Welcome everyone to the very first session of our Optimise Masterclass Series! Our first series of masterclasses, about our “CONNECT” product, received really great feedback, so Robert and I thought: let’s continue with a new series! We all know the EV ecosystem is evolving fast. But as infrastructure scales, so do the challenges. Overloaded grids, rising connection costs, tight grid operator regulations and driver frustration from slow or unreliable charging. At FlexiCharge, we believe the foundation for solving all of that is load management. And today's sessions is your blueprint for getting it right, from static versus dynamic approaches to unlocking energy flexibility and future-proofing your business. And now I'm excited to introduce the person who spent years at the core of these challenges, our CTO and co-founder, Robert Brame. He's not just an expert, he's been in the trenches, you could definitely say, building solutions that are shaping the future of EV charging. Robert, over to you, please. Yes, thanks, Chris. Yeah, thanks for the introduction, as Chris said. I'm very excited to now start talking about what FlexiCharge has been established for, or what was the core idea of establishing FlexiCharge, which is a load in energy management. And yeah, we are starting a masterclass series now on that topic. It's not just one sort of class, it will be a series of classes. And today, we are going to talk a bit about the basics. So we're going to look into what is actually the difference between load and energy management, because I can see in a sort of daily communication and discussions and meetings that often that load and energy management is just thrown in one pot. And it's not really differentiated what one or the other means. So I'm going to shed some light into that. And then we are going to talk about the architectures, load and energy management architectures, there's different ways of how you can do that. So we're going to talk about how you can structure, how you can structure, sort of set up your technology stack, your load energy management architecture from a cloud-based solution to local solutions. But what we see a lot in the market, and what actually makes a lot of sense is that there are some sort of hybrid approaches where you have local load management combined with cloud-based energy management. And that's actually what we have built with the FlexiCharge Harmony platform to be able to provide technology to actually make that happen. I look into the sort of advantages and disadvantages between cloud and local load management. There are obviously advantages, disadvantages. And then I have three, four slides where we go a bit into, not depth, but I give you an overview of what we can offer with our Harmony platform. And that should just be a teaser for the masterclasses, which will follow the one we are doing today, where we also dive deeper into topics. We'll dive deeper into specific requirements, applications, and so on and so forth. But let's start with the basics. But let's start with the basics. Really start at the beginning. And that is, what is actually the difference between load and energy management? And I mean, not only do I experience that that is often sort of a topic which is not pretty clear to our customers or also people in the industry, but also internally, but also internally, I keep teaching this to our sales department that load management is actually dealing with kilowatt. So it's dealing with power or with load, which is measured in the unit kilowatt as power load on a particular system or on a particular good connection point. And this is what load management is about. And this is what load management is about. So load management is real-time control, real-time management of power instantaneously at any point in time. So the best example is, that's typically what we see. And this is where load management systems come into play. You have a grid connection point. You have a charging infrastructure. You might have a building, maybe PV system, battery. All connected to one grid connection point. And that grid connection point has a particular limit. Let's say one megawatt. And a load management system is ensuring that any point in time, that maximum limit is not exceeded. Okay. This is sort of the core basic application of load management. Obviously, there are other, sort of applications, which you would also still sort of group under load management. For example, if you have a grid connection, if you have a soft limit, you have a contract with your energy supplier, where you have a specific limit, which is more like a soft limit, but not the hard limit, that you're not exceeding a specific maximum, average limit, but not the hard limit, but not the hard limit, but not the hard limit. So the minimum limit is also covered in load management. But still, again, the important fact, the underlying fact is load management is dealing with kilowatts. Okay. In contrast, energy management is dealing actually with kilowatt hours. And kilowatt hour, that's the unit for energy. So in energy management, it's not instantaneous, sort of per second, sort of decisions, but it's planning ahead. So it's basically trying to operate a system, an energy system, charging infrastructure, with respect to a specific objective. And that objective. And that objective is often cost optimization. Yeah. So for example, you have a fleet of EVs, which charge overnight, you know, what is the energy demand for that fleet, for each vehicle, you know, the arrival, you know, the departure time, you know, the maximum charge current per