize to shift your load, also for consumers. So in Netherlands even the pricing is changing on hourly basis. So, and there are particular, even moments in time, especially on Sundays when there's a lot of sudden wind production that consumers can even retrieve money by consuming energy, which sounds strange. But that is really interesting if you have an electric car, you can charge your car and even make money out of it. And for us it's that's great because it's shifting the load to the peak and for the customers great as well because that will save them a lot of money. And I think there's also a huge opportunity to make it more easy for customers. Nowadays it really depends on the technical abilities of the customer itself to get this up and running. And it would be, I think there's a great opportunity for making that more customer friendly and more automated. So that happens automatically. Yeah, that's a really good point to make. So for us as a great company, it's important that we guarantee great safety at all times. And especially with forecast model that other machine learning is always, that never 100% true. So there's always a margin of error you have to take into account. And for us in making sure that safety is analyzed, we make sure that there's enough space to guarantee the safe marches. But what we want to do, and we hope to improve those models to get more reliable in the future that we can, how do you say in English, blow sharper to the wind? Or I'm not sure if that's transpired well. Yeah, closer to the edge maybe. Closer to the edge. I think that's a better pronunciation. So there's no huge opportunity there. But it's something you have to keep in mind and be aware of. Yeah, and one thing to add for the example we just gave in ShapeShifter, there's also already a mechanism to cope with that. So we have the operate phase. And if say the predictions were too optimistic, there's still the opportunity to act live and control with the market to mitigate it regardless. So there is some safety system already in the protocol. Yes, question in the back. Yeah, so yes and no. So we, for example, the ShapeShifter came into existence because there was no particular standard in that particular domain. And that's one of the reasons why we make this standard openly available. And we hope to bring this standard also to, on the long run, to a more international standard. So we're really happy to see that it's also adopted in the UK. So it's gonna be a brighter adoption. Furthermore, we use internally also a lot of standards already heavily used in the energy sector. Examples are ICIM standard, which also a lot of our open source projects rely on. And many other standards like OpenADR, Project we didn't talk about today, that's a particular focus on that standard. So I think, and I believe that open standards also key enabler of the energy transition, but also particularly important for the open source projects. Okay, that brings us to the end. Thank you very much for your attendance and for your questions. Thank you.