vehicle, per charging station, you know, the grid limit, and so on and so forth. And you take all that into consideration. And then you make a plan, and then you make a plan how to charge these vehicles efficiently, cost efficient, based, for example, on a dynamic energy price tariff. So that is energy management. In the same way, energy management is you have a PV system, and you want to charge your EVs on the basis of PV prognosis. So you have them over the day, you know, during the day, you know, during the day, there's a lot of PV production, you know, how much energy will be produced from the PV, how much energy is required from the EVs. And then you try to match that, right? And to actually, in that case, the objective is, for example, to increase self consumption from the PV panel. But the fundamental difference between load and energy management is load management is about, the power, power, how much energy management is doing, shifting, moving demand into production times or into cheap energy pricing into times where energy is cheap and it's dealing with energy with kilowatt hour. So it's very important to understand this difference because often people talk, we are doing load and energy management, but at the end, they only do load management. Or some people say, okay, we do load and energy management or we do energy management, but then it's not really energy management. So I think it's really important to understand this difference also with respect to how to set up your technology stack and what architectures can actually be applied to provide true load and energy management. I have a quick summary here again for the differences. So load management controls power in real time, maintains local constraints, for example, the maximum grid connection limit, optimize charging in real time. So it typically gets the maximum out of a grid connection in any load situation. For example, you share a charging infrastructure with a building and then the available power for charging dynamically scales with the building consumption so that the maximum grid connection is never exceeded. It acts in real time to demand response, for example, DNO control. So that in any point in time, the system can respond and for example, lower the grid in times where it's maybe frequency deviations. It's optimal dispatching of available local power sources, for example, PV system. So don't mix that up with energy management. But what's also possible is to increase charge power if there suddenly is PV power available. So without doing any planning, it's still possible to increase self -consumption using load management. And typically load management is solely behind the grid connection point. So typically your load management system is behind sort of the meter providing load management, ensuring some constraints or limits are not fulfilled. Energy management, as I said, plans, demands, and sources for optimal usage. It schedules charging for, for example, for off-peak tariff windows. So when energy is cheap or it decides when to charge or discharge your on-site battery. If you use that for peak shaving, if you use that for peak shaving, it can plan in not to charge the battery immediately, no matter what the price is, to actually charge it when it's cheap. It can actually dispatch solar power or buying from the grid. So that plays along what I said with charging batteries. It can provide fleet-wide total energy forecasts, which is really, really important in terms of fleet depots. And it is, really important in terms of fleet-wide. And it can be local behind the grid connection, but often, and we'll bring up some examples where that is the case, it's also global. So then it's not for a particular site, but it's for a number of sites. This is in particular the case, for example, if you provide ancillary services, so frequency curtailment reserve, for example, with your charging infrastructures, then obviously you don't do that for a single site. A site is a number of charging stations behind the grid connection, but you actually do it for your whole sort of portfolio of your complete charging infrastructure, right? Across sites. So important to understand is that, of course, there's an interplay between energy and load management. And that is that, of course, energy management has to take into consideration that there are local constraints which are managed by the local load management system. And the local load management system is actually executing, for example, a schedule to charge a fleet of electric vehicles based on, yeah, based on a schedule which has been made by an energy management system, which, for example, gets the information from day-ahead prices of energy. So it's important to understand that load management is the fundamental base layer for energy management. And we'll see later, I'll show some examples, for example, I'll show some examples where you can clearly see that very concrete, for example, again, ancillary services. It's pretty clear that if you get, for example, the energy management decides to reduce load at this particular site, that the load management system actually needs to execute that. I will get into that in a moment. And as I said, often energy management can be or is local, but very often it's cloud-based. So then you have local control or local load management for reasons which we will see in a moment. And the energy management is on top of the load management system. And one of the main reasons is that load management often is time critical, yeah, or latency critical. If you have dynamic load management, yeah, or sort of time response for an energy management system is not as sort of tight as it is for the load management system. So now I mentioned a few times. I said local, I said cloud. So let's try to distinguish what is meant by cloud setup or by cloud-based load or energy management and what is actually meant by local load energy management and what are actually the advantages and disadvantages between the two here. I mentioned a few points. I mentioned a few points. I mentioned a few points already. So one of the, or some of the, or one of the main differentiators is of course latencies. I mean, control signal latencies, but there are a few others as well. Cloud setup means you're charging infrastructure. There's no local controller. The charging infrastructure is connected to some sort of server, typically over OCPP and the complete load management and maybe also the overlaying the load energy management is coming from the cloud. The cloud means it's basically some sort of server. It's an OCPP endpoint, an OCPP server. And then on top of that, there's built some load energy management. You see that a lot in CPMS systems, but obviously, I mean, FlexiCharge, our Harmony platform also provides it completely vendor independent. Important, no local controller. It has advantages. But it also has disadvantages, but it also has disadvantages, as we will see in a second. On the other hand, we have local load management, where you have a local controller, a local, in case of FlexiCharge, it's a local gateway. And that gateway is running all the load management algorithms, everything local. So there's, in principle, in terms of the load management. And you can also, if you have the energy management on the local controller, there's no dependency on the cloud. So the main purpose or the main differentiator here is the local controller. No dependency on the internet connection at the local site and really low latencies. So in the next slide, I have a little table to compare both what is advantages and disadvantages of cloud versus local load management. Let's start with local load management and the actual advantages. You have local control, obvious. Therefore, you have low latency control for dynamic load management. So what does dynamic load management mean? dynamic load management means that the limit for a charging infrastructure is a number of charging stations can change dynamically depending on other loads behind the grid connection point. So typical example is a building shares a grid connection with 10 high power chargers and the building load is, of course, quite dynamic, volatile. And the dynamic load management system continuously reacts or adjusts how much is actually available for the charging infrastructure. That is the dynamic load management case. And you can clearly see that latencies play a critical role here. Specifically, for example, if you have heat pumps in the building and then you have suddenly a heat pump with some inrush current, high consumption, and then the system needs to react on that. Yeah. Another advantage of local load management is if you, and that's possible with also our Harmony platform and our gateway, if you connect the chargers by Modbos, then it's completely independent of the OCPP communications. Stack where the chargers are connected to the CPMS via OCPP. And you can actually have the load management or the energy management system on a completely different communication sort of channel, which in this case would be Modbos. Obviously, what is also possible is to do local load management with OCPP. It's also an option. It's also an option. But with a Modbos or a separate connection, you have real clear independence of OCPP communication and, for example, the load management communication. Obviously, cybersecurity is very, very important effect for charging infrastructures and critical infrastructure. So obviously, a local controller is less prone to cyber security. cyber attacks. So there's less attack vectors. There's still some, but it's far less than when you do everything out of cloud. We'll see that in a second. And then, obviously, it's independent of the van. So the internet access, it can still operate dynamic load management, even thought the internet connection breaks, which is completely different or in contrast to cloud-based load. management. Yeah. For cloud-based load management. What are the advantages? It's a fast integration into existing technology stacks. If you take a load management out of a CPMS, you might even have it sort of implicit. But even if you go with an open solution like a FlexiCharge Harmony platform, then it's really, we always say, 10, 15 minutes thing to connect your charging infrastructure. So it's a fast integration into the Harmony platform via OCPP and then forward the connection to the CPMS, which you typically have. So it's really simple to integrate into existing technology stacks. Therefore, it's really fast setup. It's quick rollout. You have no hardware assets in the field. It's suitable for many load management applications, such as static load management or non-critical load management. applications where, for example, applications play play a less critical role. For example, what I said, if you want to optimize over a 15-minute time window, the average consumption of your charging infrastructure. I consider that non-critical load management because it's a soft limit, maybe a contractual limit. Of course, you still don't want to cross it and it will not cross it, but it has not like a deterministic real-time sort of time dependency as if you would do, I mean, dynamic load management, for example, measuring building consumption and so on and so forth. So a lot can be done from the cloud. Now we come to the disadvantages. Local load management, yeah, you have additional hardware in the field. You have additional hardware installation and setup work to be done. You sometimes have, I mean, if you go with Modbus, you have another sort of charger communication channel into the charger, which could be a risk factor. And when it comes to cloud-based load management, the disadvantage is it's not really suitable or only limited suitable for dynamic load management. What, for example, can be done with dynamic load management is some sort of peak shaving. If it's not too time critical, for example, with batteries, if the batteries provides a local, provides an API interface, so it can be integrated into the Harmony platform and it can be controlled for the cloud. But everything which is dynamic where you try to control different assets from the cloud, it's actually tricky. It's tricky also because of, yeah, to maintain the connections and because of latencies, yeah. Obviously, there's additional cyber security risk. I mean, you have quite a wide vector of, or many vectors of attack. It depends on the van, the right area network, so the internet connection, SIM card provider, all these things. So if that breaks or fails, it's a lot of points of failure, actually, in the system, which you might not even have under control. And in terms of fallback strategies or fallback scenarios, in cloud-based load management, you typically have only hard fallback scenarios. So that means if the connection to the charges break, they fall into some hard -coded fallback scenario, which is, for example, they go into free rending, they have a limit, a maximum charge limit, hard-coded, hard-coded, they fall back to, so you ensure that you're not exceeding, for example, the maximum grid connection limit. Yeah, that is some of the disadvantages. So now, if we go a bit more into some of the hard facts about cloud load management, sorry, local load management and cloud-based load management, so where does the load management algorithm run? And now I can only speak for FlexiCharge and our gateway. It runs completely local. So if you install a local controller, a local gateway from FlexiCharge, then the complete algorithm, everything runs local. In the cloud, obviously, everything runs out of the cloud somewhere in the internet on a server, yeah, which is a proxy, for example, in case of OCPP proxy, in case of FlexiCharge, but if it's a CPMS player or provider which has a load management, then obviously it's also running on the server. In terms of latency control loops, yeah, 50 milliseconds is quite high, but there's something which can be done specifically via Modbus, but where I say it's more like 100, 200 milliseconds up to a second. But what is more important here is that it's deterministic. So, and that is one of the major differences. It's real time. And what does actually, what does it mean real time? Real time, I mean, if I ask, and I know that from my university times, real time means that something is deterministic. So if you say you have one second round trip delay or one second round trip delay to, for example, get power meter data or to get what sent a command to the charger, it can be one second still real time, but it's deterministic. It means it's always never more than one second. And that sort of thing. And that sort of thing you cannot control in a cloud-based environment. It can be 300 milliseconds if you have a really great network connection, if you have a great SIM card provider network. Yeah, but it can also be several seconds. And it's not, you cannot control. And that is quite a challenge for specifically dynamic load management systems. Now, if we talk static, it's all good. We can cope with it. Also non-time critical load management, it's all good. But as soon as it's time critical, if it's dynamic, then that plays a crucial role. Yeah. And I know the arguments are so often, yeah, yeah, typically it's 300 milliseconds. And I say, yeah, it's okay. It's typically 300 milliseconds. But one out of a thousand times, it's not 300 milliseconds, but then it's two seconds. And that can actually cause issues. Connectivity requirements. And so, yeah, it's a really solid fallback solution in case of a network outage. And specifically, if we talk about DNO control, so that network operators can reduce load in terms of grid stabilities, or maybe even facing close to blackout. So, yeah, there's no other people. But if we talk about conditions where internet connection, maybe it's already down, there's no way around local load management. You still need to be able to control your charges in order to maintain grid stability. So that is another core advantage of local load management. For the cloud-based part, yeah, it just requires a stable internet link because you really rely on it. Reliability is a risk. Yeah, you have additional hardware in the field, no problem, completely agreed to. But typically hardware these days is solid. But still, yeah, you can have power loss of the gateways and you have hard fallback scenarios and so on and so forth. But that's the risk. Again, in a cloud-based situation, yeah, you have internet outage, you have backend downtime, you have SIM card data, CAP overruns, you have a SIM card provider outage and so on and so forth. That's a risk one needs to consider. Security risk, and briefly mentioned it, can be hit physically or from the building network. So some attacks or the middle attacks or whatever can come from the building network. You can hit it physically, so somebody can grab the gateway, rip it out or do something with it. But an advantage here is only vulnerable to failure at the single site. And in contrast, cloud-based, it can be hit remotely. So typically, so typically you have, you do not need to be at the site to attack it. And if it's hijacked, then typically it is not only one site going down, but actually it's every connected site which you have, which can be attacked. Yeah, so these are the sort of more hard facts for load and cloud-based management, local load management and cloud-based load management. Again, it's not to say one is better than the other. Important to understand is there is particular applications where local load management is required. So there's not many ways around it. And there's not many ways around it. And there's not many applications where cloud-based load management is just fine. Okay. But I think it's really, really important to be aware of which comes in when. What we do offer for FlexiCharge for our Harmony platform is we actually offer both. So we have a solution for cloud, sorry, for here on the left-hand side, cloud-based, load management. That means, as I explained, we connect the chargers to our Harmony platform. Our Harmony platform provides load and energy management. How that is happening, that will all be part of some of the next masterclasses we'll cover. But important is Harmony or FlexiCharge is not a backend. So it's not a CPMS. It's just a load energy management platform, which is open and independent when the is open and independent. So it can provide load management for every charging station. So it can provide load management for every charging infrastructure, which let's say for any CPMS. So the important here is if it's provided, if we provide cloud-based load management, then it's an OCP-based load management solution. And for every charging station, connect, we forward the connection to the CPMS. We forward the connection to the CPMS. And load management is provided by OCP-based. As said, we're going to dive deeper in one of the next masterclasses into that. In contrast, on the other side, the same algorithms, the same load management algorithms, we run in Harmony for the cloud-based load management. We can also run on the local device, which is our gateway. In that case, and there's a local controller installed, which hosts our load energy management or load management algorithms. And the complete load management is local. You can then integrate into local assets like batteries, PV systems, the chargers. Or Modbus, also batteries, there's different interfaces we get to that. But basically, we provide everything, both concept cloud and load management out of one platform. Now I want to, because basically we started talking about load and energy management. I think load management in terms of what can be, what comes from the cloud or cloud-based load management. So where cloud-based load management makes sense. So where local load management makes sense. Or what are the different sort of advantages and disadvantages or application layers there. I want to now sort of close the circle a bit where we started from. So where do load and energy management actually interplay? Because I said earlier that what is typically the case is that you have local load management or you have a load management layer. And then on top of the load management layer. And then on top you have the energy management layer. And I want to show you some examples where that is happening on the basis of our Harmony platform. And one of the examples here is, for example, scheduling of fleets of EVs based on spot market day ahead prices. In that case, at the bottom here on the left-hand side, you have charging infrastructure. So you have a charging infrastructure. Maybe for trucks, maybe for buses with some local load management system, which ensures at any point in time that local grid constraints are not exceeded or constraints are fulfilled. And then on top out of our Harmony platform, you have a cloud-based dynamic scheduling, which receives information from vehicles, arrival time, departure time, departure time, departure time, departure time, departure time, It receives information about state of charge when they arrive at the charging infrastructure from the local controller and from the charging station. And then it combines or compiles all this information and actually makes a schedule for charging out of it. It gets a day ahead price from the spot market and then actually start scheduling EVs. EVs. And this is a perfect example where load and energy management interplay where you have a local load management system and then you have an overlaying energy management system. We'll go into depth about this particular application also in one of the next master classes. So there's a lot of dynamics obviously also involved. That means if you have sent a schedule and then the prognosis is, because as I said, because as I said in energy management, there's also sort of some prognosis, some ahead planning. So if you have one vehicle, it needs to know when will the next vehicles arrive. It needs to an estimation on what will be the state of charge, the energy management, the energy demand of these vehicles and how to cope with the situation that these predictions are maybe wrong. A vehicle comes with a lower or higher state of charge and it needs to rescate. Or they're suddenly a higher state of charge. Or they're suddenly a different building consumption coming into place, which has not been predicted and so on and so forth. We'll get into that. The other example is our aggregate product, which is a virtual power plant layer, which provides a technology stack to integrate charging infrastructure into an auxiliary service. services and flexibility markets. And in that case, you have a number of charging infrastructure. So you have 100 charging infrastructure. They all have local load management, local control. And through our Harmony platform out of the cloud, all the data, all the charging utilization, charge power and so on is collected. Based on that, there's a capacity prognosis. So how much flexibility can actually be provided by the charging infrastructure across all the sites. That prognosis is actually giving to market integrators, BRPs, BSPs, which then will place bids, for example, to provide frequency control and reserve. And at some point, the system will receive a reduction. So how much flexibility. I can actually be provided by whatever one megawatt. And then the VPP layer, which is the energy management layer in Harmony in the cloud will actually distribute this reduction of one megawatt based on the flexibility prognosis of each of the sites towards the local site load management systems and the load management system will just execute it. So in the best case, I can give you an example. So in the best case, I can give you an example. There's a site, it has a typical limit of one megawatt. There's a 800 kilowatt load right now from charging buses or trucks or vehicles. And then it gets an activation to reduce the load by, let's say, 100 kilowatt. So what the system, what the local load management system will be doing is to reduce the load. So put the limit down from one megawatt. To actually 700 kilowatt to actually 700 kilowatt because it needs to reduce per hundred kilowatt. And it was previously load was 800. So it needs to reduce per hundred. And then distribute the available power, 700 kilowatt in an optimal way based on, for example, schedules based on policies, like it can be state of charge based and so on and so forth. And there you see where it's, it's, it's, it's for, for my opinion, the, the best example where load and energy management actually interplay. So as, as, as, as my last slide, I just want to give you a bit of a teaser about harmonies or our, our load and energy management platform, the open load and energy management platform, and what are the different components of, of our platform. One of the challenges, one of the challenges, and we'll talk about that in, in, in, in one of our next, uh, masterclasses. And we, uh, just to, to, to, to repeat it because we have actually had a series on, on, on connect our, our, our middleware interconnect middleware and data broker, which is an essential point in the load management, um, architecture, a load energy management architecture to ensure we can abstract all the different, different communication, uh, physical communication, uh, physical communication protocols to charging stations. So the idea here is it doesn't matter if you connect a charging station by OCPP or Modbus, the connect data broker will actually abstract it to the overlaying, uh, sort of, uh, system components optimize, which is the local load management per site and aggregate, which is the energy management VPP player, uh, layer, sitting on top of, uh, sitting on top of optimize. Yeah. Um, so the, the connect is basically the, the, the, the data, the, the, the abstraction of connection and interconnection of, of, of assets. So again, it doesn't matter if you'll connect the charter by OCPP or Modbus, a battery by API MQTT or Modbus or a power meter by Modbus, the connect layer will actually abstract it towards the, the optimize. And the beauty in this architecture is that our load management. So optimize is completely decoupled, uh, and encapsulated from the actual physical communication, for example, with a, with, with an asset, like a charging infrastructure, a charging station or a battery. Um, and then on top. So as I said, optimize that's a load management. So as I said, I'm going to integrate assets into the flexibility markets. Yeah. Um, and we're going to dive deeper into this in, in, in, in, in, in, in this series of masterclasses today, it was just to give you the first, uh, impression and lay a bit, uh, the foundation of, um, yeah, diamond, uh, sort of a cloud-based local load management, energy management, load management, what are the differences and, um, why it's so important to, you have a platform and an architecture which can actually cover both cases. Yeah. Because you do not want to have, uh, different vendors for all the different options, which are, which are available here. And, uh, with that, I think we are, we are having 45 minutes now almost filled. So this has really been a long masterclass. Um, thanks for joining and, uh, yeah. See you the next time. Yeah. And maybe just giving, uh, just a 30 seconds for eventually people in the audience to, to send the questions. I hope, uh, I didn't mention it beginning, but I think people are used to, uh, the option to ask questions, uh, no question for now. Um, otherwise you can always reach out afterwards. And again, as Robert said, it's the first of a series. Um, we haven't set the date for the next one, but we can expect to have it, uh, probably sometime in the end of May. So look forward. Uh, I think you gave a hint to a couple of topics, so I'm sure there will be, uh, a bit of everything for everyone. And, um, thank you, Robert. Thank you for joining and, uh, have a great day, everyone. Yeah. Bye for now. Bye. Bye. Bye.